v3.16.3

v3.16.2
v3.16.1
2025-09-17 23:44:27 +00:00 · 2021-05-06 14:55:03 -04:00 · 2021-04-08 18:09:31 -04:00 · 2021-03-24 18:24:20 -04:00 · 2021-03-19 15:45:32 -04:00 · 2021-03-08 22:44:44 -05:00
120 changed files with 7168 additions and 4229 deletions
--- a/145
+++ b/145
@@ -1,5 +1,9 @@
 Instructions for compiling cpuminer-opt for Windows.

+Thwaw intructions nay be out of date. Please consult the wiki for
+the latest:
+
+https://github.com/JayDDee/cpuminer-opt/wiki/Compiling-from-source

 Windows compilation using Visual Studio is not supported. Mingw64 is
 used on a Linux system (bare metal or virtual machine) to cross-compile
@@ -24,79 +28,76 @@ Refer to Linux compile instructions and install required packages.

 Additionally, install mingw-w64.

-sudo apt-get install mingw-w64
+sudo apt-get install mingw-w64 libz-mingw-w64-dev


 2. Create a local library directory for packages to be compiled in the next
   step. Suggested location is $HOME/usr/lib/

+$ mkdir $HOME/usr/lib
+
 3. Download and build other packages for mingw that don't have a mingw64
   version available in the repositories.

 Download the following source code packages from their respective and
 respected download locations, copy them to ~/usr/lib/ and uncompress them. 

-openssl
-curl
-gmp
+openssl: https://github.com/openssl/openssl/releases

-In most cases the latest vesrion is ok but it's safest to download
-the same major and minor version as included in your distribution.
+curl: https://github.com/curl/curl/releases

-Run the following commands or follow the supplied instructions.
-Do not run "make install" unless you are using ~/usr/lib, which isn't
-recommended.
+gmp: https://gmplib.org/download/gmp/

-Some instructions insist on running "make check". If make check fails
-it may still work, YMMV.
+In most cases the latest version is ok but it's safest to download the same major and minor version as included in your distribution. The following uses versions from Ubuntu 20.04. Change version numbers as required.

-You can speed up "make" by using all CPU cores available with "-j n" where
-n is the number of CPU threads you want to use.
+Run the following commands or follow the supplied instructions. Do not run "make install" unless you are using /usr/lib, which isn't recommended.
+
+Some instructions insist on running "make check". If make check fails it may still work, YMMV.
+
+You can speed up "make" by using all CPU cores available with "-j n" where n is the number of CPU threads you want to use.

 openssl:

-./Configure mingw64 shared --cross-compile-prefix=x86_64-w64-mingw32
-make
+$ ./Configure mingw64 shared --cross-compile-prefix=x86_64-w64-mingw32-
+$ make
+
+Make may fail with an ld error, just ensure libcrypto-1_1-x64.dll is created.

 curl:

-./configure --with-winssl --with-winidn --host=x86_64-w64-mingw32
-make
+$ ./configure --with-winssl --with-winidn --host=x86_64-w64-mingw32
+$ make

 gmp:

-./configure --host=x86_64-w64-mingw32 
-make
-
-
+$ ./configure --host=x86_64-w64-mingw32
+$ make

 4. Tweak the environment.

-This step is required everytime you login or the commands can be added to
-.bashrc.
+This step is required everytime you login or the commands can be added to .bashrc.

-Define some local variables to point to local library. 
+Define some local variables to point to local library.

-export LOCAL_LIB="$HOME/usr/lib"
+$ export LOCAL_LIB="$HOME/usr/lib"

-export LDFLAGS="-L$LOCAL_LIB/curl/lib/.libs -L$LOCAL_LIB/gmp/.libs -L$LOCAL_LIB/openssl"
+$ export LDFLAGS="-L$LOCAL_LIB/curl/lib/.libs -L$LOCAL_LIB/gmp/.libs -L$LOCAL_LIB/openssl"

-export CONFIGURE_ARGS="--with-curl=$LOCAL_LIB/curl --with-crypto=$LOCAL_LIB/openssl --host=x86_64-w64-mingw32"
+$ export CONFIGURE_ARGS="--with-curl=$LOCAL_LIB/curl --with-crypto=$LOCAL_LIB/openssl --host=x86_64-w64-mingw32"

-Create a release directory and copy some dll files previously built.
-This can be done outside of cpuminer-opt and only needs to be done once.
-If the release directory is in cpuminer-opt directory it needs to be
-recreated every a source package is decompressed.
+Adjust for gcc version:

-mkdir release
-cp /usr/x86_64-w64-mingw32/lib/zlib1.dll release/
-cp /usr/x86_64-w64-mingw32/lib/libwinpthread-1.dll release/
-cp /usr/lib/gcc/x86_64-w64-mingw32/7.3-win32/libstdc++-6.dll release/
-cp /usr/lib/gcc/x86_64-w64-mingw32/7.3-win32/libgcc_s_seh-1.dll release/
-cp $LOCAL_LIB/openssl/libcrypto-1_1-x64.dll release/
-cp $LOCAL_LIB/curl/lib/.libs/libcurl-4.dll release/
+$ export GCC_MINGW_LIB="/usr/lib/gcc/x86_64-w64-mingw32/9.3-win32"

+Create a release directory and copy some dll files previously built. This can be done outside of cpuminer-opt and only needs to be done once. If the release directory is in cpuminer-opt directory it needs to be recreated every time a source package is decompressed.

+$ mkdir release
+$ cp /usr/x86_64-w64-mingw32/lib/zlib1.dll release/
+$ cp /usr/x86_64-w64-mingw32/lib/libwinpthread-1.dll release/
+$ cp $GCC_MINGW_LIB/libstdc++-6.dll release/
+$ cp $GCC_MINGW_LIB/libgcc_s_seh-1.dll release/
+$ cp $LOCAL_LIB/openssl/libcrypto-1_1-x64.dll release/
+$ cp $LOCAL_LIB/curl/lib/.libs/libcurl-4.dll release/

 The following steps need to be done every time a new source package is
 opened.
@@ -110,13 +111,73 @@ https://github.com/JayDDee/cpuminer-opt/releases

 Decompress and change to the cpuminer-opt directory.

-
-
-6. Prepare to compile
+6. compile

 Create a link to the locally compiled version of gmp.h

-ln -s $LOCAL_LIB/gmp-version/gmp.h ./gmp.h
+$ ln -s $LOCAL_LIB/gmp-version/gmp.h ./gmp.h
+
+$ ./autogen.sh
+
+Configure the compiler for the CPU architecture of the host machine:
+
+CFLAGS="-O3 -march=native -Wall" ./configure $CONFIGURE_ARGS
+
+or cross compile for a specific CPU architecture:
+
+CFLAGS="-O3 -march=znver1 -Wall" ./configure $CONFIGURE_ARGS
+
+This will compile for AMD Ryzen.
+
+You can compile more generically for a set of specific CPU features if you know what features you want:
+
+CFLAGS="-O3 -maes -msse4.2 -Wall" ./configure $CONFIGURE_ARGS
+
+This will compile for an older CPU that does not have AVX.
+
+You can find several examples in README.txt
+
+If you have a CPU with more than 64 threads and Windows 7 or higher you can enable the CPU Groups feature by adding the following to CFLAGS:
+
+"-D_WIN32_WINNT=0x0601"
+
+Once you have run configure successfully run the compiler with n CPU threads:
+
+$ make -j n
+
+Copy cpuminer.exe to the release directory, compress and copy the release directory to a Windows system and run cpuminer.exe from the command line.
+
+Run cpuminer
+
+In a command windows change directories to the unzipped release folder. to get a list of all options:
+
+cpuminer.exe --help
+
+Command options are specific to where you mine. Refer to the pool's instructions on how to set them.
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Create a link to the locally compiled version of gmp.h
+
+$ ln -s $LOCAL_LIB/gmp-version/gmp.h ./gmp.h

 Edit configure.ac to fix lipthread package name.

--- a/Makefile.am
+++ b/Makefile.am
@@ -129,7 +129,7 @@ cpuminer_SOURCES = \
  algo/lyra2/allium.c \
  algo/lyra2/phi2-4way.c \
  algo/lyra2/phi2.c \
-  algo//m7m/m7m.c \
+  algo/m7m/m7m.c \
  algo/m7m/magimath.cpp \
  algo/nist5/nist5-gate.c \
  algo/nist5/nist5-4way.c \
@@ -192,6 +192,11 @@ cpuminer_SOURCES = \
  algo/sm3/sm3-hash-4way.c \
  algo/swifftx/swifftx.c \
  algo/tiger/sph_tiger.c \
+  algo/verthash/verthash-gate.c \
+  algo/verthash/Verthash.c \
+  algo/verthash/fopen_utf8.c \
+  algo/verthash/tiny_sha3/sha3.c \
+  algo/verthash/tiny_sha3/sha3-4way.c \
  algo/whirlpool/sph_whirlpool.c \
  algo/whirlpool/whirlpool-hash-4way.c \
  algo/whirlpool/whirlpool-gate.c \
--- a/README.md
+++ b/README.md
@@ -89,7 +89,7 @@ Supported Algorithms
                          lyra2h        Hppcoin
                          lyra2re       lyra2
                          lyra2rev2     lyra2v2
-                          lyra2rev3     lyrav2v3, Vertcoin
+                          lyra2rev3     lyrav2v3
                          lyra2z        
                          lyra2z330     Lyra2 330 rows, Zoin (ZOI)
                          m7m           Magi (XMG)
@@ -122,6 +122,7 @@ Supported Algorithms
                          tribus        Denarius (DNR)
                          vanilla       blake256r8vnl (VCash)
                          veltor        (VLT)
+                          verthash      Vertcoin
                          whirlpool
                          whirlpoolx
                          x11           Dash
@@ -134,7 +135,7 @@ Supported Algorithms
                          x14           X14
                          x15           X15
                          x16r          
-                          x16rv2        Ravencoin (RVN)
+                          x16rv2        
                          x16rt         Gincoin (GIN)
                          x16rt-veil    Veil (VEIL)
                          x16s          Pigeoncoin (PGN)
--- a/README.txt
+++ b/README.txt
@@ -1,6 +1,10 @@
 This file is included in the Windows binary package. Compile instructions
 for Linux and Windows can be found in RELEASE_NOTES.

+This package is officially avalable only from:
+ https://github.com/JayDDee/cpuminer-opt
+No other sources should be trusted.
+
 cpuminer is a console program that is executed from a DOS or Powershell
 prompt. There is no GUI and no mouse support.

@@ -10,7 +14,7 @@ miners. The source code is open for anyone to inspect. If you don't trust
 the software, don't use it.

 Choose the exe that best matches you CPU's features or use trial and
-error to find the fastest one that doesn't crash. Pay attention to
+error to find the fastest one that works. Pay attention to
 the features listed at cpuminer startup to ensure you are mining at
 optimum speed using the best available features.

@@ -31,26 +35,31 @@ https://en.wikipedia.org/wiki/List_of_Intel_CPU_microarchitectures
 https://en.wikipedia.org/wiki/List_of_AMD_CPU_microarchitectures


-Exe file name              Compile flags              Arch name
+Exe file name                Compile flags              Arch name

-cpuminer-sse2.exe            "-msse2"                 Core2, Nehalem   
-cpuminer-aes-sse42.exe       "-march=westmere"        Westmere
-cpuminer-avx.exe             "-march=corei7-avx"      Sandybridge, Ivybridge
-cpuminer-avx2.exe            "-march=core-avx2 -maes" Haswell*
-cpuminer-avx512.exe          "-march=skylake-avx512"  Skylake-X, Cascadelake-X
-cpuminer-zen.exe             "-march=znver1"          AMD Ryzen, Threadripper
-cpuminer-avx512-sha-vaes.exe "-march=icelake-client"  Icelake*
+cpuminer-sse2.exe            "-msse2"                   Core2, Nehalem   
+cpuminer-aes-sse42.exe       "-march=westmere"          Westmere
+cpuminer-avx.exe             "-march=corei7-avx"        Sandybridge, Ivybridge
+cpuminer-avx2.exe            "-march=core-avx2 -maes"   Haswell(1)
+cpuminer-avx512.exe          "-march=skylake-avx512"    Skylake-X, Cascadelake
+cpuminer-avx512-sha.exe      "-march=cascadelake -msha" Rocketlake(2)
+cpuminer-avx512-sha-vaes.exe "-march=icelake-client"    Icelake, Tigerlake(3)
+cpuminer-zen.exe             "-march=znver1"            AMD Zen1, Zen2
+cpuminer-zen3.exe            "-march=znver2 -mvaes"     Zen3(4)

-* Haswell includes Broadwell, Skylake, Kabylake, Coffeelake & Cometlake. 
-Icelake is only available on some laptops. Mining with a laptop is not
-recommended. The icelake build is included in anticipation of Intel eventually
-releasing a desktop CPU with a microarchitecture newer than Skylake.
+(1) Haswell includes Broadwell, Skylake, Kabylake, Coffeelake & Cometlake. 
+(2) Rocketlake build uses cascadelake+sha as a workaround until Rocketlake
+    compiler support is avalable.
+(3) Icelake & Tigerlake are only available on some laptops. Mining with a
+    laptop is not recommended.
+(4) Zen3 build uses zen2+vaes as a workaround until Zen3 compiler support is
+    available. Zen2 CPUs should use Zen1 build.

 Notes about included DLL files:

 Downloading DLL files from alternative sources presents an inherent
 security risk if their source is unknown. All DLL files included have
-been copied from the Ubuntu-20.04 instalation or compiled by me from
+been copied from the Ubuntu-20.04 installation or compiled by me from
 source code obtained from the author's official repository. The exact
 procedure is documented in the build instructions for Windows:
 https://github.com/JayDDee/cpuminer-opt/wiki/Compiling-from-source
--- a/77
+++ b/77
@@ -65,6 +65,83 @@ If not what makes it happen or not happen?
 Change Log
 ----------

+v3.16.3
+
+#313 Fix compile error with GCC 11.
+Incremental improvements to verthash.
+
+v3.16.2
+
+Verthash: midstate prehash optimization for all architectures.
+Verthash: AVX2 optimization.
+GBT: added support for Bech32 addresses.
+Linux: added CPU frequency to benchmark log.
+Fixed integer overflow in time calculations.
+
+v3.16.1
+
+New options for verthash:
+  --data-file to specify the name, and optionally the path, of the verthash
+              data file, default is "verthash.dat" in the current directory.
+  --verify to perform the data file integrity check at startup, default is
+           not to verify data file integrity.
+Support for creation of default verthash data file if:
+   1) --data-file option is not used,
+   2) no default data file is found in the current directory, and,
+   3) --verify option is used.
+More detailed logs related to verthash data file.
+Small verthash performance improvement.
+Fixed detection of corrupt stats caused by networking issues.
+
+v3.16.0
+
+Added verthash algo.
+
+v3.15.7
+
+Added accepted/stale/rejected percentage to summary log report.
+Added warning if share counters mismatch which could corrupt stats.
+Linux: CPU temperature reporting is more responsive to rising temperature.
+A few AVX2 & AVX512 tweaks.
+Removed some dead code and other cleanup.
+
+v3.15.6
+
+Implement keccak pre-hash optimization for x16* algos.
+Move conditional mining test to before get_new_work in miner thread.
+Add test for share reject reason when solo mining.
+Add support for floating point, as well as integer, "networkhasps" in
+RPC getmininginfo method.
+
+v3.15.5
+
+Fix stratum jobs lost if 2 jobs received in less than one second.
+
+v3.15.4
+
+Fixed yescryptr16 broken in v3.15.3.
+
+v3.15.3
+
+Yescrypt algos now use yespower v0.5, a little faster.
+New implementation of sha256 using SHA CPU extension.
+Replace Openssl with SPH for sha256 & sha512.
+AVX512 optimization for sha256t & sha256q.
+Faster sha256t, sha256q, x21s, x22i & x25x on CPUs with SHA without AVX512.
+AVX512+SHA build for Intel Rocketlake added to Windows binary package. 
+
+v3.15.2
+
+Zen3 AVX2+VAES optimization for x16*, x17, sonoa, xevan, x21s, x22i, x25x,
+allium.
+Zen3 (AVX2+SHA+VAES) build added to Windows binary package.
+
+v3.15.1
+
+Fix compile on AMD Zen3 CPUs with VAES.
+Force new work immediately after solving a block solo.
+
+
 v3.15.0

 Fugue optimized with AES, improves many sha3 algos.
--- a/algo-gate-api.c
+++ b/algo-gate-api.c
@@ -15,8 +15,6 @@
 #include <stdbool.h>
 #include <memory.h>
 #include <unistd.h>
-#include <openssl/sha.h>
-//#include "miner.h"
 #include "algo-gate-api.h"

 // Define null and standard functions.
@@ -279,9 +277,11 @@ void init_algo_gate( algo_gate_t* gate )
 #pragma GCC diagnostic push
 #pragma GCC diagnostic ignored "-Wimplicit-function-declaration"

-// called by each thread that uses the gate
+// Called once by main
 bool register_algo_gate( int algo, algo_gate_t *gate )
 {
+  bool rc = false;
+
  if ( NULL == gate )
  {
    applog(LOG_ERR,"FAIL: algo_gate registration failed, NULL gate\n");
@@ -290,104 +290,108 @@ bool register_algo_gate( int algo, algo_gate_t *gate )

  init_algo_gate( gate );

-  switch (algo)
+  switch ( algo )
  {
-    case ALGO_ALLIUM:        register_allium_algo        ( gate ); break;
-    case ALGO_ANIME:         register_anime_algo         ( gate ); break;
-    case ALGO_ARGON2:        register_argon2_algo        ( gate ); break;
-    case ALGO_ARGON2D250:    register_argon2d_crds_algo  ( gate ); break;
-    case ALGO_ARGON2D500:    register_argon2d_dyn_algo   ( gate ); break;
-    case ALGO_ARGON2D4096:   register_argon2d4096_algo   ( gate ); break;
-    case ALGO_AXIOM:         register_axiom_algo         ( gate ); break;
-    case ALGO_BLAKE:         register_blake_algo         ( gate ); break;
-    case ALGO_BLAKE2B:       register_blake2b_algo       ( gate ); break;
-    case ALGO_BLAKE2S:       register_blake2s_algo       ( gate ); break;
-    case ALGO_BLAKECOIN:     register_blakecoin_algo     ( gate ); break;
-    case ALGO_BMW512:        register_bmw512_algo        ( gate ); break;
-    case ALGO_C11:           register_c11_algo           ( gate ); break;
-    case ALGO_DECRED:        register_decred_algo        ( gate ); break;
-    case ALGO_DEEP:          register_deep_algo          ( gate ); break;
-    case ALGO_DMD_GR:        register_dmd_gr_algo        ( gate ); break;
-    case ALGO_GROESTL:       register_groestl_algo       ( gate ); break;
-    case ALGO_HEX:           register_hex_algo           ( gate ); break;
-    case ALGO_HMQ1725:       register_hmq1725_algo       ( gate ); break;
-    case ALGO_HODL:          register_hodl_algo          ( gate ); break;
-    case ALGO_JHA:           register_jha_algo           ( gate ); break;
-    case ALGO_KECCAK:        register_keccak_algo        ( gate ); break;
-    case ALGO_KECCAKC:       register_keccakc_algo       ( gate ); break;
-    case ALGO_LBRY:          register_lbry_algo          ( gate ); break;
-    case ALGO_LYRA2H:        register_lyra2h_algo        ( gate ); break;
-    case ALGO_LYRA2RE:       register_lyra2re_algo       ( gate ); break;
-    case ALGO_LYRA2REV2:     register_lyra2rev2_algo     ( gate ); break;
-    case ALGO_LYRA2REV3:     register_lyra2rev3_algo     ( gate ); break;
-    case ALGO_LYRA2Z:        register_lyra2z_algo        ( gate ); break;
-    case ALGO_LYRA2Z330:     register_lyra2z330_algo     ( gate ); break;
-    case ALGO_M7M:           register_m7m_algo           ( gate ); break;
-    case ALGO_MINOTAUR:      register_minotaur_algo      ( gate ); break;
-    case ALGO_MYR_GR:        register_myriad_algo        ( gate ); break;
-    case ALGO_NEOSCRYPT:     register_neoscrypt_algo     ( gate ); break;
-    case ALGO_NIST5:         register_nist5_algo         ( gate ); break;
-    case ALGO_PENTABLAKE:    register_pentablake_algo    ( gate ); break;
-    case ALGO_PHI1612:       register_phi1612_algo       ( gate ); break;
-    case ALGO_PHI2:          register_phi2_algo          ( gate ); break;
-    case ALGO_POLYTIMOS:     register_polytimos_algo     ( gate ); break;
-    case ALGO_POWER2B:       register_power2b_algo       ( gate ); break;
-    case ALGO_QUARK:         register_quark_algo         ( gate ); break;
-    case ALGO_QUBIT:         register_qubit_algo         ( gate ); break;
-    case ALGO_SCRYPT:        register_scrypt_algo        ( gate ); break;
-    case ALGO_SHA256D:       register_sha256d_algo       ( gate ); break;
-    case ALGO_SHA256Q:       register_sha256q_algo       ( gate ); break;
-    case ALGO_SHA256T:       register_sha256t_algo       ( gate ); break;
-    case ALGO_SHA3D:         register_sha3d_algo         ( gate ); break;
-    case ALGO_SHAVITE3:      register_shavite_algo       ( gate ); break;
-    case ALGO_SKEIN:         register_skein_algo         ( gate ); break;
-    case ALGO_SKEIN2:        register_skein2_algo        ( gate ); break;
-    case ALGO_SKUNK:         register_skunk_algo         ( gate ); break;
-    case ALGO_SONOA:         register_sonoa_algo         ( gate ); break;
-    case ALGO_TIMETRAVEL:    register_timetravel_algo    ( gate ); break;
-    case ALGO_TIMETRAVEL10:  register_timetravel10_algo  ( gate ); break;
-    case ALGO_TRIBUS:        register_tribus_algo        ( gate ); break;
-    case ALGO_VANILLA:       register_vanilla_algo       ( gate ); break;
-    case ALGO_VELTOR:        register_veltor_algo        ( gate ); break;
-    case ALGO_WHIRLPOOL:     register_whirlpool_algo     ( gate ); break;
-    case ALGO_WHIRLPOOLX:    register_whirlpoolx_algo    ( gate ); break;
-    case ALGO_X11:           register_x11_algo           ( gate ); break;
-    case ALGO_X11EVO:        register_x11evo_algo        ( gate ); break;
-    case ALGO_X11GOST:       register_x11gost_algo       ( gate ); break;
-    case ALGO_X12:           register_x12_algo           ( gate ); break;
-    case ALGO_X13:           register_x13_algo           ( gate ); break;
-    case ALGO_X13BCD:        register_x13bcd_algo        ( gate ); break;
-    case ALGO_X13SM3:        register_x13sm3_algo        ( gate ); break;
-    case ALGO_X14:           register_x14_algo           ( gate ); break;
-    case ALGO_X15:           register_x15_algo           ( gate ); break;
-    case ALGO_X16R:          register_x16r_algo          ( gate ); break;
-    case ALGO_X16RV2:        register_x16rv2_algo        ( gate ); break;
-    case ALGO_X16RT:         register_x16rt_algo         ( gate ); break;
-    case ALGO_X16RT_VEIL:    register_x16rt_veil_algo    ( gate ); break;
-    case ALGO_X16S:          register_x16s_algo          ( gate ); break;
-    case ALGO_X17:           register_x17_algo           ( gate ); break;
-    case ALGO_X21S:          register_x21s_algo          ( gate ); break;
-    case ALGO_X22I:          register_x22i_algo          ( gate ); break;
-    case ALGO_X25X:          register_x25x_algo          ( gate ); break;
-    case ALGO_XEVAN:         register_xevan_algo         ( gate ); break;
-    case ALGO_YESCRYPT:      register_yescrypt_algo      ( gate ); break;
-    case ALGO_YESCRYPTR8:    register_yescryptr8_algo    ( gate ); break;
-    case ALGO_YESCRYPTR8G:   register_yescryptr8g_algo   ( gate ); break;
-    case ALGO_YESCRYPTR16:   register_yescryptr16_algo   ( gate ); break;
-    case ALGO_YESCRYPTR32:   register_yescryptr32_algo   ( gate ); break;
-    case ALGO_YESPOWER:      register_yespower_algo      ( gate ); break;
-    case ALGO_YESPOWERR16:   register_yespowerr16_algo   ( gate ); break;
-    case ALGO_YESPOWER_B2B:  register_yespower_b2b_algo  ( gate ); break;
-    case ALGO_ZR5:           register_zr5_algo           ( gate ); break;
+    case ALGO_ALLIUM:       rc = register_allium_algo        ( gate ); break;
+    case ALGO_ANIME:        rc = register_anime_algo         ( gate ); break;
+    case ALGO_ARGON2:       rc = register_argon2_algo        ( gate ); break;
+    case ALGO_ARGON2D250:   rc = register_argon2d_crds_algo  ( gate ); break;
+    case ALGO_ARGON2D500:   rc = register_argon2d_dyn_algo   ( gate ); break;
+    case ALGO_ARGON2D4096:  rc = register_argon2d4096_algo   ( gate ); break;
+    case ALGO_AXIOM:        rc = register_axiom_algo         ( gate ); break;
+    case ALGO_BLAKE:        rc = register_blake_algo         ( gate ); break;
+    case ALGO_BLAKE2B:      rc = register_blake2b_algo       ( gate ); break;
+    case ALGO_BLAKE2S:      rc = register_blake2s_algo       ( gate ); break;
+    case ALGO_BLAKECOIN:    rc = register_blakecoin_algo     ( gate ); break;
+    case ALGO_BMW512:       rc = register_bmw512_algo        ( gate ); break;
+    case ALGO_C11:          rc = register_c11_algo           ( gate ); break;
+    case ALGO_DECRED:       rc = register_decred_algo        ( gate ); break;
+    case ALGO_DEEP:         rc = register_deep_algo          ( gate ); break;
+    case ALGO_DMD_GR:       rc = register_dmd_gr_algo        ( gate ); break;
+    case ALGO_GROESTL:      rc = register_groestl_algo       ( gate ); break;
+    case ALGO_HEX:          rc = register_hex_algo           ( gate ); break;
+    case ALGO_HMQ1725:      rc = register_hmq1725_algo       ( gate ); break;
+    case ALGO_HODL:         rc = register_hodl_algo          ( gate ); break;
+    case ALGO_JHA:          rc = register_jha_algo           ( gate ); break;
+    case ALGO_KECCAK:       rc = register_keccak_algo        ( gate ); break;
+    case ALGO_KECCAKC:      rc = register_keccakc_algo       ( gate ); break;
+    case ALGO_LBRY:         rc = register_lbry_algo          ( gate ); break;
+    case ALGO_LYRA2H:       rc = register_lyra2h_algo        ( gate ); break;
+    case ALGO_LYRA2RE:      rc = register_lyra2re_algo       ( gate ); break;
+    case ALGO_LYRA2REV2:    rc = register_lyra2rev2_algo     ( gate ); break;
+    case ALGO_LYRA2REV3:    rc = register_lyra2rev3_algo     ( gate ); break;
+    case ALGO_LYRA2Z:       rc = register_lyra2z_algo        ( gate ); break;
+    case ALGO_LYRA2Z330:    rc = register_lyra2z330_algo     ( gate ); break;
+    case ALGO_M7M:          rc = register_m7m_algo           ( gate ); break;
+    case ALGO_MINOTAUR:     rc = register_minotaur_algo      ( gate ); break;
+    case ALGO_MYR_GR:       rc = register_myriad_algo        ( gate ); break;
+    case ALGO_NEOSCRYPT:    rc = register_neoscrypt_algo     ( gate ); break;
+    case ALGO_NIST5:        rc = register_nist5_algo         ( gate ); break;
+    case ALGO_PENTABLAKE:   rc = register_pentablake_algo    ( gate ); break;
+    case ALGO_PHI1612:      rc = register_phi1612_algo       ( gate ); break;
+    case ALGO_PHI2:         rc = register_phi2_algo          ( gate ); break;
+    case ALGO_POLYTIMOS:    rc = register_polytimos_algo     ( gate ); break;
+    case ALGO_POWER2B:      rc = register_power2b_algo       ( gate ); break;
+    case ALGO_QUARK:        rc = register_quark_algo         ( gate ); break;
+    case ALGO_QUBIT:        rc = register_qubit_algo         ( gate ); break;
+    case ALGO_SCRYPT:       rc = register_scrypt_algo        ( gate ); break;
+    case ALGO_SHA256D:      rc = register_sha256d_algo       ( gate ); break;
+    case ALGO_SHA256Q:      rc = register_sha256q_algo       ( gate ); break;
+    case ALGO_SHA256T:      rc = register_sha256t_algo       ( gate ); break;
+    case ALGO_SHA3D:        rc = register_sha3d_algo         ( gate ); break;
+    case ALGO_SHAVITE3:     rc = register_shavite_algo       ( gate ); break;
+    case ALGO_SKEIN:        rc = register_skein_algo         ( gate ); break;
+    case ALGO_SKEIN2:       rc = register_skein2_algo        ( gate ); break;
+    case ALGO_SKUNK:        rc = register_skunk_algo         ( gate ); break;
+    case ALGO_SONOA:        rc = register_sonoa_algo         ( gate ); break;
+    case ALGO_TIMETRAVEL:   rc = register_timetravel_algo    ( gate ); break;
+    case ALGO_TIMETRAVEL10: rc = register_timetravel10_algo  ( gate ); break;
+    case ALGO_TRIBUS:       rc = register_tribus_algo        ( gate ); break;
+    case ALGO_VANILLA:      rc = register_vanilla_algo       ( gate ); break;
+    case ALGO_VELTOR:       rc = register_veltor_algo        ( gate ); break;
+    case ALGO_VERTHASH:     rc = register_verthash_algo      ( gate ); break;
+    case ALGO_WHIRLPOOL:    rc = register_whirlpool_algo     ( gate ); break;
+    case ALGO_WHIRLPOOLX:   rc = register_whirlpoolx_algo    ( gate ); break;
+    case ALGO_X11:          rc = register_x11_algo           ( gate ); break;
+    case ALGO_X11EVO:       rc = register_x11evo_algo        ( gate ); break;
+    case ALGO_X11GOST:      rc = register_x11gost_algo       ( gate ); break;
+    case ALGO_X12:          rc = register_x12_algo           ( gate ); break;
+    case ALGO_X13:          rc = register_x13_algo           ( gate ); break;
+    case ALGO_X13BCD:       rc = register_x13bcd_algo        ( gate ); break;
+    case ALGO_X13SM3:       rc = register_x13sm3_algo        ( gate ); break;
+    case ALGO_X14:          rc = register_x14_algo           ( gate ); break;
+    case ALGO_X15:          rc = register_x15_algo           ( gate ); break;
+    case ALGO_X16R:         rc = register_x16r_algo          ( gate ); break;
+    case ALGO_X16RV2:       rc = register_x16rv2_algo        ( gate ); break;
+    case ALGO_X16RT:        rc = register_x16rt_algo         ( gate ); break;
+    case ALGO_X16RT_VEIL:   rc = register_x16rt_veil_algo    ( gate ); break;
+    case ALGO_X16S:         rc = register_x16s_algo          ( gate ); break;
+    case ALGO_X17:          rc = register_x17_algo           ( gate ); break;
+    case ALGO_X21S:         rc = register_x21s_algo          ( gate ); break;
+    case ALGO_X22I:         rc = register_x22i_algo          ( gate ); break;
+    case ALGO_X25X:         rc = register_x25x_algo          ( gate ); break;
+    case ALGO_XEVAN:        rc = register_xevan_algo         ( gate ); break;
+    case ALGO_YESCRYPT:     rc = register_yescrypt_05_algo   ( gate ); break;
+//    case ALGO_YESCRYPT:      register_yescrypt_algo      ( gate ); break;
+    case ALGO_YESCRYPTR8:   rc = register_yescryptr8_05_algo ( gate ); break;
+//    case ALGO_YESCRYPTR8:    register_yescryptr8_algo    ( gate ); break;
+    case ALGO_YESCRYPTR8G:  rc = register_yescryptr8g_algo   ( gate ); break;
+    case ALGO_YESCRYPTR16:  rc = register_yescryptr16_05_algo( gate ); break;
+//    case ALGO_YESCRYPTR16:   register_yescryptr16_algo   ( gate ); break;
+    case ALGO_YESCRYPTR32:  rc = register_yescryptr32_05_algo( gate ); break;
+//    case ALGO_YESCRYPTR32:   register_yescryptr32_algo   ( gate ); break;
+    case ALGO_YESPOWER:     rc = register_yespower_algo      ( gate ); break;
+    case ALGO_YESPOWERR16:  rc = register_yespowerr16_algo   ( gate ); break;
+    case ALGO_YESPOWER_B2B: rc = register_yespower_b2b_algo  ( gate ); break;
+    case ALGO_ZR5:          rc = register_zr5_algo           ( gate ); break;
   default:
-      applog(LOG_ERR,"FAIL: algo_gate registration failed, unknown algo %s.\n", algo_names[opt_algo] );
+      applog(LOG_ERR,"BUG: unregistered algorithm %s.\n", algo_names[opt_algo] );
      return false;
  } // switch

- // ensure required functions were defined.
-  if (  gate->scanhash == (void*)&null_scanhash )
+  if ( !rc )
  {
-    applog(LOG_ERR, "FAIL: Required algo_gate functions undefined\n");
+    applog(LOG_ERR, "FAIL: %s algorithm failed to initialize\n", algo_names[opt_algo] );
    return false;
  }
  return true;
@@ -415,7 +419,6 @@ void exec_hash_function( int algo, void *output, const void *pdata )
 const char* const algo_alias_map[][2] =
 {
 //   alias                proper
-  { "argon2d-crds",      "argon2d250"     },
  { "argon2d-dyn",       "argon2d500"     },
  { "argon2d-uis",       "argon2d4096"    },
  { "bcd",               "x13bcd"         },
@@ -430,7 +433,6 @@ const char* const algo_alias_map[][2] =
  { "flax",              "c11"            },
  { "hsr",               "x13sm3"         },
  { "jackpot",           "jha"            },
-  { "jane",              "scryptjane"     }, 
  { "lyra2",             "lyra2re"        },
  { "lyra2v2",           "lyra2rev2"      },
  { "lyra2v3",           "lyra2rev3"      },
--- a/algo-gate-api.h
+++ b/algo-gate-api.h
@@ -90,10 +90,11 @@ typedef  uint32_t set_t;
 #define AES_OPT          2  
 #define SSE42_OPT        4
 #define AVX_OPT          8   // Sandybridge
-#define AVX2_OPT      0x10   // Haswell
-#define SHA_OPT       0x20   // sha256 (Ryzen, Ice Lake)
-#define AVX512_OPT    0x40   // AVX512- F, VL, DQ, BW (Skylake-X)
-#define VAES_OPT      0x80   // VAES (Ice Lake)
+#define AVX2_OPT      0x10   // Haswell, Zen1
+#define SHA_OPT       0x20   // Zen1, Icelake (sha256)
+#define AVX512_OPT    0x40   // Skylake-X (AVX512[F,VL,DQ,BW])
+#define VAES_OPT      0x80   // Icelake (VAES & AVX512)
+#define VAES256_OPT   0x100  // Zen3 (VAES without AVX512)


 // return set containing all elements from sets a & b
@@ -111,17 +112,17 @@ inline bool set_excl ( set_t a, set_t b ) { return (a & b) == 0; }
 typedef struct
 {
 // Mandatory functions, one of these is mandatory. If a generic scanhash
-// is used a custom hash function must be registered, with a custom scanhash
-// the custom hash function can be called directly and doesn't need to be
-// registered in the gate. 
+// is used a custom target hash function must be registered, with a custom
+// scanhash the target hash function can be called directly and doesn't need
+// to be registered with the gate. 
 int ( *scanhash ) ( struct work*, uint32_t, uint64_t*, struct thr_info* );

 int ( *hash )     ( void*, const void*, int );

 //optional, safe to use default in most cases

-// Allocate thread local buffers and other initialization specific to miner
-// threads.
+// Called once by each miner thread to allocate thread local buffers and
+// other initialization specific to miner threads.
 bool ( *miner_thread_init )     ( int );

 // Get thread local copy of blockheader with unique nonce.
@@ -149,7 +150,7 @@ void ( *build_stratum_request ) ( char*, struct work*, struct stratum_ctx* );

 char* ( *malloc_txs_request )   ( struct work* );

-// Big or little
+// Big endian or little endian
 void ( *set_work_data_endian )  ( struct work* );

 double ( *calc_network_diff )   ( struct work* );
@@ -161,7 +162,7 @@ bool ( *ready_to_mine )         ( struct work*, struct stratum_ctx*, int );
 bool ( *do_this_thread )        ( int );

 // After do_this_thread
-void ( *resync_threads )        ( struct work* );
+void ( *resync_threads )        ( int, struct work* );

 // No longer needed
 json_t* (*longpoll_rpc_call)      ( CURL*, int*, char* );
@@ -259,7 +260,7 @@ int scanhash_8way_64in_32out( struct work *work, uint32_t max_nonce,
 #endif

 // displays warning
-int null_hash    ();
+int null_hash();

 // optional safe targets, default listed first unless noted.

@@ -280,7 +281,7 @@ void std_be_build_stratum_request( char *req, struct work *work );

 char* std_malloc_txs_request( struct work *work );

-// Default is do_nothing (assumed LE)
+// Default is do_nothing, little endian is assumed
 void set_work_data_big_endian( struct work *work );

 double std_calc_network_diff( struct work *work );
--- a/algo/blake/blake2b-hash-4way.h
+++ b/algo/blake/blake2b-hash-4way.h
@@ -17,7 +17,7 @@

 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

-ALIGN(128) typedef struct {
+typedef struct ALIGN( 64 ) {
   __m512i b[16]; // input buffer
   __m512i h[8];  // chained state
   uint64_t t[2];  // total number of bytes
@@ -35,7 +35,7 @@ void blake2b_8way_final( blake2b_8way_ctx *ctx, void *out );
 #if defined(__AVX2__)

 // state context
-ALIGN(128) typedef struct {
+typedef struct ALIGN( 64 ) {
 	__m256i b[16]; // input buffer
 	__m256i h[8];  // chained state
 	uint64_t t[2];  // total number of bytes
--- a/algo/blake/blake2s-hash-4way.h
+++ b/algo/blake/blake2s-hash-4way.h
@@ -60,7 +60,7 @@ typedef struct __blake2s_nway_param
 } blake2s_nway_param;
 #pragma pack(pop)

-ALIGN( 64 ) typedef struct __blake2s_4way_state
+typedef struct ALIGN( 64 ) __blake2s_4way_state
 {
   __m128i h[8];
   uint8_t  buf[ BLAKE2S_BLOCKBYTES * 4 ];
@@ -80,7 +80,7 @@ int blake2s_4way_full_blocks( blake2s_4way_state *S, void *out,

 #if defined(__AVX2__)

-ALIGN( 64 ) typedef struct __blake2s_8way_state
+typedef struct ALIGN( 64 ) __blake2s_8way_state
 {
   __m256i h[8];
   uint8_t  buf[ BLAKE2S_BLOCKBYTES * 8 ];
@@ -101,7 +101,7 @@ int blake2s_8way_full_blocks( blake2s_8way_state *S, void *out,

 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

-ALIGN( 128 ) typedef struct __blake2s_16way_state
+typedef struct ALIGN( 64 ) __blake2s_16way_state
 {
   __m512i h[8];
   uint8_t  buf[ BLAKE2S_BLOCKBYTES * 16 ];
--- a/algo/blake/sph-blake2s.c
+++ b/algo/blake/sph-blake2s.c
@@ -323,7 +323,7 @@ int blake2s_final( blake2s_state *S, uint8_t *out, uint8_t outlen )

 int blake2s( uint8_t *out, const void *in, const void *key, const uint8_t outlen, const uint64_t inlen, uint8_t keylen )
 {
-	blake2s_state S[1];
+	blake2s_state S;

 	/* Verify parameters */
 	if ( NULL == in ) return -1;
@@ -334,15 +334,15 @@ int blake2s( uint8_t *out, const void *in, const void *key, const uint8_t outlen

 	if( keylen > 0 )
 	{
-		if( blake2s_init_key( S, outlen, key, keylen ) < 0 ) return -1;
+		if( blake2s_init_key( &S, outlen, key, keylen ) < 0 ) return -1;
 	}
 	else
 	{
-		if( blake2s_init( S, outlen ) < 0 ) return -1;
+		if( blake2s_init( &S, outlen ) < 0 ) return -1;
 	}

-	blake2s_update( S, ( uint8_t * )in, inlen );
-	blake2s_final( S, out, outlen );
+	blake2s_update( &S, ( uint8_t * )in, inlen );
+	blake2s_final( &S, out, outlen );
 	return 0;
 }

--- a/algo/blake/sph-blake2s.h
+++ b/algo/blake/sph-blake2s.h
@@ -116,7 +116,7 @@ extern "C" {
 		uint8_t  personal[BLAKE2S_PERSONALBYTES];  // 32
 	} blake2s_param;

-	ALIGN( 64 ) typedef struct __blake2s_state
+	typedef struct ALIGN( 64 ) __blake2s_state
 	{
 		uint32_t h[8];
 		uint32_t t[2];
--- a/algo/blake/sph_blake2b.h
+++ b/algo/blake/sph_blake2b.h
@@ -18,7 +18,7 @@
 #endif

 // state context
-ALIGN(64) typedef struct {
+typedef ALIGN(64) struct {
 	uint8_t b[128]; // input buffer
 	uint64_t h[8];  // chained state
 	uint64_t t[2];  // total number of bytes
--- a/algo/echo/aes_ni/hash.c
+++ b/algo/echo/aes_ni/hash.c
@@ -55,8 +55,8 @@ MYALIGN const unsigned int	mul2ipt[]		= {0x728efc00, 0x6894e61a, 0x3fc3b14d, 0x2

 #define ECHO_SUBBYTES(state, i, j) \
 	state[i][j] = _mm_aesenc_si128(state[i][j], k1);\
-	state[i][j] = _mm_aesenc_si128(state[i][j], M128(zero));\
-	k1 = _mm_add_epi32(k1, M128(const1))
+   k1 = _mm_add_epi32(k1, M128(const1));\
+	state[i][j] = _mm_aesenc_si128(state[i][j], M128(zero))

 #define ECHO_MIXBYTES(state1, state2, j, t1, t2, s2) \
 	s2 = _mm_add_epi8(state1[0][j], state1[0][j]);\
--- a/algo/echo/echo-hash-4way.c
+++ b/algo/echo/echo-hash-4way.c
@@ -1,5 +1,4 @@
-//#if 0
-#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__VAES__)

 #include "simd-utils.h"
 #include "echo-hash-4way.h"
@@ -11,18 +10,20 @@ static const unsigned int mul2ipt[] __attribute__ ((aligned (64))) =
   0xfd5ba600, 0x2a8c71d7, 0x1eb845e3, 0xc96f9234
 };
 */
-// do these need to be reversed?

-#define mul2mask \
-   _mm512_set4_epi32( 0, 0, 0, 0x00001b00 ) 
-//   _mm512_set4_epi32( 0x00001b00, 0, 0, 0 )  
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

-#define lsbmask    m512_const1_32( 0x01010101 ) 
+
+//#define mul2mask    m512_const2_64( 0, 0x00001b00 )
+//_mm512_set4_epi32( 0, 0, 0, 0x00001b00 ) 
+//_mm512_set4_epi32( 0x00001b00, 0, 0, 0 )  
+
+//#define lsbmask    m512_const1_32( 0x01010101 ) 

 #define ECHO_SUBBYTES( state, i, j ) \
 	state[i][j] = _mm512_aesenc_epi128( state[i][j], k1 ); \
-	state[i][j] = _mm512_aesenc_epi128( state[i][j], m512_zero ); \
-	k1 = _mm512_add_epi32( k1, m512_one_128 );
+   k1 = _mm512_add_epi32( k1, one ); \
+	state[i][j] = _mm512_aesenc_epi128( state[i][j], m512_zero );

 #define ECHO_MIXBYTES( state1, state2, j, t1, t2, s2 ) do \
 { \
@@ -30,87 +31,87 @@ static const unsigned int mul2ipt[] __attribute__ ((aligned (64))) =
   const int j2 = ( (j)+2 ) & 3; \
   const int j3 = ( (j)+3 ) & 3; \
   s2 = _mm512_add_epi8( state1[ 0 ] [j ], state1[ 0 ][ j ] ); \
-	t1 = _mm512_srli_epi16( state1[ 0 ][ j ], 7 ); \
-	t1 = _mm512_and_si512( t1, lsbmask );\
-	t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
-	s2 = _mm512_xor_si512( s2, t2 ); \
-	state2[ 0 ] [j ] = s2; \
-	state2[ 1 ] [j ] = state1[ 0 ][ j ]; \
-	state2[ 2 ] [j ] = state1[ 0 ][ j ]; \
-	state2[ 3 ] [j ] = _mm512_xor_si512( s2, state1[ 0 ][ j ] );\
-	s2 = _mm512_add_epi8( state1[ 1 ][ j1 ], state1[ 1 ][ j1 ] ); \
-	t1 = _mm512_srli_epi16( state1[ 1 ][ j1 ], 7 ); \
-	t1 = _mm512_and_si512( t1, lsbmask ); \
-	t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
-	s2 = _mm512_xor_si512( s2, t2 );\
-	state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], \
-                            _mm512_xor_si512( s2, state1[ 1 ][ j1 ] ) ); \
-	state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], s2 ); \
-	state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], state1[ 1 ][ j1 ] ); \
-	state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3 ][ j ], state1[ 1 ][ j1 ] ); \
-	s2 = _mm512_add_epi8( state1[ 2 ][ j2 ], state1[ 2 ][ j2 ] ); \
-	t1 = _mm512_srli_epi16( state1[ 2 ][ j2 ], 7 ); \
-	t1 = _mm512_and_si512( t1, lsbmask ); \
-	t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
-	s2 = _mm512_xor_si512( s2, t2 ); \
-	state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], state1[ 2 ][ j2 ] ); \
-	state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], \
+   t1 = _mm512_srli_epi16( state1[ 0 ][ j ], 7 ); \
+   t1 = _mm512_and_si512( t1, lsbmask );\
+   t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
+   s2 = _mm512_xor_si512( s2, t2 ); \
+   state2[ 0 ] [j ] = s2; \
+   state2[ 1 ] [j ] = state1[ 0 ][ j ]; \
+   state2[ 2 ] [j ] = state1[ 0 ][ j ]; \
+   state2[ 3 ] [j ] = _mm512_xor_si512( s2, state1[ 0 ][ j ] );\
+   s2 = _mm512_add_epi8( state1[ 1 ][ j1 ], state1[ 1 ][ j1 ] ); \
+   t1 = _mm512_srli_epi16( state1[ 1 ][ j1 ], 7 ); \
+   t1 = _mm512_and_si512( t1, lsbmask ); \
+   t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
+   s2 = _mm512_xor_si512( s2, t2 );\
+   state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], \
+                              _mm512_xor_si512( s2, state1[ 1 ][ j1 ] ) ); \
+   state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], s2 ); \
+   state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], state1[ 1 ][ j1 ] ); \
+   state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3 ][ j ], state1[ 1 ][ j1 ] ); \
+   s2 = _mm512_add_epi8( state1[ 2 ][ j2 ], state1[ 2 ][ j2 ] ); \
+   t1 = _mm512_srli_epi16( state1[ 2 ][ j2 ], 7 ); \
+   t1 = _mm512_and_si512( t1, lsbmask ); \
+   t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
+   s2 = _mm512_xor_si512( s2, t2 ); \
+   state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], state1[ 2 ][ j2 ] ); \
+   state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], \
                            _mm512_xor_si512( s2, state1[ 2 ][ j2 ] ) ); \
-	state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], s2 ); \
-	state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3][ j ], state1[ 2 ][ j2 ] ); \
-	s2 = _mm512_add_epi8( state1[ 3 ][ j3 ], state1[ 3 ][ j3 ] ); \
-	t1 = _mm512_srli_epi16( state1[ 3 ][ j3 ], 7 ); \
-	t1 = _mm512_and_si512( t1, lsbmask ); \
-	t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
-	s2 = _mm512_xor_si512( s2, t2 ); \
-	state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], state1[ 3 ][ j3 ] ); \
-	state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], state1[ 3 ][ j3 ] ); \
-	state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], \
+   state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], s2 ); \
+   state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3][ j ], state1[ 2 ][ j2 ] ); \
+   s2 = _mm512_add_epi8( state1[ 3 ][ j3 ], state1[ 3 ][ j3 ] ); \
+   t1 = _mm512_srli_epi16( state1[ 3 ][ j3 ], 7 ); \
+   t1 = _mm512_and_si512( t1, lsbmask ); \
+   t2 = _mm512_shuffle_epi8( mul2mask, t1 ); \
+   s2 = _mm512_xor_si512( s2, t2 ); \
+   state2[ 0 ][ j ] = _mm512_xor_si512( state2[ 0 ][ j ], state1[ 3 ][ j3 ] ); \
+   state2[ 1 ][ j ] = _mm512_xor_si512( state2[ 1 ][ j ], state1[ 3 ][ j3 ] ); \
+   state2[ 2 ][ j ] = _mm512_xor_si512( state2[ 2 ][ j ], \
                            _mm512_xor_si512( s2, state1[ 3 ][ j3] ) ); \
-	state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3 ][ j ], s2 ); \
+   state2[ 3 ][ j ] = _mm512_xor_si512( state2[ 3 ][ j ], s2 ); \
 } while(0)

 #define ECHO_ROUND_UNROLL2 \
-	ECHO_SUBBYTES(_state, 0, 0);\
+   ECHO_SUBBYTES(_state, 0, 0);\
   ECHO_SUBBYTES(_state, 1, 0);\
-	ECHO_SUBBYTES(_state, 2, 0);\
-	ECHO_SUBBYTES(_state, 3, 0);\
-	ECHO_SUBBYTES(_state, 0, 1);\
-	ECHO_SUBBYTES(_state, 1, 1);\
-	ECHO_SUBBYTES(_state, 2, 1);\
-	ECHO_SUBBYTES(_state, 3, 1);\
-	ECHO_SUBBYTES(_state, 0, 2);\
-	ECHO_SUBBYTES(_state, 1, 2);\
-	ECHO_SUBBYTES(_state, 2, 2);\
-	ECHO_SUBBYTES(_state, 3, 2);\
-	ECHO_SUBBYTES(_state, 0, 3);\
-	ECHO_SUBBYTES(_state, 1, 3);\
-	ECHO_SUBBYTES(_state, 2, 3);\
-	ECHO_SUBBYTES(_state, 3, 3);\
-	ECHO_MIXBYTES(_state, _state2, 0, t1, t2, s2);\
-	ECHO_MIXBYTES(_state, _state2, 1, t1, t2, s2);\
-	ECHO_MIXBYTES(_state, _state2, 2, t1, t2, s2);\
-	ECHO_MIXBYTES(_state, _state2, 3, t1, t2, s2);\
-	ECHO_SUBBYTES(_state2, 0, 0);\
-	ECHO_SUBBYTES(_state2, 1, 0);\
-	ECHO_SUBBYTES(_state2, 2, 0);\
-	ECHO_SUBBYTES(_state2, 3, 0);\
-	ECHO_SUBBYTES(_state2, 0, 1);\
-	ECHO_SUBBYTES(_state2, 1, 1);\
-	ECHO_SUBBYTES(_state2, 2, 1);\
-	ECHO_SUBBYTES(_state2, 3, 1);\
-	ECHO_SUBBYTES(_state2, 0, 2);\
-	ECHO_SUBBYTES(_state2, 1, 2);\
-	ECHO_SUBBYTES(_state2, 2, 2);\
-	ECHO_SUBBYTES(_state2, 3, 2);\
-	ECHO_SUBBYTES(_state2, 0, 3);\
-	ECHO_SUBBYTES(_state2, 1, 3);\
-	ECHO_SUBBYTES(_state2, 2, 3);\
-	ECHO_SUBBYTES(_state2, 3, 3);\
-	ECHO_MIXBYTES(_state2, _state, 0, t1, t2, s2);\
-	ECHO_MIXBYTES(_state2, _state, 1, t1, t2, s2);\
-	ECHO_MIXBYTES(_state2, _state, 2, t1, t2, s2);\
-	ECHO_MIXBYTES(_state2, _state, 3, t1, t2, s2)
+   ECHO_SUBBYTES(_state, 2, 0);\
+   ECHO_SUBBYTES(_state, 3, 0);\
+   ECHO_SUBBYTES(_state, 0, 1);\
+   ECHO_SUBBYTES(_state, 1, 1);\
+   ECHO_SUBBYTES(_state, 2, 1);\
+   ECHO_SUBBYTES(_state, 3, 1);\
+   ECHO_SUBBYTES(_state, 0, 2);\
+   ECHO_SUBBYTES(_state, 1, 2);\
+   ECHO_SUBBYTES(_state, 2, 2);\
+   ECHO_SUBBYTES(_state, 3, 2);\
+   ECHO_SUBBYTES(_state, 0, 3);\
+   ECHO_SUBBYTES(_state, 1, 3);\
+   ECHO_SUBBYTES(_state, 2, 3);\
+   ECHO_SUBBYTES(_state, 3, 3);\
+   ECHO_MIXBYTES(_state, _state2, 0, t1, t2, s2);\
+   ECHO_MIXBYTES(_state, _state2, 1, t1, t2, s2);\
+   ECHO_MIXBYTES(_state, _state2, 2, t1, t2, s2);\
+   ECHO_MIXBYTES(_state, _state2, 3, t1, t2, s2);\
+   ECHO_SUBBYTES(_state2, 0, 0);\
+   ECHO_SUBBYTES(_state2, 1, 0);\
+   ECHO_SUBBYTES(_state2, 2, 0);\
+   ECHO_SUBBYTES(_state2, 3, 0);\
+   ECHO_SUBBYTES(_state2, 0, 1);\
+   ECHO_SUBBYTES(_state2, 1, 1);\
+   ECHO_SUBBYTES(_state2, 2, 1);\
+   ECHO_SUBBYTES(_state2, 3, 1);\
+   ECHO_SUBBYTES(_state2, 0, 2);\
+   ECHO_SUBBYTES(_state2, 1, 2);\
+   ECHO_SUBBYTES(_state2, 2, 2);\
+   ECHO_SUBBYTES(_state2, 3, 2);\
+   ECHO_SUBBYTES(_state2, 0, 3);\
+   ECHO_SUBBYTES(_state2, 1, 3);\
+   ECHO_SUBBYTES(_state2, 2, 3);\
+   ECHO_SUBBYTES(_state2, 3, 3);\
+   ECHO_MIXBYTES(_state2, _state, 0, t1, t2, s2);\
+   ECHO_MIXBYTES(_state2, _state, 1, t1, t2, s2);\
+   ECHO_MIXBYTES(_state2, _state, 2, t1, t2, s2);\
+   ECHO_MIXBYTES(_state2, _state, 3, t1, t2, s2)

 #define SAVESTATE(dst, src)\
 	dst[0][0] = src[0][0];\
@@ -137,6 +138,9 @@ void echo_4way_compress( echo_4way_context *ctx, const __m512i *pmsg,
  unsigned int r, b, i, j;
  __m512i t1, t2, s2, k1;
  __m512i _state[4][4], _state2[4][4], _statebackup[4][4]; 
+  __m512i one = m512_one_128;
+  __m512i mul2mask = m512_const2_64( 0, 0x00001b00 );
+  __m512i lsbmask  = m512_const1_32( 0x01010101 ); 

  _state[ 0 ][ 0 ] = ctx->state[ 0 ][ 0 ];
  _state[ 0 ][ 1 ] = ctx->state[ 0 ][ 1 ];
@@ -224,43 +228,43 @@ void echo_4way_compress( echo_4way_context *ctx, const __m512i *pmsg,

 int echo_4way_init( echo_4way_context *ctx, int nHashSize )
 {
-	int i, j;
+   int i, j;

   ctx->k = m512_zero; 
-	ctx->processed_bits = 0;
-	ctx->uBufferBytes = 0;
+   ctx->processed_bits = 0;
+   ctx->uBufferBytes = 0;

-	switch( nHashSize )
-	{
-		case 256:
-			ctx->uHashSize = 256;
-			ctx->uBlockLength = 192;
-			ctx->uRounds = 8;
-			ctx->hashsize = _mm512_set4_epi32( 0, 0, 0, 0x100 );
-			ctx->const1536 = _mm512_set4_epi32( 0, 0, 0, 0x600 );
-			break;
+   switch( nHashSize )
+   {
+	case 256:
+		ctx->uHashSize = 256;
+		ctx->uBlockLength = 192;
+		ctx->uRounds = 8;
+		ctx->hashsize = m512_const2_64( 0, 0x100 );
+		ctx->const1536 = m512_const2_64( 0, 0x600 );
+		break;

-		case 512:
-			ctx->uHashSize = 512;
-			ctx->uBlockLength = 128;
-			ctx->uRounds = 10;
-			ctx->hashsize = _mm512_set4_epi32( 0, 0, 0, 0x200 );
-			ctx->const1536 = _mm512_set4_epi32( 0, 0, 0, 0x400);
-			break;
+	case 512:
+		ctx->uHashSize = 512;
+		ctx->uBlockLength = 128;
+		ctx->uRounds = 10;
+		ctx->hashsize = m512_const2_64( 0, 0x200 );
+		ctx->const1536 = m512_const2_64( 0, 0x400);
+		break;

-		default:
-			return 1;
-	}
+	default:
+        	return 1;
+   }

-	for( i = 0; i < 4; i++ )
-		for( j = 0; j < nHashSize / 256; j++ )
-			ctx->state[ i ][ j ] = ctx->hashsize;
+   for( i = 0; i < 4; i++ )
+	for( j = 0; j < nHashSize / 256; j++ )
+		ctx->state[ i ][ j ] = ctx->hashsize;

-	for( i = 0; i < 4; i++ )
-		for( j = nHashSize / 256; j < 4; j++ )
-			ctx->state[ i ][ j ] = m512_zero;
+   for( i = 0; i < 4; i++ )
+	for( j = nHashSize / 256; j < 4; j++ )
+		ctx->state[ i ][ j ] = m512_zero;

-	return 0;
+   return 0;
 }

 int echo_4way_update_close( echo_4way_context *state, void *hashval,
@@ -285,17 +289,13 @@ int echo_4way_update_close( echo_4way_context *state, void *hashval,
      vlen = databitlen / 128;  // * 4 lanes / 128 bits per lane
      memcpy_512( state->buffer, data, vlen );
      state->processed_bits += (unsigned int)( databitlen );
-      remainingbits = _mm512_set4_epi32( 0, 0, 0, databitlen );
-
+      remainingbits = m512_const2_64( 0, (uint64_t)databitlen );
   }

-   state->buffer[ vlen ] = _mm512_set4_epi32( 0, 0, 0, 0x80 );
+   state->buffer[ vlen ] = m512_const2_64( 0, 0x80 );
   memset_zero_512( state->buffer + vlen + 1, vblen - vlen - 2 );
-   state->buffer[ vblen-2 ] =
-                _mm512_set4_epi32( (uint32_t)state->uHashSize << 16, 0, 0, 0 );
-   state->buffer[ vblen-1 ] =
-                   _mm512_set4_epi64( 0, state->processed_bits,
-                                      0, state->processed_bits );  
+   state->buffer[ vblen-2 ] = m512_const2_64( (uint64_t)state->uHashSize << 48, 0 );
+   state->buffer[ vblen-1 ] = m512_const2_64( 0, state->processed_bits);

   state->k = _mm512_add_epi64( state->k, remainingbits );
   state->k = _mm512_sub_epi64( state->k, state->const1536 );
@@ -328,16 +328,16 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
         ctx->uHashSize = 256;
         ctx->uBlockLength = 192;
         ctx->uRounds = 8;
-         ctx->hashsize = _mm512_set4_epi32( 0, 0, 0, 0x100 );
-         ctx->const1536 = _mm512_set4_epi32( 0, 0, 0, 0x600 );
+         ctx->hashsize = m512_const2_64( 0, 0x100 );
+         ctx->const1536 = m512_const2_64( 0, 0x600 );
         break;

      case 512:
         ctx->uHashSize = 512;
         ctx->uBlockLength = 128;
         ctx->uRounds = 10;
-         ctx->hashsize = _mm512_set4_epi32( 0, 0, 0, 0x200 );
-         ctx->const1536 = _mm512_set4_epi32( 0, 0, 0, 0x400);
+         ctx->hashsize = m512_const2_64( 0, 0x200 );
+         ctx->const1536 = m512_const2_64( 0, 0x400 );
         break;

      default:
@@ -372,17 +372,14 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
      vlen = databitlen / 128;  // * 4 lanes / 128 bits per lane
      memcpy_512( ctx->buffer, data, vlen );
      ctx->processed_bits += (unsigned int)( databitlen );
-      remainingbits = _mm512_set4_epi32( 0, 0, 0, databitlen );
-
+      remainingbits = m512_const2_64( 0, databitlen );
   }

-   ctx->buffer[ vlen ] = _mm512_set4_epi32( 0, 0, 0, 0x80 );
+   ctx->buffer[ vlen ] = m512_const2_64( 0, 0x80 );
   memset_zero_512( ctx->buffer + vlen + 1, vblen - vlen - 2 );
   ctx->buffer[ vblen-2 ] =
-                _mm512_set4_epi32( (uint32_t)ctx->uHashSize << 16, 0, 0, 0 );
-   ctx->buffer[ vblen-1 ] =
-                   _mm512_set4_epi64( 0, ctx->processed_bits,
-                                      0, ctx->processed_bits );
+                     m512_const2_64( (uint64_t)ctx->uHashSize << 48, 0 );
+   ctx->buffer[ vblen-1 ] = m512_const2_64( 0, ctx->processed_bits);

   ctx->k = _mm512_add_epi64( ctx->k, remainingbits );
   ctx->k = _mm512_sub_epi64( ctx->k, ctx->const1536 );
@@ -400,5 +397,380 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
   return 0;
 }

+#endif  // AVX512

-#endif
+// AVX2 + VAES
+
+#define mul2mask_2way   m256_const2_64( 0, 0x0000000000001b00 ) 
+
+#define lsbmask_2way    m256_const1_32( 0x01010101 ) 
+
+#define ECHO_SUBBYTES_2WAY( state, i, j ) \
+        state[i][j] = _mm256_aesenc_epi128( state[i][j], k1 ); \
+        k1 = _mm256_add_epi32( k1, m256_one_128 ); \
+        state[i][j] = _mm256_aesenc_epi128( state[i][j], m256_zero ); \
+
+#define ECHO_MIXBYTES_2WAY( state1, state2, j, t1, t2, s2 ) do \
+{ \
+   const int j1 = ( (j)+1 ) & 3; \
+   const int j2 = ( (j)+2 ) & 3; \
+   const int j3 = ( (j)+3 ) & 3; \
+   s2 = _mm256_add_epi8( state1[ 0 ] [j ], state1[ 0 ][ j ] ); \
+   t1 = _mm256_srli_epi16( state1[ 0 ][ j ], 7 ); \
+   t1 = _mm256_and_si256( t1, lsbmask_2way );\
+   t2 = _mm256_shuffle_epi8( mul2mask_2way, t1 ); \
+   s2 = _mm256_xor_si256( s2, t2 ); \
+   state2[ 0 ] [j ] = s2; \
+   state2[ 1 ] [j ] = state1[ 0 ][ j ]; \
+   state2[ 2 ] [j ] = state1[ 0 ][ j ]; \
+   state2[ 3 ] [j ] = _mm256_xor_si256( s2, state1[ 0 ][ j ] );\
+   s2 = _mm256_add_epi8( state1[ 1 ][ j1 ], state1[ 1 ][ j1 ] ); \
+   t1 = _mm256_srli_epi16( state1[ 1 ][ j1 ], 7 ); \
+   t1 = _mm256_and_si256( t1, lsbmask_2way ); \
+   t2 = _mm256_shuffle_epi8( mul2mask_2way, t1 ); \
+   s2 = _mm256_xor_si256( s2, t2 );\
+   state2[ 0 ][ j ] = _mm256_xor_si256( state2[ 0 ][ j ], \
+                              _mm256_xor_si256( s2, state1[ 1 ][ j1 ] ) ); \
+   state2[ 1 ][ j ] = _mm256_xor_si256( state2[ 1 ][ j ], s2 ); \
+   state2[ 2 ][ j ] = _mm256_xor_si256( state2[ 2 ][ j ], state1[ 1 ][ j1 ] ); \
+   state2[ 3 ][ j ] = _mm256_xor_si256( state2[ 3 ][ j ], state1[ 1 ][ j1 ] ); \
+   s2 = _mm256_add_epi8( state1[ 2 ][ j2 ], state1[ 2 ][ j2 ] ); \
+   t1 = _mm256_srli_epi16( state1[ 2 ][ j2 ], 7 ); \
+   t1 = _mm256_and_si256( t1, lsbmask_2way ); \
+   t2 = _mm256_shuffle_epi8( mul2mask_2way, t1 ); \
+   s2 = _mm256_xor_si256( s2, t2 ); \
+   state2[ 0 ][ j ] = _mm256_xor_si256( state2[ 0 ][ j ], state1[ 2 ][ j2 ] ); \
+   state2[ 1 ][ j ] = _mm256_xor_si256( state2[ 1 ][ j ], \
+                            _mm256_xor_si256( s2, state1[ 2 ][ j2 ] ) ); \
+   state2[ 2 ][ j ] = _mm256_xor_si256( state2[ 2 ][ j ], s2 ); \
+   state2[ 3 ][ j ] = _mm256_xor_si256( state2[ 3][ j ], state1[ 2 ][ j2 ] ); \
+   s2 = _mm256_add_epi8( state1[ 3 ][ j3 ], state1[ 3 ][ j3 ] ); \
+   t1 = _mm256_srli_epi16( state1[ 3 ][ j3 ], 7 ); \
+   t1 = _mm256_and_si256( t1, lsbmask_2way ); \
+   t2 = _mm256_shuffle_epi8( mul2mask_2way, t1 ); \
+   s2 = _mm256_xor_si256( s2, t2 ); \
+   state2[ 0 ][ j ] = _mm256_xor_si256( state2[ 0 ][ j ], state1[ 3 ][ j3 ] ); \
+   state2[ 1 ][ j ] = _mm256_xor_si256( state2[ 1 ][ j ], state1[ 3 ][ j3 ] ); \
+   state2[ 2 ][ j ] = _mm256_xor_si256( state2[ 2 ][ j ], \
+                            _mm256_xor_si256( s2, state1[ 3 ][ j3] ) ); \
+   state2[ 3 ][ j ] = _mm256_xor_si256( state2[ 3 ][ j ], s2 ); \
+} while(0)
+
+#define ECHO_ROUND_UNROLL2_2WAY \
+   ECHO_SUBBYTES_2WAY(_state, 0, 0);\
+   ECHO_SUBBYTES_2WAY(_state, 1, 0);\
+   ECHO_SUBBYTES_2WAY(_state, 2, 0);\
+   ECHO_SUBBYTES_2WAY(_state, 3, 0);\
+   ECHO_SUBBYTES_2WAY(_state, 0, 1);\
+   ECHO_SUBBYTES_2WAY(_state, 1, 1);\
+   ECHO_SUBBYTES_2WAY(_state, 2, 1);\
+   ECHO_SUBBYTES_2WAY(_state, 3, 1);\
+   ECHO_SUBBYTES_2WAY(_state, 0, 2);\
+   ECHO_SUBBYTES_2WAY(_state, 1, 2);\
+   ECHO_SUBBYTES_2WAY(_state, 2, 2);\
+   ECHO_SUBBYTES_2WAY(_state, 3, 2);\
+   ECHO_SUBBYTES_2WAY(_state, 0, 3);\
+   ECHO_SUBBYTES_2WAY(_state, 1, 3);\
+   ECHO_SUBBYTES_2WAY(_state, 2, 3);\
+   ECHO_SUBBYTES_2WAY(_state, 3, 3);\
+   ECHO_MIXBYTES_2WAY(_state, _state2, 0, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state, _state2, 1, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state, _state2, 2, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state, _state2, 3, t1, t2, s2);\
+   ECHO_SUBBYTES_2WAY(_state2, 0, 0);\
+   ECHO_SUBBYTES_2WAY(_state2, 1, 0);\
+   ECHO_SUBBYTES_2WAY(_state2, 2, 0);\
+   ECHO_SUBBYTES_2WAY(_state2, 3, 0);\
+   ECHO_SUBBYTES_2WAY(_state2, 0, 1);\
+   ECHO_SUBBYTES_2WAY(_state2, 1, 1);\
+   ECHO_SUBBYTES_2WAY(_state2, 2, 1);\
+   ECHO_SUBBYTES_2WAY(_state2, 3, 1);\
+   ECHO_SUBBYTES_2WAY(_state2, 0, 2);\
+   ECHO_SUBBYTES_2WAY(_state2, 1, 2);\
+   ECHO_SUBBYTES_2WAY(_state2, 2, 2);\
+   ECHO_SUBBYTES_2WAY(_state2, 3, 2);\
+   ECHO_SUBBYTES_2WAY(_state2, 0, 3);\
+   ECHO_SUBBYTES_2WAY(_state2, 1, 3);\
+   ECHO_SUBBYTES_2WAY(_state2, 2, 3);\
+   ECHO_SUBBYTES_2WAY(_state2, 3, 3);\
+   ECHO_MIXBYTES_2WAY(_state2, _state, 0, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state2, _state, 1, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state2, _state, 2, t1, t2, s2);\
+   ECHO_MIXBYTES_2WAY(_state2, _state, 3, t1, t2, s2)
+
+#define SAVESTATE_2WAY(dst, src)\
+        dst[0][0] = src[0][0];\
+        dst[0][1] = src[0][1];\
+        dst[0][2] = src[0][2];\
+        dst[0][3] = src[0][3];\
+        dst[1][0] = src[1][0];\
+        dst[1][1] = src[1][1];\
+        dst[1][2] = src[1][2];\
+        dst[1][3] = src[1][3];\
+        dst[2][0] = src[2][0];\
+        dst[2][1] = src[2][1];\
+        dst[2][2] = src[2][2];\
+        dst[2][3] = src[2][3];\
+        dst[3][0] = src[3][0];\
+        dst[3][1] = src[3][1];\
+        dst[3][2] = src[3][2];\
+        dst[3][3] = src[3][3]
+
+// blockcount always 1
+void echo_2way_compress( echo_2way_context *ctx, const __m256i *pmsg,
+               unsigned int uBlockCount )
+{
+  unsigned int r, b, i, j;
+  __m256i t1, t2, s2, k1;
+  __m256i _state[4][4], _state2[4][4], _statebackup[4][4];
+
+  _state[ 0 ][ 0 ] = ctx->state[ 0 ][ 0 ];
+  _state[ 0 ][ 1 ] = ctx->state[ 0 ][ 1 ];
+  _state[ 0 ][ 2 ] = ctx->state[ 0 ][ 2 ];
+  _state[ 0 ][ 3 ] = ctx->state[ 0 ][ 3 ];
+  _state[ 1 ][ 0 ] = ctx->state[ 1 ][ 0 ];
+  _state[ 1 ][ 1 ] = ctx->state[ 1 ][ 1 ];
+  _state[ 1 ][ 2 ] = ctx->state[ 1 ][ 2 ];
+  _state[ 1 ][ 3 ] = ctx->state[ 1 ][ 3 ];
+  _state[ 2 ][ 0 ] = ctx->state[ 2 ][ 0 ];
+  _state[ 2 ][ 1 ] = ctx->state[ 2 ][ 1 ];
+  _state[ 2 ][ 2 ] = ctx->state[ 2 ][ 2 ];
+  _state[ 2 ][ 3 ] = ctx->state[ 2 ][ 3 ];
+  _state[ 3 ][ 0 ] = ctx->state[ 3 ][ 0 ];
+  _state[ 3 ][ 1 ] = ctx->state[ 3 ][ 1 ];
+  _state[ 3 ][ 2 ] = ctx->state[ 3 ][ 2 ];
+  _state[ 3 ][ 3 ] = ctx->state[ 3 ][ 3 ];
+
+  for ( b = 0; b < uBlockCount; b++ )
+  {
+    ctx->k = _mm256_add_epi64( ctx->k, ctx->const1536 );
+
+    for( j = ctx->uHashSize / 256; j < 4; j++ )
+    {
+      for ( i = 0; i < 4; i++ )
+      {
+        _state[ i ][ j ] = _mm256_load_si256(
+                     pmsg + 4 * (j - (ctx->uHashSize / 256)) + i );
+      }
+   }
+
+   // save state
+   SAVESTATE_2WAY( _statebackup, _state );
+
+   k1 = ctx->k;
+
+   for ( r = 0; r < ctx->uRounds / 2; r++ )
+   {
+       ECHO_ROUND_UNROLL2_2WAY;
+   }
+
+   if ( ctx->uHashSize == 256 )
+   {
+      for ( i = 0; i < 4; i++ )
+      {
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _state[ i ][ 1 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _state[ i ][ 2 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _state[ i ][ 3 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _statebackup[ i ][ 0 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _statebackup[ i ][ 1 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _statebackup[ i ][ 2 ] ) ;
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _statebackup[ i ][ 3 ] );
+       }
+    }
+    else
+    {
+       for ( i = 0; i < 4; i++ )
+       {
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _state[ i ][ 2 ] );
+         _state[ i ][ 1 ] = _mm256_xor_si256( _state[ i ][ 1 ],
+                                              _state[ i ][ 3 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ][ 0 ],
+                                              _statebackup[ i ][ 0 ] );
+         _state[ i ][ 0 ] = _mm256_xor_si256( _state[ i ] [0 ],
+                                              _statebackup[ i ][ 2 ] );
+         _state[ i ][ 1 ] = _mm256_xor_si256( _state[ i ][ 1 ],
+                                              _statebackup[ i ][ 1 ] );
+         _state[ i ][ 1 ] = _mm256_xor_si256( _state[ i ][ 1 ],
+                                              _statebackup[ i ][ 3 ] );
+      }
+    }
+    pmsg += ctx->uBlockLength;
+  }
+  SAVESTATE_2WAY(ctx->state, _state);
+
+}
+int echo_2way_init( echo_2way_context *ctx, int nHashSize )
+{
+        int i, j;
+
+   ctx->k = m256_zero;
+   ctx->processed_bits = 0;
+   ctx->uBufferBytes = 0;
+
+   switch( nHashSize )
+   {
+                case 256:
+                        ctx->uHashSize = 256;
+                        ctx->uBlockLength = 192;
+                        ctx->uRounds = 8;
+                        ctx->hashsize = m256_const2_64( 0, 0x100 );
+                        ctx->const1536 = m256_const2_64( 0, 0x600 );
+                        break;
+
+                case 512:
+                        ctx->uHashSize = 512;
+                        ctx->uBlockLength = 128;
+                        ctx->uRounds = 10;
+                        ctx->hashsize = m256_const2_64( 0, 0x200 );
+                        ctx->const1536 = m256_const2_64( 0, 0x400 );
+                        break;
+
+                default:
+                        return 1;
+        }
+
+        for( i = 0; i < 4; i++ )
+                for( j = 0; j < nHashSize / 256; j++ )
+                        ctx->state[ i ][ j ] = ctx->hashsize;
+
+        for( i = 0; i < 4; i++ )
+                for( j = nHashSize / 256; j < 4; j++ )
+                        ctx->state[ i ][ j ] = m256_zero;
+
+        return 0;
+}
+
+int echo_2way_update_close( echo_2way_context *state, void *hashval,
+                              const void *data, int databitlen )
+{
+// bytelen is either 32 (maybe), 64 or 80 or 128!
+// all are less than full block.
+
+   int vlen = databitlen / 128;  // * 4 lanes / 128 bits per lane
+   const int vblen = state->uBlockLength / 16; //  16 bytes per lane
+   __m256i remainingbits;
+
+   if ( databitlen == 1024 )
+   {
+      echo_2way_compress( state, data, 1 );
+      state->processed_bits = 1024;
+      remainingbits = m256_const2_64( 0, -1024 );
+      vlen = 0;
+   }
+   else
+   {
+      memcpy_256( state->buffer, data, vlen );
+      state->processed_bits += (unsigned int)( databitlen );
+      remainingbits = m256_const2_64( 0, databitlen );
+   }
+
+   state->buffer[ vlen ] = m256_const2_64( 0, 0x80 );
+   memset_zero_256( state->buffer + vlen + 1, vblen - vlen - 2 );
+   state->buffer[ vblen-2 ] = m256_const2_64( (uint64_t)state->uHashSize << 48, 0 );
+   state->buffer[ vblen-1 ] = m256_const2_64( 0, state->processed_bits );
+
+   state->k = _mm256_add_epi64( state->k, remainingbits );
+   state->k = _mm256_sub_epi64( state->k, state->const1536 );
+
+   echo_2way_compress( state, state->buffer, 1 );
+
+   _mm256_store_si256( (__m256i*)hashval + 0, state->state[ 0 ][ 0] );
+   _mm256_store_si256( (__m256i*)hashval + 1, state->state[ 1 ][ 0] );
+
+   if ( state->uHashSize == 512 )
+   {
+      _mm256_store_si256( (__m256i*)hashval + 2, state->state[ 2 ][ 0 ] );
+      _mm256_store_si256( (__m256i*)hashval + 3, state->state[ 3 ][ 0 ] );
+   }
+   return 0;
+}
+
+int echo_2way_full( echo_2way_context *ctx, void *hashval, int nHashSize,
+                    const void *data, int datalen )
+{
+   int i, j;
+   int databitlen = datalen * 8;
+   ctx->k = m256_zero;
+   ctx->processed_bits = 0;
+   ctx->uBufferBytes = 0;
+
+   switch( nHashSize )
+   {
+      case 256:
+         ctx->uHashSize = 256;
+         ctx->uBlockLength = 192;
+         ctx->uRounds = 8;
+         ctx->hashsize = m256_const2_64( 0, 0x100 );
+         ctx->const1536 = m256_const2_64( 0, 0x600 );
+         break;
+
+      case 512:
+         ctx->uHashSize = 512;
+         ctx->uBlockLength = 128;
+         ctx->uRounds = 10;
+         ctx->hashsize = m256_const2_64( 0, 0x200 );
+         ctx->const1536 = m256_const2_64( 0, 0x400 );
+         break;
+
+      default:
+         return 1;
+   }
+
+   for( i = 0; i < 4; i++ )
+      for( j = 0; j < nHashSize / 256; j++ )
+         ctx->state[ i ][ j ] = ctx->hashsize;
+
+   for( i = 0; i < 4; i++ )
+      for( j = nHashSize / 256; j < 4; j++ )
+         ctx->state[ i ][ j ] = m256_zero;
+
+   int vlen = datalen / 32;
+   const int vblen = ctx->uBlockLength / 16; //  16 bytes per lane
+   __m256i remainingbits;
+
+   if ( databitlen == 1024 )
+   {
+      echo_2way_compress( ctx, data, 1 );
+      ctx->processed_bits = 1024;
+      remainingbits = m256_const2_64( 0, -1024 );
+      vlen = 0;
+   }
+   else
+   {
+      vlen = databitlen / 128;  // * 4 lanes / 128 bits per lane
+      memcpy_256( ctx->buffer, data, vlen );
+      ctx->processed_bits += (unsigned int)( databitlen );
+      remainingbits = m256_const2_64( 0, databitlen );
+   }
+
+   ctx->buffer[ vlen ] = m256_const2_64( 0, 0x80 );
+   memset_zero_256( ctx->buffer + vlen + 1, vblen - vlen - 2 );
+   ctx->buffer[ vblen-2 ] = m256_const2_64( (uint64_t)ctx->uHashSize << 48, 0 );
+   ctx->buffer[ vblen-1 ] = m256_const2_64( 0, ctx->processed_bits );
+
+   ctx->k = _mm256_add_epi64( ctx->k, remainingbits );
+   ctx->k = _mm256_sub_epi64( ctx->k, ctx->const1536 );
+
+   echo_2way_compress( ctx, ctx->buffer, 1 );
+
+   _mm256_store_si256( (__m256i*)hashval + 0, ctx->state[ 0 ][ 0] );
+   _mm256_store_si256( (__m256i*)hashval + 1, ctx->state[ 1 ][ 0] );
+
+   if ( ctx->uHashSize == 512 )
+   {
+      _mm256_store_si256( (__m256i*)hashval + 2, ctx->state[ 2 ][ 0 ] );
+      _mm256_store_si256( (__m256i*)hashval + 3, ctx->state[ 3 ][ 0 ] );
+   }
+   return 0;
+}
+
+
+#endif   // VAES
--- a/algo/echo/echo-hash-4way.h
+++ b/algo/echo/echo-hash-4way.h
@@ -1,10 +1,12 @@
 #if !defined(ECHO_HASH_4WAY_H__)
 #define ECHO_HASH_4WAY_H__ 1

-#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__VAES__)

 #include "simd-utils.h"

+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
 typedef struct
 {
   __m512i    state[4][4];
@@ -20,6 +22,7 @@ typedef struct
   unsigned int   processed_bits;

 } echo_4way_context __attribute__ ((aligned (64)));
+#define echo512_4way_context echo_4way_context

 int echo_4way_init( echo_4way_context *state, int hashbitlen );
 #define echo512_4way_init( state ) echo_4way_init( state, 512 )
@@ -29,8 +32,8 @@ int echo_4way_update( echo_4way_context *state, const void *data,
    unsigned int databitlen);
 #define echo512_4way_update echo_4way_update

-int echo_close( echo_4way_context *state, void *hashval );
-#define echo512_4way_close echo_4way_close
+// int echo_4way_close( echo_4way_context *state, void *hashval );
+// #define echo512_4way_close echo_4way_close

 int echo_4way_update_close( echo_4way_context *state, void *hashval,
                              const void *data, int databitlen );
@@ -43,5 +46,45 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
 #define echo256_4way_full( state, hashval, data, datalen ) \
           echo_4way_full( state, hashval, 256, data, datalen )

-#endif 
-#endif
+#endif   // AVX512
+
+typedef struct
+{
+   __m256i    state[4][4];
+   __m256i    buffer[ 4 * 192 / 16 ];  // 4x128 interleaved 192 bytes
+   __m256i    k;
+   __m256i    hashsize;
+   __m256i    const1536;
+
+   unsigned int   uRounds;
+   unsigned int   uHashSize;
+   unsigned int   uBlockLength;
+   unsigned int   uBufferBytes;
+   unsigned int   processed_bits;
+
+} echo_2way_context __attribute__ ((aligned (64)));
+#define echo512_2way_context echo_2way_context
+
+int echo_2way_init( echo_2way_context *state, int hashbitlen );
+#define echo512_2way_init( state ) echo_2way_init( state, 512 )
+#define echo256_2way_init( state ) echo_2way_init( state, 256 )
+
+int echo_2way_update( echo_2way_context *state, const void *data,
+    unsigned int databitlen);
+#define echo512_2way_update echo_2way_update
+
+int echo_2way_update_close( echo_2way_context *state, void *hashval,
+                              const void *data, int databitlen );
+#define echo512_2way_update_close echo_2way_update_close
+
+int echo_2way_full( echo_2way_context *ctx, void *hashval, int nHashSize,
+                    const void *data, int datalen );
+#define echo512_2way_full( state, hashval, data, datalen ) \
+           echo_2way_full( state, hashval, 512, data, datalen )
+#define echo256_2way_full( state, hashval, data, datalen ) \
+           echo_2way_full( state, hashval, 256, data, datalen )
+
+
+#endif  // VAES
+
+#endif   // ECHO_HASH_4WAY_H__
--- a/algo/fugue/fugue-aesni.c
+++ b/algo/fugue/fugue-aesni.c
@@ -34,13 +34,11 @@ MYALIGN const unsigned long long _supermix4c[]	= {0x0706050403020000, 0x03020000
 MYALIGN const unsigned long long _supermix7a[]	= {0x010c0b060d080702, 0x0904030e03000104};
 MYALIGN const unsigned long long _supermix7b[]	= {0x8080808080808080, 0x0504070605040f06};
 MYALIGN const unsigned long long _k_n[] = {0x4E4E4E4E4E4E4E4E, 0x1B1B1B1B0E0E0E0E};
-MYALIGN const unsigned int _maskd3n[] = {0xffffffff, 0xffffffff, 0xffffffff, 0x00000000};
 MYALIGN const unsigned char _shift_one_mask[]   = {7, 4, 5, 6, 11, 8, 9, 10, 15, 12, 13, 14, 3, 0, 1, 2};
 MYALIGN const unsigned char _shift_four_mask[]  = {13, 14, 15, 12, 1, 2, 3, 0, 5, 6, 7, 4, 9, 10, 11, 8};
 MYALIGN const unsigned char _shift_seven_mask[] = {10, 11, 8, 9, 14, 15, 12, 13, 2, 3, 0, 1, 6, 7, 4, 5};
 MYALIGN const unsigned char _aes_shift_rows[]   = {0, 5, 10, 15, 4, 9, 14, 3, 8, 13, 2, 7, 12, 1, 6, 11};
 MYALIGN const unsigned int _inv_shift_rows[] = {0x070a0d00, 0x0b0e0104, 0x0f020508, 0x0306090c};
-MYALIGN const unsigned int _zero[] = {0x00000000, 0x00000000, 0x00000000, 0x00000000};
 MYALIGN const unsigned int _mul2mask[] = {0x1b1b0000, 0x00000000, 0x00000000, 0x00000000};
 MYALIGN const unsigned int _mul4mask[] = {0x2d361b00, 0x00000000, 0x00000000, 0x00000000};
 MYALIGN const unsigned int _lsbmask2[] = {0x03030303, 0x03030303, 0x03030303, 0x03030303};
@@ -61,7 +59,7 @@ MYALIGN const unsigned int _IV512[] = {

 #define UNPACK_S0(s0, s1, t1)\
   s1 = _mm_castps_si128(_mm_insert_ps(_mm_castsi128_ps(s1), _mm_castsi128_ps(s0), 0xc0));\
-   s0 = _mm_and_si128(s0, M128(_maskd3n))
+   s0 = mm128_mask_32( s0, 8 )

 #define CMIX(s1, s2, r1, r2, t1, t2)\
   t1 = s1;\
@@ -78,7 +76,7 @@ MYALIGN const unsigned int _IV512[] = {
 #define UNPACK_S0(s0, s1, t1)\
   t1 = _mm_shuffle_epi32(s0, _MM_SHUFFLE(3, 3, 3, 3));\
   s1 = _mm_castps_si128(_mm_move_ss(_mm_castsi128_ps(s1), _mm_castsi128_ps(t1)));\
-   s0 = _mm_and_si128(s0, M128(_maskd3n))
+   s0 = mm128_mask_32( s0, 8 )

 #define CMIX(s1, s2, r1, r2, t1, t2)\
   t1 = _mm_shuffle_epi32(s1, 0xf9);\
@@ -138,7 +136,7 @@ MYALIGN const unsigned int _IV512[] = {

 #define SUBSTITUTE(r0, _t1, _t2, _t3, _t0)\
 	_t2 = _mm_shuffle_epi8(r0, M128(_inv_shift_rows));\
-	_t2 = _mm_aesenclast_si128(_t2, M128(_zero))
+	_t2 = _mm_aesenclast_si128( _t2, m128_zero )
 	
 #define SUPERMIX(t0, t1, t2, t3, t4)\
 	PRESUPERMIX(t0, t1, t2, t3, t4);\
@@ -181,14 +179,14 @@ MYALIGN const unsigned int _IV512[] = {
 	SUPERMIX(_t2, _t3, _t0, _t1, r1c);\
 	_t0 = _mm_shuffle_epi32(r1c, 0x39);\
 	r2c = _mm_xor_si128(r2c, _t0);\
-	_t0 = _mm_and_si128(_t0, M128(_maskd3n));\
+   _t0 = mm128_mask_32( _t0, 8 ); \
 	r2d = _mm_xor_si128(r2d, _t0);\
 	UNPACK_S0(r1c, r1a, _t3);\
 	SUBSTITUTE(r2c, _t1, _t2, _t3, _t0);\
 	SUPERMIX(_t2, _t3, _t0, _t1, r2c);\
 	_t0 = _mm_shuffle_epi32(r2c, 0x39);\
 	r3c = _mm_xor_si128(r3c, _t0);\
-	_t0 = _mm_and_si128(_t0, M128(_maskd3n));\
+   _t0 = mm128_mask_32( _t0, 8 ); \
 	r3d = _mm_xor_si128(r3d, _t0);\
 	UNPACK_S0(r2c, r2a, _t3);\
 	SUBSTITUTE(r3c, _t1, _t2, _t3, _t0);\
@@ -203,21 +201,21 @@ MYALIGN const unsigned int _IV512[] = {
 	SUPERMIX(_t2, _t3, _t0, _t1, r1c);\
 	_t0 = _mm_shuffle_epi32(r1c, 0x39);\
 	r2c = _mm_xor_si128(r2c, _t0);\
-	_t0 = _mm_and_si128(_t0, M128(_maskd3n));\
+   _t0 = mm128_mask_32( _t0, 8 ); \
 	r2d = _mm_xor_si128(r2d, _t0);\
 	UNPACK_S0(r1c, r1a, _t3);\
 	SUBSTITUTE(r2c, _t1, _t2, _t3, _t0);\
 	SUPERMIX(_t2, _t3, _t0, _t1, r2c);\
 	_t0 = _mm_shuffle_epi32(r2c, 0x39);\
 	r3c = _mm_xor_si128(r3c, _t0);\
-	_t0 = _mm_and_si128(_t0, M128(_maskd3n));\
+   _t0 = mm128_mask_32( _t0, 8 ); \
 	r3d = _mm_xor_si128(r3d, _t0);\
 	UNPACK_S0(r2c, r2a, _t3);\
 	SUBSTITUTE(r3c, _t1, _t2, _t3, _t0);\
 	SUPERMIX(_t2, _t3, _t0, _t1, r3c);\
 	_t0 = _mm_shuffle_epi32(r3c, 0x39);\
 	r4c = _mm_xor_si128(r4c, _t0);\
-	_t0 = _mm_and_si128(_t0, M128(_maskd3n));\
+   _t0 = mm128_mask_32( _t0, 8 ); \
 	r4d = _mm_xor_si128(r4d, _t0);\
 	UNPACK_S0(r3c, r3a, _t3);\
 	SUBSTITUTE(r4c, _t1, _t2, _t3, _t0);\
@@ -462,7 +460,7 @@ HashReturn fugue512_Init(hashState_fugue *ctx, int nHashSize)
 	ctx->uBlockLength = 4;

 	for(i = 0; i < 6; i++)
-		ctx->state[i] = _mm_setzero_si128();
+		ctx->state[i] = m128_zero;

 	ctx->state[6]  = _mm_load_si128((__m128i*)_IV512 + 0);
 	ctx->state[7]  = _mm_load_si128((__m128i*)_IV512 + 1);
--- a/algo/fugue/fugue-aesni.h
+++ b/algo/fugue/fugue-aesni.h
@@ -14,10 +14,14 @@
 #ifndef FUGUE_HASH_API_H
 #define FUGUE_HASH_API_H

-#if defined(__AES__)
+#if defined(__AES__) 
+
+#if !defined(__SSE4_1__)
+#error "Unsupported configuration, AES needs SSE4.1. Compile without AES."
+#endif

 #include "algo/sha/sha3_common.h"
-#include <x86intrin.h>
+#include "simd-utils.h"


 typedef struct
--- a/algo/groestl/groestl256-hash-4way.c
+++ b/algo/groestl/groestl256-hash-4way.c
@@ -15,7 +15,9 @@
 #include "miner.h"
 #include "simd-utils.h"

-#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__AVX2__) && defined(__VAES__)
+
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)


 int groestl256_4way_init( groestl256_4way_context* ctx, uint64_t hashlen )
@@ -43,13 +45,13 @@ int groestl256_4way_init( groestl256_4way_context* ctx, uint64_t hashlen )
 }

 int groestl256_4way_full( groestl256_4way_context* ctx, void* output,
-                                const void* input, uint64_t databitlen )
+                                const void* input, uint64_t datalen )
 {
-   const int len = (int)databitlen / 128;
-   const int hashlen_m128i = 32 / 16;   // bytes to __m128i
+   const int len = (int)datalen >> 4;
+   const int hashlen_m128i = 32 >> 4;   // bytes to __m128i
   const int hash_offset = SIZE256 - hashlen_m128i;
   int rem = ctx->rem_ptr;
-   int blocks = len / SIZE256;
+   uint64_t blocks = len / SIZE256;
   __m512i* in = (__m512i*)input;
   int i;

@@ -87,21 +89,21 @@ int groestl256_4way_full( groestl256_4way_context* ctx, void* output,
   if ( i == SIZE256 - 1 )
   {        
       // only 1 vector left in buffer, all padding at once
-      ctx->buffer[i] = m512_const2_64( (uint64_t)blocks << 56, 0x80 ); 
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0x80 ); 
   }   
   else
   {
       // add first padding
-       ctx->buffer[i] = m512_const4_64( 0, 0x80, 0, 0x80 );
+       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
       // add zero padding
       for ( i += 1; i < SIZE256 - 1; i++ )
           ctx->buffer[i] = m512_zero;

       // add length padding, second last byte is zero unless blocks > 255
-      ctx->buffer[i] = m512_const2_64( (uint64_t)blocks << 56, 0 );
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
   }

-// digest final padding block and do output transform
+   // digest final padding block and do output transform
   TF512_4way( ctx->chaining, ctx->buffer );

   OF512_4way( ctx->chaining );
@@ -120,7 +122,7 @@ int groestl256_4way_update_close( groestl256_4way_context* ctx, void* output,
   const int hashlen_m128i = ctx->hashlen / 16;   // bytes to __m128i
   const int hash_offset = SIZE256 - hashlen_m128i;
   int rem = ctx->rem_ptr;
-   int blocks = len / SIZE256;
+   uint64_t blocks = len / SIZE256;
   __m512i* in = (__m512i*)input;
   int i;

@@ -144,20 +146,18 @@ int groestl256_4way_update_close( groestl256_4way_context* ctx, void* output,
   if ( i == SIZE256 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m512_const1_128( _mm_set_epi8(
-                      blocks, blocks>>8,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80 ) );
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0x80 );
   }
   else
   {
       // add first padding
-       ctx->buffer[i] = m512_const4_64( 0, 0x80, 0, 0x80 );
+       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
       // add zero padding
       for ( i += 1; i < SIZE256 - 1; i++ )
           ctx->buffer[i] = m512_zero;

       // add length padding, second last byte is zero unless blocks > 255
-       ctx->buffer[i] = m512_const1_128( _mm_set_epi8(
-                   blocks, blocks>>8, 0,0, 0,0, 0,0, 0,0, 0,0, 0,0, 0,0 ) );
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
   }

 // digest final padding block and do output transform
@@ -172,5 +172,159 @@ int groestl256_4way_update_close( groestl256_4way_context* ctx, void* output,
   return 0;
 }

-#endif   // VAES
+#endif   // AVX512

+// AVX2 + VAES
+
+int groestl256_2way_init( groestl256_2way_context* ctx, uint64_t hashlen )
+{
+  int i;
+
+  ctx->hashlen = hashlen;
+
+  if (ctx->chaining == NULL || ctx->buffer == NULL)
+    return 1;
+
+  for ( i = 0; i < SIZE256; i++ )
+  {
+     ctx->chaining[i] = m256_zero;
+     ctx->buffer[i]   = m256_zero;
+  }
+
+  // The only non-zero in the IV is len. It can be hard coded.
+  ctx->chaining[ 3 ] = m256_const2_64( 0, 0x0100000000000000 );
+
+  ctx->buf_ptr = 0;
+  ctx->rem_ptr = 0;
+
+  return 0;
+}
+
+int groestl256_2way_full( groestl256_2way_context* ctx, void* output,
+                                const void* input, uint64_t datalen )
+{
+   const int len = (int)datalen >> 4;
+   const int hashlen_m128i = 32 >> 4;   // bytes to __m128i
+   const int hash_offset = SIZE256 - hashlen_m128i;
+   int rem = ctx->rem_ptr;
+   uint64_t blocks = len / SIZE256;
+   __m256i* in = (__m256i*)input;
+   int i;
+
+   if (ctx->chaining == NULL || ctx->buffer == NULL)
+     return 1;
+
+   for ( i = 0; i < SIZE256; i++ )
+   {
+     ctx->chaining[i] = m256_zero;
+     ctx->buffer[i]   = m256_zero;
+   }
+
+   // The only non-zero in the IV is len. It can be hard coded.
+   ctx->chaining[ 3 ] = m256_const2_64( 0, 0x0100000000000000 );
+   ctx->buf_ptr = 0;
+   ctx->rem_ptr = 0;
+
+   // --- update ---
+
+   // digest any full blocks, process directly from input 
+   for ( i = 0; i < blocks; i++ )
+      TF512_2way( ctx->chaining, &in[ i * SIZE256 ] );
+   ctx->buf_ptr = blocks * SIZE256;
+
+   // copy any remaining data to buffer, it may already contain data
+   // from a previous update for a midstate precalc
+   for ( i = 0; i < len % SIZE256; i++ )
+       ctx->buffer[ rem + i ] = in[ ctx->buf_ptr + i ];
+   i += rem;    // use i as rem_ptr in final
+
+   //--- final ---
+
+   blocks++;      // adjust for final block
+
+   if ( i == SIZE256 - 1 )
+   {
+       // only 1 vector left in buffer, all padding at once
+      ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+   }
+   else
+   {
+       // add first padding
+       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       // add zero padding
+       for ( i += 1; i < SIZE256 - 1; i++ )
+           ctx->buffer[i] = m256_zero;
+
+       // add length padding, second last byte is zero unless blocks > 255
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+   }
+
+   // digest final padding block and do output transform
+   TF512_2way( ctx->chaining, ctx->buffer );
+
+   OF512_2way( ctx->chaining );
+
+   // store hash result in output 
+   for ( i = 0; i < hashlen_m128i; i++ )
+      casti_m256i( output, i ) = ctx->chaining[ hash_offset + i ];
+
+   return 0;
+}
+int groestl256_2way_update_close( groestl256_2way_context* ctx, void* output,
+                                const void* input, uint64_t databitlen )
+{
+   const int len = (int)databitlen / 128;
+   const int hashlen_m128i = ctx->hashlen / 16;   // bytes to __m128i
+   const int hash_offset = SIZE256 - hashlen_m128i;
+   int rem = ctx->rem_ptr;
+   uint64_t blocks = len / SIZE256;
+   __m256i* in = (__m256i*)input;
+   int i;
+
+   // --- update ---
+
+   // digest any full blocks, process directly from input 
+   for ( i = 0; i < blocks; i++ )
+      TF512_2way( ctx->chaining, &in[ i * SIZE256 ] );
+   ctx->buf_ptr = blocks * SIZE256;
+
+   // copy any remaining data to buffer, it may already contain data
+   // from a previous update for a midstate precalc
+   for ( i = 0; i < len % SIZE256; i++ )
+       ctx->buffer[ rem + i ] = in[ ctx->buf_ptr + i ];
+   i += rem;    // use i as rem_ptr in final
+
+   //--- final ---
+
+   blocks++;      // adjust for final block
+
+   if ( i == SIZE256 - 1 )
+   {
+       // only 1 vector left in buffer, all padding at once
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+   }
+   else
+   {
+       // add first padding
+       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       // add zero padding
+       for ( i += 1; i < SIZE256 - 1; i++ )
+           ctx->buffer[i] = m256_zero;
+
+       // add length padding, second last byte is zero unless blocks > 255
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+   }
+
+// digest final padding block and do output transform
+   TF512_2way( ctx->chaining, ctx->buffer );
+
+   OF512_2way( ctx->chaining );
+
+   // store hash result in output 
+   for ( i = 0; i < hashlen_m128i; i++ )
+      casti_m256i( output, i ) = ctx->chaining[ hash_offset + i ];
+
+   return 0;
+}
+
+#endif  // VAES
--- a/algo/groestl/groestl256-hash-4way.h
+++ b/algo/groestl/groestl256-hash-4way.h
@@ -18,8 +18,8 @@
 #endif
 #include <stdlib.h>

-#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
-   
+#if defined(__AVX2__) && defined(__VAES__)
+
 #define LENGTH (256)

 //#include "brg_endian.h"
@@ -48,6 +48,8 @@

 #define SIZE256 (SIZE_512/16)

+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
 typedef struct {
  __attribute__ ((aligned (128))) __m512i chaining[SIZE256];
  __attribute__ ((aligned (64))) __m512i buffer[SIZE256];
@@ -55,7 +57,7 @@ typedef struct {
  int blk_count;     // SIZE_m128i
  int buf_ptr;       // __m128i offset
  int rem_ptr;
-  int databitlen;    // bits
+//  int databitlen;    // bits
 } groestl256_4way_context;


@@ -74,5 +76,25 @@ int groestl256_4way_update_close( groestl256_4way_context*,  void*,
 int groestl256_4way_full( groestl256_4way_context*, void*,
                          const void*, uint64_t );

-#endif
-#endif 
+#endif  // AVX512
+
+typedef struct {
+  __attribute__ ((aligned (128))) __m256i chaining[SIZE256];
+  __attribute__ ((aligned (64))) __m256i buffer[SIZE256];
+  int hashlen;       // byte
+  int blk_count;     // SIZE_m128i
+  int buf_ptr;       // __m128i offset
+  int rem_ptr;
+//  int databitlen;    // bits
+} groestl256_2way_context;
+
+int groestl256_2way_init( groestl256_2way_context*, uint64_t );
+
+int groestl256_2way_update_close( groestl256_2way_context*,  void*,
+                                        const void*, uint64_t );
+
+int groestl256_2way_full( groestl256_2way_context*, void*,
+                          const void*, uint64_t );
+
+#endif  // VAES
+#endif  // GROESTL256_HASH_4WAY_H__
--- a/algo/groestl/groestl256-intr-4way.h
+++ b/algo/groestl/groestl256-intr-4way.h
@@ -7,13 +7,13 @@
 * This code is placed in the public domain
 */

-
 #if !defined(GROESTL256_INTR_4WAY_H__)
 #define GROESTL256_INTR_4WAY_H__ 1
      
 #include "groestl256-hash-4way.h"

-#if defined(__VAES__)
+#if defined(__AVX2__) && defined(__VAES__)
+
 static const __m128i round_const_l0[] __attribute__ ((aligned (64))) =
 {
   { 0x7060504030201000, 0xffffffffffffffff },
@@ -42,6 +42,8 @@ static const __m128i round_const_l7[] __attribute__ ((aligned (64))) =
   { 0x0000000000000000, 0x8696a6b6c6d6e6f6 }
 };

+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
 static const __m512i TRANSP_MASK = { 0x0d0509010c040800, 0x0f070b030e060a02,
                                     0x1d1519111c141810, 0x1f171b131e161a12,
                                     0x2d2529212c242820, 0x2f272b232e262a22,
@@ -499,5 +501,398 @@ void OF512_4way( __m512i* chaining )
  chaining[3] = xmm11;
 }

+#endif  // AVX512
+
+static const __m256i TRANSP_MASK_2WAY =
+             { 0x0d0509010c040800, 0x0f070b030e060a02,
+               0x1d1519111c141810, 0x1f171b131e161a12 };
+
+static const __m256i SUBSH_MASK0_2WAY =
+             { 0x0c0f0104070b0e00, 0x03060a0d08020509,
+               0x1c1f1114171b1e10, 0x13161a1d18121519 };
+
+static const __m256i SUBSH_MASK1_2WAY =
+             { 0x0e090205000d0801, 0x04070c0f0a03060b,
+               0x1e191215101d1801, 0x14171c1f1a13161b };
+
+static const __m256i SUBSH_MASK2_2WAY =
+               { 0x080b0306010f0a02, 0x05000e090c04070d,
+                 0x181b1316111f1a12, 0x15101e191c14171d };
+
+static const __m256i SUBSH_MASK3_2WAY =
+               { 0x0a0d040702090c03, 0x0601080b0e05000f,
+                 0x1a1d141712191c13, 0x1611181b1e15101f };
+
+static const __m256i SUBSH_MASK4_2WAY =
+               { 0x0b0e0500030a0d04, 0x0702090c0f060108,
+                 0x1b1e1510131a1d14, 0x1712191c1f161118 };
+
+static const __m256i SUBSH_MASK5_2WAY =
+               { 0x0d080601040c0f05, 0x00030b0e0907020a,
+                 0x1d181611141c1f15, 0x10131b1e1917121a };
+
+static const __m256i SUBSH_MASK6_2WAY =
+               { 0x0f0a0702050e0906, 0x01040d080b00030c,
+                 0x1f1a1712151e1916, 0x11141d181b10131c };
+
+static const __m256i SUBSH_MASK7_2WAY =
+               { 0x090c000306080b07, 0x02050f0a0d01040e,
+                 0x191c101316181b17, 0x12151f1a1d11141e, };
+
+#define tos(a)    #a
+#define tostr(a)  tos(a)
+
+/* xmm[i] will be multiplied by 2
+ * xmm[j] will be lost
+ * xmm[k] has to be all 0x1b */
+#define MUL2_2WAY(i, j, k){\
+  j = _mm256_xor_si256(j, j);\
+  j = _mm256_cmpgt_epi8(j, i );\
+  i = _mm256_add_epi8(i, i);\
+  j = _mm256_and_si256(j, k);\
+  i = _mm256_xor_si256(i, j);\
+}
+
+#define MixBytes_2way(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
+  /* t_i = a_i + a_{i+1} */\
+  b6 = a0;\
+  b7 = a1;\
+  a0 = _mm256_xor_si256(a0, a1);\
+  b0 = a2;\
+  a1 = _mm256_xor_si256(a1, a2);\
+  b1 = a3;\
+  a2 = _mm256_xor_si256(a2, a3);\
+  b2 = a4;\
+  a3 = _mm256_xor_si256(a3, a4);\
+  b3 = a5;\
+  a4 = _mm256_xor_si256(a4, a5);\
+  b4 = a6;\
+  a5 = _mm256_xor_si256(a5, a6);\
+  b5 = a7;\
+  a6 = _mm256_xor_si256(a6, a7);\
+  a7 = _mm256_xor_si256(a7, b6);\
+  \
+  /* build y4 y5 y6 ... in regs xmm8, xmm9, xmm10 by adding t_i*/\
+  b0 = _mm256_xor_si256(b0, a4);\
+  b6 = _mm256_xor_si256(b6, a4);\
+  b1 = _mm256_xor_si256(b1, a5);\
+  b7 = _mm256_xor_si256(b7, a5);\
+  b2 = _mm256_xor_si256(b2, a6);\
+  b0 = _mm256_xor_si256(b0, a6);\
+  /* spill values y_4, y_5 to memory */\
+  TEMP0 = b0;\
+  b3 = _mm256_xor_si256(b3, a7);\
+  b1 = _mm256_xor_si256(b1, a7);\
+  TEMP1 = b1;\
+  b4 = _mm256_xor_si256(b4, a0);\
+  b2 = _mm256_xor_si256(b2, a0);\
+  /* save values t0, t1, t2 to xmm8, xmm9 and memory */\
+  b0 = a0;\
+  b5 = _mm256_xor_si256(b5, a1);\
+  b3 = _mm256_xor_si256(b3, a1);\
+  b1 = a1;\
+  b6 = _mm256_xor_si256(b6, a2);\
+  b4 = _mm256_xor_si256(b4, a2);\
+  TEMP2 = a2;\
+  b7 = _mm256_xor_si256(b7, a3);\
+  b5 = _mm256_xor_si256(b5, a3);\
+  \
+  /* compute x_i = t_i + t_{i+3} */\
+  a0 = _mm256_xor_si256(a0, a3);\
+  a1 = _mm256_xor_si256(a1, a4);\
+  a2 = _mm256_xor_si256(a2, a5);\
+  a3 = _mm256_xor_si256(a3, a6);\
+  a4 = _mm256_xor_si256(a4, a7);\
+  a5 = _mm256_xor_si256(a5, b0);\
+  a6 = _mm256_xor_si256(a6, b1);\
+  a7 = _mm256_xor_si256(a7, TEMP2);\
+  \
+  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
+  /* compute w_i : add y_{i+4} */\
+  b1 = m256_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  MUL2_2WAY(a0, b0, b1);\
+  a0 = _mm256_xor_si256(a0, TEMP0);\
+  MUL2_2WAY(a1, b0, b1);\
+  a1 = _mm256_xor_si256(a1, TEMP1);\
+  MUL2_2WAY(a2, b0, b1);\
+  a2 = _mm256_xor_si256(a2, b2);\
+  MUL2_2WAY(a3, b0, b1);\
+  a3 = _mm256_xor_si256(a3, b3);\
+  MUL2_2WAY(a4, b0, b1);\
+  a4 = _mm256_xor_si256(a4, b4);\
+  MUL2_2WAY(a5, b0, b1);\
+  a5 = _mm256_xor_si256(a5, b5);\
+  MUL2_2WAY(a6, b0, b1);\
+  a6 = _mm256_xor_si256(a6, b6);\
+  MUL2_2WAY(a7, b0, b1);\
+  a7 = _mm256_xor_si256(a7, b7);\
+  \
+  /* compute v_i : double w_i      */\
+  /* add to y_4 y_5 .. v3, v4, ... */\
+  MUL2_2WAY(a0, b0, b1);\
+  b5 = _mm256_xor_si256(b5, a0);\
+  MUL2_2WAY(a1, b0, b1);\
+  b6 = _mm256_xor_si256(b6, a1);\
+  MUL2_2WAY(a2, b0, b1);\
+  b7 = _mm256_xor_si256(b7, a2);\
+  MUL2_2WAY(a5, b0, b1);\
+  b2 = _mm256_xor_si256(b2, a5);\
+  MUL2_2WAY(a6, b0, b1);\
+  b3 = _mm256_xor_si256(b3, a6);\
+  MUL2_2WAY(a7, b0, b1);\
+  b4 = _mm256_xor_si256(b4, a7);\
+  MUL2_2WAY(a3, b0, b1);\
+  MUL2_2WAY(a4, b0, b1);\
+  b0 = TEMP0;\
+  b1 = TEMP1;\
+  b0 = _mm256_xor_si256(b0, a3);\
+  b1 = _mm256_xor_si256(b1, a4);\
+}/*MixBytes*/
+
+#define ROUND_2WAY(i, a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
+  /* AddRoundConstant */\
+  b1 = m256_const2_64( 0xffffffffffffffff, 0 ); \
+  a0 = _mm256_xor_si256( a0, m256_const1_128( round_const_l0[i] ) );\
+  a1 = _mm256_xor_si256( a1, b1 );\
+  a2 = _mm256_xor_si256( a2, b1 );\
+  a3 = _mm256_xor_si256( a3, b1 );\
+  a4 = _mm256_xor_si256( a4, b1 );\
+  a5 = _mm256_xor_si256( a5, b1 );\
+  a6 = _mm256_xor_si256( a6, b1 );\
+  a7 = _mm256_xor_si256( a7, m256_const1_128( round_const_l7[i] ) );\
+  \
+  /* ShiftBytes + SubBytes (interleaved) */\
+  b0 = _mm256_xor_si256( b0, b0 );\
+  a0 = _mm256_shuffle_epi8( a0, SUBSH_MASK0_2WAY );\
+  a0 = _mm256_aesenclast_epi128(a0, b0 );\
+  a1 = _mm256_shuffle_epi8( a1, SUBSH_MASK1_2WAY );\
+  a1 = _mm256_aesenclast_epi128(a1, b0 );\
+  a2 = _mm256_shuffle_epi8( a2, SUBSH_MASK2_2WAY );\
+  a2 = _mm256_aesenclast_epi128(a2, b0 );\
+  a3 = _mm256_shuffle_epi8( a3, SUBSH_MASK3_2WAY );\
+  a3 = _mm256_aesenclast_epi128(a3, b0 );\
+  a4 = _mm256_shuffle_epi8( a4, SUBSH_MASK4_2WAY );\
+  a4 = _mm256_aesenclast_epi128(a4, b0 );\
+  a5 = _mm256_shuffle_epi8( a5, SUBSH_MASK5_2WAY );\
+  a5 = _mm256_aesenclast_epi128(a5, b0 );\
+  a6 = _mm256_shuffle_epi8( a6, SUBSH_MASK6_2WAY );\
+  a6 = _mm256_aesenclast_epi128(a6, b0 );\
+  a7 = _mm256_shuffle_epi8( a7, SUBSH_MASK7_2WAY );\
+  a7 = _mm256_aesenclast_epi128( a7, b0 );\
+  \
+  /* MixBytes */\
+  MixBytes_2way(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7);\
+\
+}
+
+/* 10 rounds, P and Q in parallel */
+#define ROUNDS_P_Q_2WAY(){\
+  ROUND_2WAY(0, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+  ROUND_2WAY(1, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  ROUND_2WAY(2, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+  ROUND_2WAY(3, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  ROUND_2WAY(4, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+  ROUND_2WAY(5, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  ROUND_2WAY(6, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+  ROUND_2WAY(7, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  ROUND_2WAY(8, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+  ROUND_2WAY(9, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+}
+
+#define Matrix_Transpose_A_2way(i0, i1, i2, i3, o1, o2, o3, t0){\
+  t0 = TRANSP_MASK_2WAY;\
+  \
+  i0 = _mm256_shuffle_epi8( i0, t0 );\
+  i1 = _mm256_shuffle_epi8( i1, t0 );\
+  i2 = _mm256_shuffle_epi8( i2, t0 );\
+  i3 = _mm256_shuffle_epi8( i3, t0 );\
+  \
+  o1 = i0;\
+  t0 = i2;\
+  \
+  i0 = _mm256_unpacklo_epi16( i0, i1 );\
+  o1 = _mm256_unpackhi_epi16( o1, i1 );\
+  i2 = _mm256_unpacklo_epi16( i2, i3 );\
+  t0 = _mm256_unpackhi_epi16( t0, i3 );\
+  \
+  i0 = _mm256_shuffle_epi32( i0, 216 );\
+  o1 = _mm256_shuffle_epi32( o1, 216 );\
+  i2 = _mm256_shuffle_epi32( i2, 216 );\
+  t0 = _mm256_shuffle_epi32( t0, 216 );\
+  \
+  o2 = i0;\
+  o3 = o1;\
+  \
+  i0 = _mm256_unpacklo_epi32( i0, i2 );\
+  o1 = _mm256_unpacklo_epi32( o1, t0 );\
+  o2 = _mm256_unpackhi_epi32( o2, i2 );\
+  o3 = _mm256_unpackhi_epi32( o3, t0 );\
+}/**/
+
+#define Matrix_Transpose_B_2way(i0, i1, i2, i3, i4, i5, i6, i7, o1, o2, o3, o4, o5, o6, o7){\
+  o1 = i0;\
+  o2 = i1;\
+  i0 = _mm256_unpacklo_epi64( i0, i4 );\
+  o1 = _mm256_unpackhi_epi64( o1, i4 );\
+  o3 = i1;\
+  o4 = i2;\
+  o2 = _mm256_unpacklo_epi64( o2, i5 );\
+  o3 = _mm256_unpackhi_epi64( o3, i5 );\
+  o5 = i2;\
+  o6 = i3;\
+  o4 = _mm256_unpacklo_epi64( o4, i6 );\
+  o5 = _mm256_unpackhi_epi64( o5, i6 );\
+  o7 = i3;\
+  o6 = _mm256_unpacklo_epi64( o6, i7 );\
+  o7 = _mm256_unpackhi_epi64( o7, i7 );\
+}/**/
+
+#define Matrix_Transpose_B_INV_2way(i0, i1, i2, i3, i4, i5, i6, i7, o0, o1, o2, o3){\
+  o0 = i0;\
+  i0 = _mm256_unpacklo_epi64( i0, i1 );\
+  o0 = _mm256_unpackhi_epi64( o0, i1 );\
+  o1 = i2;\
+  i2 = _mm256_unpacklo_epi64( i2, i3 );\
+  o1 = _mm256_unpackhi_epi64( o1, i3 );\
+  o2 = i4;\
+  i4 = _mm256_unpacklo_epi64( i4, i5 );\
+  o2 = _mm256_unpackhi_epi64( o2, i5 );\
+  o3 = i6;\
+  i6 = _mm256_unpacklo_epi64( i6, i7 );\
+  o3 = _mm256_unpackhi_epi64( o3, i7 );\
+}/**/
+
+#define Matrix_Transpose_O_B_2way(i0, i1, i2, i3, i4, i5, i6, i7, t0){\
+  t0 = _mm256_xor_si256( t0, t0 );\
+  i1 = i0;\
+  i3 = i2;\
+  i5 = i4;\
+  i7 = i6;\
+  i0 = _mm256_unpacklo_epi64( i0, t0 );\
+  i1 = _mm256_unpackhi_epi64( i1, t0 );\
+  i2 = _mm256_unpacklo_epi64( i2, t0 );\
+  i3 = _mm256_unpackhi_epi64( i3, t0 );\
+  i4 = _mm256_unpacklo_epi64( i4, t0 );\
+  i5 = _mm256_unpackhi_epi64( i5, t0 );\
+  i6 = _mm256_unpacklo_epi64( i6, t0 );\
+  i7 = _mm256_unpackhi_epi64( i7, t0 );\
+}/**/
+
+#define Matrix_Transpose_O_B_INV_2way(i0, i1, i2, i3, i4, i5, i6, i7){\
+  i0 = _mm256_unpacklo_epi64( i0, i1 );\
+  i2 = _mm256_unpacklo_epi64( i2, i3 );\
+  i4 = _mm256_unpacklo_epi64( i4, i5 );\
+  i6 = _mm256_unpacklo_epi64( i6, i7 );\
+}/**/
+
+void TF512_2way( __m256i* chaining, __m256i* message )
+{
+  static __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  static __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  static __m256i TEMP0;
+  static __m256i TEMP1;
+  static __m256i TEMP2;
+
+  /* load message into registers xmm12 - xmm15 */
+  xmm12 = message[0];
+  xmm13 = message[1];
+  xmm14 = message[2];
+  xmm15 = message[3];
+
+  /* transform message M from column ordering into row ordering */
+  /* we first put two rows (64 bit) of the message into one 128-bit xmm register */
+  Matrix_Transpose_A_2way(xmm12, xmm13, xmm14, xmm15, xmm2, xmm6, xmm7, xmm0);
+
+  /* load previous chaining value */
+  /* we first put two rows (64 bit) of the CV into one 128-bit xmm register */
+  xmm8 = chaining[0];
+  xmm0 = chaining[1];
+  xmm4 = chaining[2];
+  xmm5 = chaining[3];
+
+  /* xor message to CV get input of P */
+  /* result: CV+M in xmm8, xmm0, xmm4, xmm5 */
+  xmm8 = _mm256_xor_si256( xmm8, xmm12 );
+  xmm0 = _mm256_xor_si256( xmm0, xmm2 );
+  xmm4 = _mm256_xor_si256( xmm4, xmm6 );
+  xmm5 = _mm256_xor_si256( xmm5, xmm7 );
+
+  /* there are now 2 rows of the Groestl state (P and Q) in each xmm register */
+  /* unpack to get 1 row of P (64 bit) and Q (64 bit) into one xmm register */
+  /* result: the 8 rows of P and Q in xmm8 - xmm12 */
+  Matrix_Transpose_B_2way(xmm8, xmm0, xmm4, xmm5, xmm12, xmm2, xmm6, xmm7, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);
+
+  /* compute the two permutations P and Q in parallel */
+  ROUNDS_P_Q_2WAY();
+
+  /* unpack again to get two rows of P or two rows of Q in one xmm register */
+  Matrix_Transpose_B_INV_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3);
+
+  /* xor output of P and Q */
+  /* result: P(CV+M)+Q(M) in xmm0...xmm3 */
+  xmm0 = _mm256_xor_si256( xmm0, xmm8 );
+  xmm1 = _mm256_xor_si256( xmm1, xmm10 );
+  xmm2 = _mm256_xor_si256( xmm2, xmm12 );
+  xmm3 = _mm256_xor_si256( xmm3, xmm14 );
+
+  /* xor CV (feed-forward) */
+  /* result: P(CV+M)+Q(M)+CV in xmm0...xmm3 */
+  xmm0 = _mm256_xor_si256( xmm0, (chaining[0]) );
+  xmm1 = _mm256_xor_si256( xmm1, (chaining[1]) );
+  xmm2 = _mm256_xor_si256( xmm2, (chaining[2]) );
+  xmm3 = _mm256_xor_si256( xmm3, (chaining[3]) );
+
+  /* store CV */
+  chaining[0] = xmm0;
+  chaining[1] = xmm1;
+  chaining[2] = xmm2;
+  chaining[3] = xmm3;
+
+  return;
+}
+  
+void OF512_2way( __m256i* chaining )
+{
+  static __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  static __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  static __m256i TEMP0;
+  static __m256i TEMP1;
+  static __m256i TEMP2;
+
+  /* load CV into registers xmm8, xmm10, xmm12, xmm14 */
+  xmm8 = chaining[0];
+  xmm10 = chaining[1];
+  xmm12 = chaining[2];
+  xmm14 = chaining[3];
+
+  /* there are now 2 rows of the CV in one xmm register */
+  /* unpack to get 1 row of P (64 bit) into one half of an xmm register */
+  /* result: the 8 input rows of P in xmm8 - xmm15 */
+  Matrix_Transpose_O_B_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0);
+
+  /* compute the permutation P */
+  /* result: the output of P(CV) in xmm8 - xmm15 */
+  ROUNDS_P_Q_2WAY();
+
+  /* unpack again to get two rows of P in one xmm register */
+  /* result: P(CV) in xmm8, xmm10, xmm12, xmm14 */
+  Matrix_Transpose_O_B_INV_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);
+
+  /* xor CV to P output (feed-forward) */
+  /* result: P(CV)+CV in xmm8, xmm10, xmm12, xmm14 */
+  xmm8  = _mm256_xor_si256( xmm8,  (chaining[0]) );
+  xmm10 = _mm256_xor_si256( xmm10, (chaining[1]) );
+  xmm12 = _mm256_xor_si256( xmm12, (chaining[2]) );
+  xmm14 = _mm256_xor_si256( xmm14, (chaining[3]) );
+
+  /* transform state back from row ordering into column ordering */
+  /* result: final hash value in xmm9, xmm11 */
+  Matrix_Transpose_A_2way(xmm8, xmm10, xmm12, xmm14, xmm4, xmm9, xmm11, xmm0);
+
+  /* we only need to return the truncated half of the state */
+  chaining[2] = xmm9;
+  chaining[3] = xmm11;
+}
+
 #endif  // VAES
-#endif  // GROESTL512_INTR_4WAY_H__
+#endif  // GROESTL256_INTR_4WAY_H__
--- a/algo/groestl/groestl512-hash-4way.c
+++ b/algo/groestl/groestl512-hash-4way.c
@@ -15,7 +15,9 @@
 #include "miner.h"
 #include "simd-utils.h"

-#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__AVX2__) && defined(__VAES__)
+
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

 int groestl512_4way_init( groestl512_4way_context* ctx, uint64_t hashlen )
 {
@@ -41,7 +43,7 @@ int groestl512_4way_update_close( groestl512_4way_context* ctx, void* output,
   const int hashlen_m128i = 64 / 16;   // bytes to __m128i
   const int hash_offset = SIZE512 - hashlen_m128i;
   int rem = ctx->rem_ptr;
-   int blocks = len / SIZE512;
+   uint64_t blocks = len / SIZE512;
   __m512i* in = (__m512i*)input;
   int i;

@@ -62,16 +64,14 @@ int groestl512_4way_update_close( groestl512_4way_context* ctx, void* output,
   if ( i == SIZE512 - 1 )
   {        
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m512_const1_128( _mm_set_epi8(
-                      blocks, blocks>>8,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0x80 ) );
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0x80 );
   }   
   else
   {
-       ctx->buffer[i] = m512_const4_64( 0, 0x80, 0, 0x80 );
+       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m512_zero;
-       ctx->buffer[i] = m512_const1_128( _mm_set_epi8(
-                   blocks, blocks>>8, 0,0, 0,0, 0,0, 0,0, 0,0, 0,0, 0,0 ) );
+       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
   }

   TF1024_4way( ctx->chaining, ctx->buffer );
@@ -122,7 +122,7 @@ int groestl512_4way_full( groestl512_4way_context* ctx, void* output,
   }
   else
   {
-       ctx->buffer[i] = m512_const4_64( 0, 0x80, 0, 0x80 );
+       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m512_zero;
       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
@@ -137,5 +137,128 @@ int groestl512_4way_full( groestl512_4way_context* ctx, void* output,
   return 0;
 }

+#endif   // AVX512
+
+
+// AVX2 + VAES
+
+int groestl512_2way_init( groestl512_2way_context* ctx, uint64_t hashlen )
+{
+  if (ctx->chaining == NULL || ctx->buffer == NULL)
+    return 1;
+
+  memset_zero_256( ctx->chaining, SIZE512 );
+  memset_zero_256( ctx->buffer, SIZE512 );
+
+  // The only non-zero in the IV is len. It can be hard coded.
+  ctx->chaining[ 6 ] = m256_const2_64( 0x0200000000000000, 0 );
+
+  ctx->buf_ptr = 0;
+  ctx->rem_ptr = 0;
+
+  return 0;
+}
+
+int groestl512_2way_update_close( groestl512_2way_context* ctx, void* output,
+                                const void* input, uint64_t databitlen )
+{
+   const int len = (int)databitlen / 128;
+   const int hashlen_m128i = 64 / 16;   // bytes to __m128i
+   const int hash_offset = SIZE512 - hashlen_m128i;
+   int rem = ctx->rem_ptr;
+   uint64_t blocks = len / SIZE512;
+   __m256i* in = (__m256i*)input;
+   int i;
+
+   // --- update ---
+
+   for ( i = 0; i < blocks; i++ )
+      TF1024_2way( ctx->chaining, &in[ i * SIZE512 ] );
+   ctx->buf_ptr = blocks * SIZE512;
+
+   for ( i = 0; i < len % SIZE512; i++ )
+       ctx->buffer[ rem + i ] = in[ ctx->buf_ptr + i ];
+   i += rem;
+
+   //--- final ---
+
+   blocks++;      // adjust for final block
+
+   if ( i == SIZE512 - 1 )
+   {
+       // only 1 vector left in buffer, all padding at once
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+   }
+   else
+   {
+       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       for ( i += 1; i < SIZE512 - 1; i++ )
+           ctx->buffer[i] = m256_zero;
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+   }
+
+   TF1024_2way( ctx->chaining, ctx->buffer );
+   OF1024_2way( ctx->chaining );
+
+   for ( i = 0; i < hashlen_m128i; i++ )
+      casti_m256i( output, i ) = ctx->chaining[ hash_offset + i ];
+
+   return 0;
+}
+
+int groestl512_2way_full( groestl512_2way_context* ctx, void* output,
+                          const void* input, uint64_t datalen )
+{
+   const int len = (int)datalen >> 4;
+   const int hashlen_m128i = 64 >> 4;   // bytes to __m128i
+   const int hash_offset = SIZE512 - hashlen_m128i;
+   uint64_t blocks = len / SIZE512;
+   __m256i* in = (__m256i*)input;
+   int i;
+
+   // --- init ---
+
+   memset_zero_256( ctx->chaining, SIZE512 );
+   memset_zero_256( ctx->buffer, SIZE512 );
+   ctx->chaining[ 6 ] = m256_const2_64( 0x0200000000000000, 0 );
+   ctx->buf_ptr = 0;
+   ctx->rem_ptr = 0;
+
+   // --- update ---
+
+   for ( i = 0; i < blocks; i++ )
+      TF1024_2way( ctx->chaining, &in[ i * SIZE512 ] );
+   ctx->buf_ptr = blocks * SIZE512;
+
+   for ( i = 0; i < len % SIZE512; i++ )
+       ctx->buffer[ ctx->rem_ptr + i ] = in[ ctx->buf_ptr + i ];
+   i += ctx->rem_ptr;
+
+   // --- close ---
+
+   blocks++;
+
+   if ( i == SIZE512 - 1 )
+   {
+       // only 1 vector left in buffer, all padding at once
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+   }
+   else
+   {
+       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       for ( i += 1; i < SIZE512 - 1; i++ )
+           ctx->buffer[i] = m256_zero;
+       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+   }
+
+   TF1024_2way( ctx->chaining, ctx->buffer );
+   OF1024_2way( ctx->chaining );
+
+   for ( i = 0; i < hashlen_m128i; i++ )
+      casti_m256i( output, i ) = ctx->chaining[ hash_offset + i ];
+
+   return 0;
+}
+   
 #endif   // VAES

--- a/algo/groestl/groestl512-hash-4way.h
+++ b/algo/groestl/groestl512-hash-4way.h
@@ -10,7 +10,7 @@
 #endif
 #include <stdlib.h>

-#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__AVX2__) && defined(__VAES__)

 #define LENGTH (512)

@@ -36,20 +36,19 @@

 #define SIZE512 (SIZE_1024/16)

+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
 typedef struct {
  __attribute__ ((aligned (128))) __m512i chaining[SIZE512];
  __attribute__ ((aligned (64))) __m512i buffer[SIZE512];
  int blk_count;     // SIZE_m128i
  int buf_ptr;       // __m128i offset
  int rem_ptr;
-  int databitlen;    // bits
 } groestl512_4way_context;


 int groestl512_4way_init( groestl512_4way_context*, uint64_t );

-//int reinit_groestl( hashState_groestl* );
-
 int groestl512_4way_update( groestl512_4way_context*, const void*,
                              uint64_t );
 int groestl512_4way_close( groestl512_4way_context*, void* );
@@ -58,5 +57,29 @@ int groestl512_4way_update_close( groestl512_4way_context*,  void*,
 int groestl512_4way_full( groestl512_4way_context*,  void*,
                          const void*, uint64_t );

+#endif   // AVX512
+
+// AVX2 + VAES
+
+typedef struct {
+  __attribute__ ((aligned (128))) __m256i chaining[SIZE512];
+  __attribute__ ((aligned (64))) __m256i buffer[SIZE512];
+  int blk_count;     // SIZE_m128i
+  int buf_ptr;       // __m128i offset
+  int rem_ptr;
+} groestl512_2way_context;
+
+
+int groestl512_2way_init( groestl512_2way_context*, uint64_t );
+
+int groestl512_2way_update( groestl512_2way_context*, const void*,
+                              uint64_t );
+int groestl512_2way_close( groestl512_2way_context*, void* );
+int groestl512_2way_update_close( groestl512_2way_context*,  void*,
+                                        const void*, uint64_t );
+int groestl512_2way_full( groestl512_2way_context*,  void*,
+                          const void*, uint64_t );
+
+
 #endif   // VAES
 #endif   // GROESTL512_HASH_4WAY_H__
--- a/algo/groestl/groestl512-intr-4way.h
+++ b/algo/groestl/groestl512-intr-4way.h
@@ -7,13 +7,12 @@
 * This code is placed in the public domain
 */

-
 #if !defined(GROESTL512_INTR_4WAY_H__)
 #define GROESTL512_INTR_4WAY_H__ 1
      
 #include "groestl512-hash-4way.h"

-#if defined(__VAES__)
+#if defined(__AVX2__) && defined(__VAES__)

 static const __m128i round_const_p[] __attribute__ ((aligned (64))) =
 {
@@ -51,6 +50,8 @@ static const __m128i round_const_q[] __attribute__ ((aligned (64))) =
   { 0x8292a2b2c2d2e2f2, 0x0212223242526272 }
 };

+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
 static const __m512i TRANSP_MASK = { 0x0d0509010c040800, 0x0f070b030e060a02,
                                     0x1d1519111c141810, 0x1f171b131e161a12,
                                     0x2d2529212c242820, 0x2f272b232e262a22,
@@ -661,5 +662,578 @@ void OF1024_4way( __m512i* chaining )
  return;
 }

+#endif  // AVX512
+
+// AVX2 + VAES
+
+static const __m256i TRANSP_MASK_2WAY =
+    { 0x0d0509010c040800, 0x0f070b030e060a02,
+      0x1d1519111c141810, 0x1f171b131e161a12 };
+
+static const __m256i SUBSH_MASK0_2WAY =
+    { 0x0b0e0104070a0d00, 0x0306090c0f020508,
+      0x1b1e1114171a1d10, 0x1316191c1f121518 };
+
+static const __m256i SUBSH_MASK1_2WAY =
+    { 0x0c0f0205080b0e01, 0x04070a0d00030609,
+      0x1c1f1215181b1e11, 0x14171a1d10131619 };
+
+static const __m256i SUBSH_MASK2_2WAY =
+    { 0x0d000306090c0f02, 0x05080b0e0104070a,
+      0x1d101316191c1f12, 0x15181b1e1114171a };
+
+static const __m256i SUBSH_MASK3_2WAY =
+    { 0x0e0104070a0d0003, 0x06090c0f0205080b,
+      0x1e1114171a1d1013, 0x16191c1f1215181b };
+
+static const __m256i SUBSH_MASK4_2WAY = 
+    { 0x0f0205080b0e0104, 0x070a0d000306090c,
+      0x1f1215181b1e1114, 0x171a1d101316191c };
+
+static const __m256i SUBSH_MASK5_2WAY =
+    { 0x000306090c0f0205, 0x080b0e0104070a0d,
+      0x101316191c1f1215, 0x181b1e1114171a1d };
+
+static const __m256i SUBSH_MASK6_2WAY =
+    { 0x0104070a0d000306, 0x090c0f0205080b0e,
+      0x1114171a1d101316, 0x191c1f1215181b1e };
+
+static const __m256i SUBSH_MASK7_2WAY =
+    { 0x06090c0f0205080b, 0x0e0104070a0d0003,
+      0x16191c1f1215181b, 0x1e1114171a1d1013 };
+
+#define tos(a)    #a
+#define tostr(a)  tos(a)
+
+/* xmm[i] will be multiplied by 2
+ * xmm[j] will be lost
+ * xmm[k] has to be all 0x1b */
+#define MUL2_2WAY(i, j, k){\
+  j = _mm256_xor_si256(j, j);\
+  j = _mm256_cmpgt_epi8(j, i );\
+  i = _mm256_add_epi8(i, i);\
+  j = _mm256_and_si256(j, k);\
+  i = _mm256_xor_si256(i, j);\
+}
+
+#define MixBytes_2way(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
+  /* t_i = a_i + a_{i+1} */\
+  b6 = a0;\
+  b7 = a1;\
+  a0 = _mm256_xor_si256(a0, a1);\
+  b0 = a2;\
+  a1 = _mm256_xor_si256(a1, a2);\
+  b1 = a3;\
+  a2 = _mm256_xor_si256(a2, a3);\
+  b2 = a4;\
+  a3 = _mm256_xor_si256(a3, a4);\
+  b3 = a5;\
+  a4 = _mm256_xor_si256(a4, a5);\
+  b4 = a6;\
+  a5 = _mm256_xor_si256(a5, a6);\
+  b5 = a7;\
+  a6 = _mm256_xor_si256(a6, a7);\
+  a7 = _mm256_xor_si256(a7, b6);\
+  \
+  /* build y4 y5 y6 ... in regs xmm8, xmm9, xmm10 by adding t_i*/\
+  b0 = _mm256_xor_si256(b0, a4);\
+  b6 = _mm256_xor_si256(b6, a4);\
+  b1 = _mm256_xor_si256(b1, a5);\
+  b7 = _mm256_xor_si256(b7, a5);\
+  b2 = _mm256_xor_si256(b2, a6);\
+  b0 = _mm256_xor_si256(b0, a6);\
+  /* spill values y_4, y_5 to memory */\
+  TEMP0 = b0;\
+  b3 = _mm256_xor_si256(b3, a7);\
+  b1 = _mm256_xor_si256(b1, a7);\
+  TEMP1 = b1;\
+  b4 = _mm256_xor_si256(b4, a0);\
+  b2 = _mm256_xor_si256(b2, a0);\
+  /* save values t0, t1, t2 to xmm8, xmm9 and memory */\
+  b0 = a0;\
+  b5 = _mm256_xor_si256(b5, a1);\
+  b3 = _mm256_xor_si256(b3, a1);\
+  b1 = a1;\
+  b6 = _mm256_xor_si256(b6, a2);\
+  b4 = _mm256_xor_si256(b4, a2);\
+  TEMP2 = a2;\
+  b7 = _mm256_xor_si256(b7, a3);\
+  b5 = _mm256_xor_si256(b5, a3);\
+  \
+  /* compute x_i = t_i + t_{i+3} */\
+  a0 = _mm256_xor_si256(a0, a3);\
+  a1 = _mm256_xor_si256(a1, a4);\
+  a2 = _mm256_xor_si256(a2, a5);\
+  a3 = _mm256_xor_si256(a3, a6);\
+  a4 = _mm256_xor_si256(a4, a7);\
+  a5 = _mm256_xor_si256(a5, b0);\
+  a6 = _mm256_xor_si256(a6, b1);\
+  a7 = _mm256_xor_si256(a7, TEMP2);\
+  \
+  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
+  /* compute w_i : add y_{i+4} */\
+  b1 = m256_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  MUL2_2WAY(a0, b0, b1);\
+  a0 = _mm256_xor_si256(a0, TEMP0);\
+  MUL2_2WAY(a1, b0, b1);\
+  a1 = _mm256_xor_si256(a1, TEMP1);\
+  MUL2_2WAY(a2, b0, b1);\
+  a2 = _mm256_xor_si256(a2, b2);\
+  MUL2_2WAY(a3, b0, b1);\
+  a3 = _mm256_xor_si256(a3, b3);\
+  MUL2_2WAY(a4, b0, b1);\
+  a4 = _mm256_xor_si256(a4, b4);\
+  MUL2_2WAY(a5, b0, b1);\
+  a5 = _mm256_xor_si256(a5, b5);\
+  MUL2_2WAY(a6, b0, b1);\
+  a6 = _mm256_xor_si256(a6, b6);\
+  MUL2_2WAY(a7, b0, b1);\
+  a7 = _mm256_xor_si256(a7, b7);\
+  \
+  /* compute v_i : double w_i      */\
+  /* add to y_4 y_5 .. v3, v4, ... */\
+  MUL2_2WAY(a0, b0, b1);\
+  b5 = _mm256_xor_si256(b5, a0);\
+  MUL2_2WAY(a1, b0, b1);\
+  b6 = _mm256_xor_si256(b6, a1);\
+  MUL2_2WAY(a2, b0, b1);\
+  b7 = _mm256_xor_si256(b7, a2);\
+  MUL2_2WAY(a5, b0, b1);\
+  b2 = _mm256_xor_si256(b2, a5);\
+  MUL2_2WAY(a6, b0, b1);\
+  b3 = _mm256_xor_si256(b3, a6);\
+  MUL2_2WAY(a7, b0, b1);\
+  b4 = _mm256_xor_si256(b4, a7);\
+  MUL2_2WAY(a3, b0, b1);\
+  MUL2_2WAY(a4, b0, b1);\
+  b0 = TEMP0;\
+  b1 = TEMP1;\
+  b0 = _mm256_xor_si256(b0, a3);\
+  b1 = _mm256_xor_si256(b1, a4);\
+}/*MixBytes*/
+
+/* one round
+ * a0-a7 = input rows
+ * b0-b7 = output rows
+ */
+#define SUBMIX_2WAY(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
+  /* SubBytes */\
+  b0 = _mm256_xor_si256( b0, b0 );\
+  a0 = _mm256_aesenclast_epi128( a0, b0 );\
+  a1 = _mm256_aesenclast_epi128( a1, b0 );\
+  a2 = _mm256_aesenclast_epi128( a2, b0 );\
+  a3 = _mm256_aesenclast_epi128( a3, b0 );\
+  a4 = _mm256_aesenclast_epi128( a4, b0 );\
+  a5 = _mm256_aesenclast_epi128( a5, b0 );\
+  a6 = _mm256_aesenclast_epi128( a6, b0 );\
+  a7 = _mm256_aesenclast_epi128( a7, b0 );\
+  /* MixBytes */\
+  MixBytes_2way(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7);\
+}
+
+#define ROUNDS_P_2WAY(){\
+  uint8_t round_counter = 0;\
+  for ( round_counter = 0; round_counter < 14; round_counter += 2 ) \
+  { \
+    /* AddRoundConstant P1024 */\
+    xmm8 = _mm256_xor_si256( xmm8, m256_const1_128( \
+             casti_m128i( round_const_p, round_counter ) ) ); \
+    /* ShiftBytes P1024 + pre-AESENCLAST */\
+    xmm8  = _mm256_shuffle_epi8( xmm8,  SUBSH_MASK0_2WAY ); \
+    xmm9  = _mm256_shuffle_epi8( xmm9,  SUBSH_MASK1_2WAY );\
+    xmm10 = _mm256_shuffle_epi8( xmm10, SUBSH_MASK2_2WAY );\
+    xmm11 = _mm256_shuffle_epi8( xmm11, SUBSH_MASK3_2WAY );\
+    xmm12 = _mm256_shuffle_epi8( xmm12, SUBSH_MASK4_2WAY );\
+    xmm13 = _mm256_shuffle_epi8( xmm13, SUBSH_MASK5_2WAY );\
+    xmm14 = _mm256_shuffle_epi8( xmm14, SUBSH_MASK6_2WAY );\
+    xmm15 = _mm256_shuffle_epi8( xmm15, SUBSH_MASK7_2WAY );\
+    /* SubBytes + MixBytes */\
+    SUBMIX_2WAY(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+    \
+     /* AddRoundConstant P1024 */\
+    xmm0 = _mm256_xor_si256( xmm0, m256_const1_128( \
+             casti_m128i( round_const_p, round_counter+1 ) ) ); \
+    /* ShiftBytes P1024 + pre-AESENCLAST */\
+    xmm0 = _mm256_shuffle_epi8( xmm0, SUBSH_MASK0_2WAY );\
+    xmm1 = _mm256_shuffle_epi8( xmm1, SUBSH_MASK1_2WAY );\
+    xmm2 = _mm256_shuffle_epi8( xmm2, SUBSH_MASK2_2WAY );\
+    xmm3 = _mm256_shuffle_epi8( xmm3, SUBSH_MASK3_2WAY );\
+    xmm4 = _mm256_shuffle_epi8( xmm4, SUBSH_MASK4_2WAY );\
+    xmm5 = _mm256_shuffle_epi8( xmm5, SUBSH_MASK5_2WAY );\
+    xmm6 = _mm256_shuffle_epi8( xmm6, SUBSH_MASK6_2WAY );\
+    xmm7 = _mm256_shuffle_epi8( xmm7, SUBSH_MASK7_2WAY );\
+    /* SubBytes + MixBytes */\
+     SUBMIX_2WAY(xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  }\
+}
+
+#define ROUNDS_Q_2WAY(){\
+  uint8_t round_counter = 0;\
+  for ( round_counter = 0; round_counter < 14; round_counter += 2) \
+  { \
+    /* AddRoundConstant Q1024 */\
+    xmm1 = m256_neg1;\
+    xmm8  = _mm256_xor_si256( xmm8,  xmm1 );\
+    xmm9  = _mm256_xor_si256( xmm9,  xmm1 );\
+    xmm10 = _mm256_xor_si256( xmm10, xmm1 );\
+    xmm11 = _mm256_xor_si256( xmm11, xmm1 );\
+    xmm12 = _mm256_xor_si256( xmm12, xmm1 );\
+    xmm13 = _mm256_xor_si256( xmm13, xmm1 );\
+    xmm14 = _mm256_xor_si256( xmm14, xmm1 );\
+    xmm15 = _mm256_xor_si256( xmm15, m256_const1_128( \
+                 casti_m128i( round_const_q, round_counter ) ) ); \
+    /* ShiftBytes Q1024 + pre-AESENCLAST */\
+    xmm8  = _mm256_shuffle_epi8( xmm8,  SUBSH_MASK1_2WAY );\
+    xmm9  = _mm256_shuffle_epi8( xmm9,  SUBSH_MASK3_2WAY );\
+    xmm10 = _mm256_shuffle_epi8( xmm10, SUBSH_MASK5_2WAY );\
+    xmm11 = _mm256_shuffle_epi8( xmm11, SUBSH_MASK7_2WAY );\
+    xmm12 = _mm256_shuffle_epi8( xmm12, SUBSH_MASK0_2WAY );\
+    xmm13 = _mm256_shuffle_epi8( xmm13, SUBSH_MASK2_2WAY );\
+    xmm14 = _mm256_shuffle_epi8( xmm14, SUBSH_MASK4_2WAY );\
+    xmm15 = _mm256_shuffle_epi8( xmm15, SUBSH_MASK6_2WAY );\
+    /* SubBytes + MixBytes */\
+    SUBMIX_2WAY(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
+    \
+    /* AddRoundConstant Q1024 */\
+    xmm9 = m256_neg1;\
+    xmm0 = _mm256_xor_si256( xmm0, xmm9 );\
+    xmm1 = _mm256_xor_si256( xmm1, xmm9 );\
+    xmm2 = _mm256_xor_si256( xmm2, xmm9 );\
+    xmm3 = _mm256_xor_si256( xmm3, xmm9 );\
+    xmm4 = _mm256_xor_si256( xmm4, xmm9 );\
+    xmm5 = _mm256_xor_si256( xmm5, xmm9 );\
+    xmm6 = _mm256_xor_si256( xmm6, xmm9 );\
+    xmm7 = _mm256_xor_si256( xmm7, m256_const1_128( \
+             casti_m128i( round_const_q, round_counter+1 ) ) ); \
+    /* ShiftBytes Q1024 + pre-AESENCLAST */\
+    xmm0 = _mm256_shuffle_epi8( xmm0, SUBSH_MASK1_2WAY );\
+    xmm1 = _mm256_shuffle_epi8( xmm1, SUBSH_MASK3_2WAY );\
+    xmm2 = _mm256_shuffle_epi8( xmm2, SUBSH_MASK5_2WAY );\
+    xmm3 = _mm256_shuffle_epi8( xmm3, SUBSH_MASK7_2WAY );\
+    xmm4 = _mm256_shuffle_epi8( xmm4, SUBSH_MASK0_2WAY );\
+    xmm5 = _mm256_shuffle_epi8( xmm5, SUBSH_MASK2_2WAY );\
+    xmm6 = _mm256_shuffle_epi8( xmm6, SUBSH_MASK4_2WAY );\
+    xmm7 = _mm256_shuffle_epi8( xmm7, SUBSH_MASK6_2WAY );\
+    /* SubBytes + MixBytes */\
+    SUBMIX_2WAY(xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7, xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15);\
+  }\
+}
+
+#define Matrix_Transpose_2way(i0, i1, i2, i3, i4, i5, i6, i7, t0, t1, t2, t3, t4, t5, t6, t7){\
+  t0 = TRANSP_MASK_2WAY;\
+\
+  i6 = _mm256_shuffle_epi8(i6, t0);\
+  i0 = _mm256_shuffle_epi8(i0, t0);\
+  i1 = _mm256_shuffle_epi8(i1, t0);\
+  i2 = _mm256_shuffle_epi8(i2, t0);\
+  i3 = _mm256_shuffle_epi8(i3, t0);\
+  t1 = i2;\
+  i4 = _mm256_shuffle_epi8(i4, t0);\
+  i5 = _mm256_shuffle_epi8(i5, t0);\
+  t2 = i4;\
+  t3 = i6;\
+  i7 = _mm256_shuffle_epi8(i7, t0);\
+\
+  /* continue with unpack using 4 temp registers */\
+  t0 = i0;\
+  t2 = _mm256_unpackhi_epi16(t2, i5);\
+  i4 = _mm256_unpacklo_epi16(i4, i5);\
+  t3 = _mm256_unpackhi_epi16(t3, i7);\
+  i6 = _mm256_unpacklo_epi16(i6, i7);\
+  t0 = _mm256_unpackhi_epi16(t0, i1);\
+  t1 = _mm256_unpackhi_epi16(t1, i3);\
+  i2 = _mm256_unpacklo_epi16(i2, i3);\
+  i0 = _mm256_unpacklo_epi16(i0, i1);\
+\
+  /* shuffle with immediate */\
+  t0 = _mm256_shuffle_epi32(t0, 216);\
+  t1 = _mm256_shuffle_epi32(t1, 216);\
+  t2 = _mm256_shuffle_epi32(t2, 216);\
+  t3 = _mm256_shuffle_epi32(t3, 216);\
+  i0 = _mm256_shuffle_epi32(i0, 216);\
+  i2 = _mm256_shuffle_epi32(i2, 216);\
+  i4 = _mm256_shuffle_epi32(i4, 216);\
+  i6 = _mm256_shuffle_epi32(i6, 216);\
+\
+  /* continue with unpack */\
+  t4 = i0;\
+  i0 = _mm256_unpacklo_epi32(i0, i2);\
+  t4 = _mm256_unpackhi_epi32(t4, i2);\
+  t5 = t0;\
+  t0 = _mm256_unpacklo_epi32(t0, t1);\
+  t5 = _mm256_unpackhi_epi32(t5, t1);\
+  t6 = i4;\
+  i4 = _mm256_unpacklo_epi32(i4, i6);\
+  t7 = t2;\
+  t6 = _mm256_unpackhi_epi32(t6, i6);\
+  i2 = t0;\
+  t2 = _mm256_unpacklo_epi32(t2, t3);\
+  i3 = t0;\
+  t7 = _mm256_unpackhi_epi32(t7, t3);\
+\
+  /* there are now 2 rows in each xmm */\
+  /* unpack to get 1 row of CV in each xmm */\
+  i1 = i0;\
+  i1 = _mm256_unpackhi_epi64(i1, i4);\
+  i0 = _mm256_unpacklo_epi64(i0, i4);\
+  i4 = t4;\
+  i3 = _mm256_unpackhi_epi64(i3, t2);\
+  i5 = t4;\
+  i2 = _mm256_unpacklo_epi64(i2, t2);\
+  i6 = t5;\
+  i5 = _mm256_unpackhi_epi64(i5, t6);\
+  i7 = t5;\
+  i4 = _mm256_unpacklo_epi64(i4, t6);\
+  i7 = _mm256_unpackhi_epi64(i7, t7);\
+  i6 = _mm256_unpacklo_epi64(i6, t7);\
+  /* transpose done */\
+}/**/
+
+#define Matrix_Transpose_INV_2way(i0, i1, i2, i3, i4, i5, i6, i7, o0, o1, o2, t0, t1, t2, t3, t4){\
+  /*  transpose matrix to get output format */\
+  o1 = i0;\
+  i0 = _mm256_unpacklo_epi64(i0, i1);\
+  o1 = _mm256_unpackhi_epi64(o1, i1);\
+  t0 = i2;\
+  i2 = _mm256_unpacklo_epi64(i2, i3);\
+  t0 = _mm256_unpackhi_epi64(t0, i3);\
+  t1 = i4;\
+  i4 = _mm256_unpacklo_epi64(i4, i5);\
+  t1 = _mm256_unpackhi_epi64(t1, i5);\
+  t2 = i6;\
+  o0 = TRANSP_MASK_2WAY;\
+  i6 = _mm256_unpacklo_epi64(i6, i7);\
+  t2 = _mm256_unpackhi_epi64(t2, i7);\
+  /* load transpose mask into a register, because it will be used 8 times */\
+  i0 = _mm256_shuffle_epi8(i0, o0);\
+  i2 = _mm256_shuffle_epi8(i2, o0);\
+  i4 = _mm256_shuffle_epi8(i4, o0);\
+  i6 = _mm256_shuffle_epi8(i6, o0);\
+  o1 = _mm256_shuffle_epi8(o1, o0);\
+  t0 = _mm256_shuffle_epi8(t0, o0);\
+  t1 = _mm256_shuffle_epi8(t1, o0);\
+  t2 = _mm256_shuffle_epi8(t2, o0);\
+  /* continue with unpack using 4 temp registers */\
+  t3 = i4;\
+  o2 = o1;\
+  o0 = i0;\
+  t4 = t1;\
+  \
+  t3 = _mm256_unpackhi_epi16(t3, i6);\
+  i4 = _mm256_unpacklo_epi16(i4, i6);\
+  o0 = _mm256_unpackhi_epi16(o0, i2);\
+  i0 = _mm256_unpacklo_epi16(i0, i2);\
+  o2 = _mm256_unpackhi_epi16(o2, t0);\
+  o1 = _mm256_unpacklo_epi16(o1, t0);\
+  t4 = _mm256_unpackhi_epi16(t4, t2);\
+  t1 = _mm256_unpacklo_epi16(t1, t2);\
+  /* shuffle with immediate */\
+  i4 = _mm256_shuffle_epi32(i4, 216);\
+  t3 = _mm256_shuffle_epi32(t3, 216);\
+  o1 = _mm256_shuffle_epi32(o1, 216);\
+  o2 = _mm256_shuffle_epi32(o2, 216);\
+  i0 = _mm256_shuffle_epi32(i0, 216);\
+  o0 = _mm256_shuffle_epi32(o0, 216);\
+  t1 = _mm256_shuffle_epi32(t1, 216);\
+  t4 = _mm256_shuffle_epi32(t4, 216);\
+  /* continue with unpack */\
+  i1 = i0;\
+  i3 = o0;\
+  i5 = o1;\
+  i7 = o2;\
+  i0 = _mm256_unpacklo_epi32(i0, i4);\
+  i1 = _mm256_unpackhi_epi32(i1, i4);\
+  o0 = _mm256_unpacklo_epi32(o0, t3);\
+  i3 = _mm256_unpackhi_epi32(i3, t3);\
+  o1 = _mm256_unpacklo_epi32(o1, t1);\
+  i5 = _mm256_unpackhi_epi32(i5, t1);\
+  o2 = _mm256_unpacklo_epi32(o2, t4);\
+  i7 = _mm256_unpackhi_epi32(i7, t4);\
+  /* transpose done */\
+}/**/
+
+void INIT_2way( __m256i *chaining )
+{
+  static __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  static __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+
+  /* load IV into registers xmm8 - xmm15 */
+  xmm8 = chaining[0];
+  xmm9 = chaining[1];
+  xmm10 = chaining[2];
+  xmm11 = chaining[3];
+  xmm12 = chaining[4];
+  xmm13 = chaining[5];
+  xmm14 = chaining[6];
+  xmm15 = chaining[7];
+
+  /* transform chaining value from column ordering into row ordering */
+  Matrix_Transpose_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);
+
+  /* store transposed IV */
+  chaining[0] = xmm8;
+  chaining[1] = xmm9;
+  chaining[2] = xmm10;
+  chaining[3] = xmm11;
+  chaining[4] = xmm12;
+  chaining[5] = xmm13;
+  chaining[6] = xmm14;
+  chaining[7] = xmm15;
+}
+
+void TF1024_2way( __m256i *chaining, const __m256i *message )
+{
+  static __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  static __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  static __m256i QTEMP[8];
+  static __m256i TEMP0;
+  static __m256i TEMP1;
+  static __m256i TEMP2;
+
+  /* load message into registers xmm8 - xmm15 (Q = message) */
+  xmm8 = message[0];
+  xmm9 = message[1];
+  xmm10 = message[2];
+  xmm11 = message[3];
+  xmm12 = message[4];
+  xmm13 = message[5];
+  xmm14 = message[6];
+  xmm15 = message[7];
+
+  /* transform message M from column ordering into row ordering */
+  Matrix_Transpose_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);
+
+  /* store message M (Q input) for later */
+  QTEMP[0] = xmm8;
+  QTEMP[1] = xmm9;
+  QTEMP[2] = xmm10;
+  QTEMP[3] = xmm11;
+  QTEMP[4] = xmm12;
+  QTEMP[5] = xmm13;
+  QTEMP[6] = xmm14;
+  QTEMP[7] = xmm15;
+
+  /* xor CV to message to get P input */
+  /* result: CV+M in xmm8...xmm15 */
+  xmm8 = _mm256_xor_si256( xmm8,  (chaining[0]) );
+  xmm9 = _mm256_xor_si256( xmm9,  (chaining[1]) );
+  xmm10 = _mm256_xor_si256( xmm10, (chaining[2]) );
+  xmm11 = _mm256_xor_si256( xmm11, (chaining[3]) );
+  xmm12 = _mm256_xor_si256( xmm12, (chaining[4]) );
+  xmm13 = _mm256_xor_si256( xmm13, (chaining[5]) );
+  xmm14 = _mm256_xor_si256( xmm14, (chaining[6]) );
+  xmm15 = _mm256_xor_si256( xmm15, (chaining[7]) );
+
+  /* compute permutation P */
+  /* result: P(CV+M) in xmm8...xmm15 */
+  ROUNDS_P_2WAY();
+
+  /* xor CV to P output (feed-forward) */
+  /* result: P(CV+M)+CV in xmm8...xmm15 */
+  xmm8 = _mm256_xor_si256( xmm8,  (chaining[0]) );
+  xmm9 = _mm256_xor_si256( xmm9,  (chaining[1]) );
+  xmm10 = _mm256_xor_si256( xmm10, (chaining[2]) );
+  xmm11 = _mm256_xor_si256( xmm11, (chaining[3]) );
+  xmm12 = _mm256_xor_si256( xmm12, (chaining[4]) );
+  xmm13 = _mm256_xor_si256( xmm13, (chaining[5]) );
+  xmm14 = _mm256_xor_si256( xmm14, (chaining[6]) );
+  xmm15 = _mm256_xor_si256( xmm15, (chaining[7]) );
+
+  /* store P(CV+M)+CV */
+  chaining[0] = xmm8;
+  chaining[1] = xmm9;
+  chaining[2] = xmm10;
+  chaining[3] = xmm11;
+  chaining[4] = xmm12;
+  chaining[5] = xmm13;
+  chaining[6] = xmm14;
+  chaining[7] = xmm15;
+
+  /* load message M (Q input) into xmm8-15 */
+  xmm8 = QTEMP[0];
+  xmm9 = QTEMP[1];
+  xmm10 = QTEMP[2];
+  xmm11 = QTEMP[3];
+  xmm12 = QTEMP[4];
+  xmm13 = QTEMP[5];
+  xmm14 = QTEMP[6];
+  xmm15 = QTEMP[7];
+
+  /* compute permutation Q */
+  /* result: Q(M) in xmm8...xmm15 */
+  ROUNDS_Q_2WAY();
+
+  /* xor Q output */
+  /* result: P(CV+M)+CV+Q(M) in xmm8...xmm15 */
+  xmm8 = _mm256_xor_si256( xmm8,  (chaining[0]) );
+  xmm9 = _mm256_xor_si256( xmm9,  (chaining[1]) );
+  xmm10 = _mm256_xor_si256( xmm10, (chaining[2]) );
+  xmm11 = _mm256_xor_si256( xmm11, (chaining[3]) );
+  xmm12 = _mm256_xor_si256( xmm12, (chaining[4]) );
+  xmm13 = _mm256_xor_si256( xmm13, (chaining[5]) );
+  xmm14 = _mm256_xor_si256( xmm14, (chaining[6]) );
+  xmm15 = _mm256_xor_si256( xmm15, (chaining[7]) );
+
+  /* store CV */
+  chaining[0] = xmm8;
+  chaining[1] = xmm9;
+  chaining[2] = xmm10;
+  chaining[3] = xmm11;
+  chaining[4] = xmm12;
+  chaining[5] = xmm13;
+  chaining[6] = xmm14;
+  chaining[7] = xmm15;
+
+  return;
+}
+
+void OF1024_2way( __m256i* chaining )
+{
+  static __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  static __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  static __m256i TEMP0;
+  static __m256i TEMP1;
+  static __m256i TEMP2;
+
+  /* load CV into registers xmm8 - xmm15 */
+  xmm8 = chaining[0];
+  xmm9 = chaining[1];
+  xmm10 = chaining[2];
+  xmm11 = chaining[3];
+  xmm12 = chaining[4];
+  xmm13 = chaining[5];
+  xmm14 = chaining[6];
+  xmm15 = chaining[7];
+
+  /* compute permutation P */
+  /* result: P(CV) in xmm8...xmm15 */
+  ROUNDS_P_2WAY();
+
+  /* xor CV to P output (feed-forward) */
+  /* result: P(CV)+CV in xmm8...xmm15 */
+  xmm8 = _mm256_xor_si256( xmm8,  (chaining[0]) );
+  xmm9 = _mm256_xor_si256( xmm9,  (chaining[1]) );
+  xmm10 = _mm256_xor_si256( xmm10, (chaining[2]) );
+  xmm11 = _mm256_xor_si256( xmm11, (chaining[3]) );
+  xmm12 = _mm256_xor_si256( xmm12, (chaining[4]) );
+  xmm13 = _mm256_xor_si256( xmm13, (chaining[5]) );
+  xmm14 = _mm256_xor_si256( xmm14, (chaining[6]) );
+  xmm15 = _mm256_xor_si256( xmm15, (chaining[7]) );
+
+  /* transpose CV back from row ordering to column ordering */
+  /* result: final hash value in xmm0, xmm6, xmm13, xmm15 */
+  Matrix_Transpose_INV_2way(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm4, xmm0, xmm6, xmm1, xmm2, xmm3, xmm5, xmm7);
+
+  /* we only need to return the truncated half of the state */
+  chaining[4] = xmm0;
+  chaining[5] = xmm6;
+  chaining[6] = xmm13;
+  chaining[7] = xmm15;
+
+  return;
+}
+
+
+
 #endif  // VAES
 #endif  // GROESTL512_INTR_4WAY_H__
--- a/algo/groestl/myr-groestl.c
+++ b/algo/groestl/myr-groestl.c
@@ -11,7 +11,7 @@
 #else
  #include "sph_groestl.h"
 #endif
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"

 typedef struct {
 #ifdef __AES__
@@ -19,7 +19,7 @@ typedef struct {
 #else
    sph_groestl512_context  groestl;
 #endif
-    SHA256_CTX              sha;
+    sph_sha256_context      sha;
 } myrgr_ctx_holder;

 myrgr_ctx_holder myrgr_ctx;
@@ -31,7 +31,7 @@ void init_myrgr_ctx()
 #else
     sph_groestl512_init( &myrgr_ctx.groestl );
 #endif
-     SHA256_Init( &myrgr_ctx.sha );
+     sph_sha256_init( &myrgr_ctx.sha );
 }

 void myriad_hash(void *output, const void *input)
@@ -39,54 +39,55 @@ void myriad_hash(void *output, const void *input)
   myrgr_ctx_holder ctx;
   memcpy( &ctx, &myrgr_ctx, sizeof(myrgr_ctx) );

- 	uint32_t _ALIGN(32) hash[16];
+   uint32_t _ALIGN(32) hash[16];

 #ifdef __AES__
   update_groestl( &ctx.groestl, (char*)input, 640 );
   final_groestl( &ctx.groestl, (char*)hash);
 #else
-	sph_groestl512(&ctx.groestl, input, 80);
-	sph_groestl512_close(&ctx.groestl, hash);
+   sph_groestl512(&ctx.groestl, input, 80);
+   sph_groestl512_close(&ctx.groestl, hash);
 #endif

-   SHA256_Update( &ctx.sha, (unsigned char*)hash, 64 );
-   SHA256_Final( (unsigned char*)hash, &ctx.sha );
+   sph_sha256( &ctx.sha, hash, 64 );
+   sph_sha256_close( &ctx.sha, hash );

-	memcpy(output, hash, 32);
+   memcpy(output, hash, 32);
 }

 int scanhash_myriad( struct work *work, uint32_t max_nonce,
                     uint64_t *hashes_done, struct thr_info *mythr )
 {
-	uint32_t _ALIGN(64) endiandata[20];
+   uint32_t _ALIGN(64) endiandata[20];
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
-	const uint32_t first_nonce = pdata[19];
-	uint32_t nonce = first_nonce;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   const uint32_t first_nonce = pdata[19];
+   uint32_t nonce = first_nonce;
+   int thr_id = mythr->id;

-	if (opt_benchmark)
-		((uint32_t*)ptarget)[7] = 0x0000ff;
+   if (opt_benchmark)
+      ((uint32_t*)ptarget)[7] = 0x0000ff;

-        swab32_array( endiandata, pdata, 20 );
+   swab32_array( endiandata, pdata, 20 );

-	do {
-		const uint32_t Htarg = ptarget[7];
-		uint32_t hash[8];
-		be32enc(&endiandata[19], nonce);
-		myriad_hash(hash, endiandata);
+   do {
+      const uint32_t Htarg = ptarget[7];
+      uint32_t hash[8];
+      be32enc(&endiandata[19], nonce);
+      myriad_hash(hash, endiandata);

-		if (hash[7] <= Htarg && fulltest(hash, ptarget)) {
-			pdata[19] = nonce;
-			*hashes_done = pdata[19] - first_nonce;
-			return 1;
-		}
-		nonce++;
+      if (hash[7] <= Htarg && fulltest(hash, ptarget))
+      {
+         pdata[19] = nonce;
+         *hashes_done = pdata[19] - first_nonce;
+         return 1;
+      }
+      nonce++;

-	} while (nonce < max_nonce && !work_restart[thr_id].restart);
+   } while (nonce < max_nonce && !work_restart[thr_id].restart);

-	pdata[19] = nonce;
-	*hashes_done = pdata[19] - first_nonce + 1;
-	return 0;
+   pdata[19] = nonce;
+   *hashes_done = pdata[19] - first_nonce + 1;
+   return 0;
 }
 #endif
--- a/algo/hamsi/hamsi-hash-4way.c
+++ b/algo/hamsi/hamsi-hash-4way.c
@@ -548,7 +548,7 @@ static const sph_u32 T512[64][16] = {

 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

-// Hamsi 8 way 
+// Hamsi 8 way AVX512 

 #define INPUT_BIG8 \
 do { \
@@ -849,13 +849,11 @@ void hamsi512_8way_update( hamsi_8way_big_context *sc, const void *data,
 void hamsi512_8way_close( hamsi_8way_big_context *sc, void *dst )
 {
   __m512i pad[1];
-   int ch, cl;
+   uint32_t ch, cl;

   sph_enc32be( &ch, sc->count_high );
   sph_enc32be( &cl, sc->count_low + ( sc->partial_len << 3 ) );
-   pad[0] =  _mm512_set_epi32( cl, ch, cl, ch, cl, ch, cl, ch,
-                               cl, ch, cl, ch, cl, ch, cl, ch );
-//   pad[0] =  m512_const2_32( cl, ch );
+   pad[0] = _mm512_set1_epi64( ((uint64_t)cl << 32 ) | (uint64_t)ch );
   sc->buf[0] = m512_const1_64( 0x80 );
   hamsi_8way_big( sc, sc->buf, 1 );
   hamsi_8way_big_final( sc, pad );
@@ -863,11 +861,9 @@ void hamsi512_8way_close( hamsi_8way_big_context *sc, void *dst )
   mm512_block_bswap_32( (__m512i*)dst, sc->h );
 }

-
 #endif // AVX512

-
-// Hamsi 4 way
+// Hamsi 4 way AVX2

 #define INPUT_BIG \
 do { \
@@ -1186,14 +1182,12 @@ void hamsi512_4way_update( hamsi_4way_big_context *sc, const void *data,
 void hamsi512_4way_close( hamsi_4way_big_context *sc, void *dst )
 {
   __m256i pad[1];
-   int ch, cl;
+   uint32_t ch, cl;

   sph_enc32be( &ch, sc->count_high );
   sph_enc32be( &cl, sc->count_low + ( sc->partial_len << 3 ) );
-   pad[0] =  _mm256_set_epi32( cl, ch, cl, ch, cl, ch, cl, ch );
+   pad[0] = _mm256_set1_epi64x( ((uint64_t)cl << 32 ) | (uint64_t)ch );
   sc->buf[0] = m256_const1_64( 0x80 );
-//      sc->buf[0] = _mm256_set_epi32( 0UL, 0x80UL, 0UL, 0x80UL,
-//                                  0UL, 0x80UL, 0UL, 0x80UL );
   hamsi_big( sc, sc->buf, 1 );
   hamsi_big_final( sc, pad );

--- a/algo/hodl/hodl-gate.c
+++ b/algo/hodl/hodl-gate.c
@@ -99,13 +99,13 @@ void hodl_build_block_header( struct work* g_work, uint32_t version,
 // called only by thread 0, saves a backup of g_work
 void hodl_get_new_work( struct work* work, struct work* g_work)
 {
-   pthread_rwlock_rdlock( &g_work_lock );
+//   pthread_rwlock_rdlock( &g_work_lock );

   work_free( &hodl_work );
   work_copy( &hodl_work, g_work );
   hodl_work.data[ algo_gate.nonce_index ] = ( clock() + rand() ) % 9999;

-   pthread_rwlock_unlock( &g_work_lock );
+//   pthread_rwlock_unlock( &g_work_lock );
 }

 json_t *hodl_longpoll_rpc_call( CURL *curl, int *err, char* lp_url )
@@ -125,7 +125,7 @@ json_t *hodl_longpoll_rpc_call( CURL *curl, int *err, char* lp_url )
 }

 // called by every thread, copies the backup to each thread's work.
-void hodl_resync_threads( struct work* work )
+void hodl_resync_threads( int thr_id, struct work* work )
 {
   int nonce_index = algo_gate.nonce_index;
   pthread_barrier_wait( &hodl_barrier );
@@ -135,6 +135,7 @@ void hodl_resync_threads( struct work* work )
      work_copy( work, &hodl_work );
   }
   work->data[ nonce_index ] = swab32( hodl_work.data[ nonce_index ] );
+   work_restart[thr_id].restart = 0;
 }

 bool hodl_do_this_thread( int thr_id )
--- a/algo/keccak/keccak-macros.c
+++ b/algo/keccak/keccak-macros.c
@@ -134,65 +134,47 @@
        do { \
                DECL64(c0); \
                DECL64(c1); \
-                DECL64(c2); \
-                DECL64(c3); \
-                DECL64(c4); \
                DECL64(bnn); \
                NOT64(bnn, b20); \
                KHI_XO(c0, b00, b10, b20); \
                KHI_XO(c1, b10, bnn, b30); \
-                KHI_XA(c2, b20, b30, b40); \
-                KHI_XO(c3, b30, b40, b00); \
-                KHI_XA(c4, b40, b00, b10); \
+                KHI_XA(b20, b20, b30, b40); \
+                KHI_XO(b30, b30, b40, b00); \
+                KHI_XA(b40, b40, b00, b10); \
                MOV64(b00, c0); \
                MOV64(b10, c1); \
-                MOV64(b20, c2); \
-                MOV64(b30, c3); \
-                MOV64(b40, c4); \
                NOT64(bnn, b41); \
                KHI_XO(c0, b01, b11, b21); \
                KHI_XA(c1, b11, b21, b31); \
-                KHI_XO(c2, b21, b31, bnn); \
-                KHI_XO(c3, b31, b41, b01); \
-                KHI_XA(c4, b41, b01, b11); \
+                KHI_XO(b21, b21, b31, bnn); \
+                KHI_XO(b31, b31, b41, b01); \
+                KHI_XA(b41, b41, b01, b11); \
                MOV64(b01, c0); \
                MOV64(b11, c1); \
-                MOV64(b21, c2); \
-                MOV64(b31, c3); \
-                MOV64(b41, c4); \
                NOT64(bnn, b32); \
                KHI_XO(c0, b02, b12, b22); \
                KHI_XA(c1, b12, b22, b32); \
-                KHI_XA(c2, b22, bnn, b42); \
-                KHI_XO(c3, bnn, b42, b02); \
-                KHI_XA(c4, b42, b02, b12); \
+                KHI_XA(b22, b22, bnn, b42); \
+                KHI_XO(b32, bnn, b42, b02); \
+                KHI_XA(b42, b42, b02, b12); \
                MOV64(b02, c0); \
                MOV64(b12, c1); \
-                MOV64(b22, c2); \
-                MOV64(b32, c3); \
-                MOV64(b42, c4); \
                NOT64(bnn, b33); \
                KHI_XA(c0, b03, b13, b23); \
                KHI_XO(c1, b13, b23, b33); \
-                KHI_XO(c2, b23, bnn, b43); \
-                KHI_XA(c3, bnn, b43, b03); \
-                KHI_XO(c4, b43, b03, b13); \
+                KHI_XO(b23, b23, bnn, b43); \
+                KHI_XA(b33, bnn, b43, b03); \
+                KHI_XO(b43, b43, b03, b13); \
                MOV64(b03, c0); \
                MOV64(b13, c1); \
-                MOV64(b23, c2); \
-                MOV64(b33, c3); \
-                MOV64(b43, c4); \
                NOT64(bnn, b14); \
                KHI_XA(c0, b04, bnn, b24); \
                KHI_XO(c1, bnn, b24, b34); \
-                KHI_XA(c2, b24, b34, b44); \
-                KHI_XO(c3, b34, b44, b04); \
-                KHI_XA(c4, b44, b04, b14); \
+                KHI_XA(b24, b24, b34, b44); \
+                KHI_XO(b34, b34, b44, b04); \
+                KHI_XA(b44, b44, b04, b14); \
                MOV64(b04, c0); \
                MOV64(b14, c1); \
-                MOV64(b24, c2); \
-                MOV64(b34, c3); \
-                MOV64(b44, c4); \
        } while (0)

 #ifdef IOTA
@@ -201,6 +183,7 @@
 #define IOTA(r)   XOR64_IOTA(a00, a00, r)

 #ifdef P0
+#undef P0
 #undef P1
 #undef P2
 #undef P3
--- a/algo/luffa/luffa-hash-2way.c
+++ b/algo/luffa/luffa-hash-2way.c
@@ -66,6 +66,17 @@ static const uint32 CNS_INIT[128] __attribute((aligned(64))) = {
    a = _mm512_xor_si512(a,c0);\
    b = _mm512_xor_si512(b,c1);

+#define MULT24W( a0, a1 ) \
+do { \
+  __m512i b = _mm512_xor_si512( a0, \
+                     _mm512_maskz_shuffle_epi32( 0xbbbb, a1, 16 ) ); \
+  a0 = _mm512_or_si512( _mm512_bsrli_epi128(  b, 4 ), \
+                        _mm512_bslli_epi128( a1,12 ) ); \
+  a1 = _mm512_or_si512( _mm512_bsrli_epi128( a1, 4 ), \
+                        _mm512_bslli_epi128(  b,12 ) ); \
+} while(0)
+
+/*
 #define MULT24W( a0, a1, mask ) \
 do { \
  __m512i b = _mm512_xor_si512( a0, \
@@ -73,6 +84,7 @@ do { \
  a0 = _mm512_or_si512( _mm512_bsrli_epi128(b,4), _mm512_bslli_epi128(a1,12) );\
  a1 = _mm512_or_si512( _mm512_bsrli_epi128(a1,4), _mm512_bslli_epi128(b,12) );\
 } while(0)
+*/

 // confirm pointer arithmetic
 // ok but use array indexes
@@ -235,7 +247,6 @@ void rnd512_4way( luffa_4way_context *state, __m512i *msg )
    __m512i msg0, msg1;
    __m512i tmp[2];
    __m512i x[8];
-    const __m512i MASK = m512_const2_64( 0, 0x00000000ffffffff );

    t0 = chainv[0];
    t1 = chainv[1];
@@ -249,7 +260,7 @@ void rnd512_4way( luffa_4way_context *state, __m512i *msg )
    t0 = _mm512_xor_si512( t0, chainv[8] );
    t1 = _mm512_xor_si512( t1, chainv[9] );

-    MULT24W( t0, t1, MASK );
+    MULT24W( t0, t1 );

    msg0 = _mm512_shuffle_epi32( msg[0], 27 );
    msg1 = _mm512_shuffle_epi32( msg[1], 27 );
@@ -268,68 +279,67 @@ void rnd512_4way( luffa_4way_context *state, __m512i *msg )
    t0 = chainv[0];
    t1 = chainv[1];

-    MULT24W( chainv[0], chainv[1], MASK );
+    MULT24W( chainv[0], chainv[1] );
    chainv[0] = _mm512_xor_si512( chainv[0], chainv[2] );
    chainv[1] = _mm512_xor_si512( chainv[1], chainv[3] );

-    MULT24W( chainv[2], chainv[3], MASK );
+    MULT24W( chainv[2], chainv[3] );
    chainv[2] = _mm512_xor_si512(chainv[2], chainv[4]);
    chainv[3] = _mm512_xor_si512(chainv[3], chainv[5]);

-    MULT24W( chainv[4], chainv[5], MASK );
+    MULT24W( chainv[4], chainv[5] );
    chainv[4] = _mm512_xor_si512(chainv[4], chainv[6]);
    chainv[5] = _mm512_xor_si512(chainv[5], chainv[7]);

-    MULT24W( chainv[6], chainv[7], MASK );
+    MULT24W( chainv[6], chainv[7] );
    chainv[6] = _mm512_xor_si512(chainv[6], chainv[8]);
    chainv[7] = _mm512_xor_si512(chainv[7], chainv[9]);

-    MULT24W( chainv[8], chainv[9], MASK );
+    MULT24W( chainv[8], chainv[9] );
    chainv[8] = _mm512_xor_si512( chainv[8], t0 );
    chainv[9] = _mm512_xor_si512( chainv[9], t1 );

    t0 = chainv[8];
    t1 = chainv[9];

-    MULT24W( chainv[8], chainv[9], MASK );
+    MULT24W( chainv[8], chainv[9] );
    chainv[8] = _mm512_xor_si512( chainv[8], chainv[6] );
    chainv[9] = _mm512_xor_si512( chainv[9], chainv[7] );

-    MULT24W( chainv[6], chainv[7], MASK );
+    MULT24W( chainv[6], chainv[7] );
    chainv[6] = _mm512_xor_si512( chainv[6], chainv[4] );
    chainv[7] = _mm512_xor_si512( chainv[7], chainv[5] );

-    MULT24W( chainv[4], chainv[5], MASK );
+    MULT24W( chainv[4], chainv[5] );
    chainv[4] = _mm512_xor_si512( chainv[4], chainv[2] );
    chainv[5] = _mm512_xor_si512( chainv[5], chainv[3] );

-    MULT24W( chainv[2], chainv[3], MASK );
+    MULT24W( chainv[2], chainv[3] );
    chainv[2] = _mm512_xor_si512( chainv[2], chainv[0] );
    chainv[3] = _mm512_xor_si512( chainv[3], chainv[1] );

-    MULT24W( chainv[0], chainv[1], MASK );
+    MULT24W( chainv[0], chainv[1] );
    chainv[0] = _mm512_xor_si512( _mm512_xor_si512( chainv[0], t0 ), msg0 );
    chainv[1] = _mm512_xor_si512( _mm512_xor_si512( chainv[1], t1 ), msg1 );

-    MULT24W( msg0, msg1, MASK );
+    MULT24W( msg0, msg1 );
    chainv[2] = _mm512_xor_si512( chainv[2], msg0 );
    chainv[3] = _mm512_xor_si512( chainv[3], msg1 );

-    MULT24W( msg0, msg1, MASK );
+    MULT24W( msg0, msg1 );
    chainv[4] = _mm512_xor_si512( chainv[4], msg0 );
    chainv[5] = _mm512_xor_si512( chainv[5], msg1 );

-    MULT24W( msg0, msg1, MASK );
+    MULT24W( msg0, msg1 );
    chainv[6] = _mm512_xor_si512( chainv[6], msg0 );
    chainv[7] = _mm512_xor_si512( chainv[7], msg1 );

-    MULT24W( msg0, msg1, MASK );
+    MULT24W( msg0, msg1);
    chainv[8] = _mm512_xor_si512( chainv[8], msg0 );
    chainv[9] = _mm512_xor_si512( chainv[9], msg1 );

-    MULT24W( msg0, msg1, MASK );
+    MULT24W( msg0, msg1 );

-    // replace with ror
    chainv[3] = _mm512_rol_epi32( chainv[3], 1 );
    chainv[5] = _mm512_rol_epi32( chainv[5], 2 );
    chainv[7] = _mm512_rol_epi32( chainv[7], 3 );
@@ -496,7 +506,7 @@ int luffa_4way_update( luffa_4way_context *state, const void *data,
    {
      // remaining data bytes
      buffer[0] = _mm512_shuffle_epi8( vdata[0], shuff_bswap32 );
-      buffer[1] = m512_const2_64( 0, 0x0000000080000000 );
+      buffer[1] = m512_const1_i128(  0x0000000080000000 );
    }
    return 0;
 }
@@ -520,7 +530,7 @@ int luffa_4way_close( luffa_4way_context *state, void *hashval )
      rnd512_4way( state, buffer );
    else
    {     // empty pad block, constant data
-      msg[0] = m512_const2_64( 0, 0x0000000080000000 );
+      msg[0] = m512_const1_i128(  0x0000000080000000 );
      msg[1] = m512_zero;
      rnd512_4way( state, msg );
    }
@@ -583,13 +593,13 @@ int luffa512_4way_full( luffa_4way_context *state, void *output,
    {
       // padding of partial block
       msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
-       msg[1] = m512_const2_64( 0, 0x0000000080000000 );
+       msg[1] = m512_const1_i128(  0x0000000080000000 );
       rnd512_4way( state, msg );
    }
    else
    {
       // empty pad block
-       msg[0] = m512_const2_64( 0, 0x0000000080000000 );
+       msg[0] = m512_const1_i128( 0x0000000080000000 );
       msg[1] = m512_zero;
       rnd512_4way( state, msg );
    }
@@ -631,13 +641,13 @@ int luffa_4way_update_close( luffa_4way_context *state,
    {
       // padding of partial block
       msg[0] = _mm512_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
-       msg[1] = m512_const2_64( 0, 0x0000000080000000 );
+       msg[1] = m512_const1_i128( 0x0000000080000000 );
       rnd512_4way( state, msg );
    }
    else
    {
       // empty pad block
-       msg[0] = m512_const2_64( 0, 0x0000000080000000 );
+       msg[0] = m512_const1_i128( 0x0000000080000000 );
       msg[1] = m512_zero;
       rnd512_4way( state, msg );
    }
@@ -832,7 +842,7 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
    __m256i msg0, msg1;
    __m256i tmp[2];
    __m256i x[8];
-    const __m256i MASK = m256_const2_64( 0, 0x00000000ffffffff );
+    const __m256i MASK = m256_const1_i128( 0x00000000ffffffff );

    t0 = chainv[0];
    t1 = chainv[1];
@@ -1088,7 +1098,7 @@ int luffa_2way_update( luffa_2way_context *state, const void *data,
    {
      // remaining data bytes
      buffer[0] = _mm256_shuffle_epi8( vdata[0], shuff_bswap32 );
-      buffer[1] = m256_const2_64( 0, 0x0000000080000000 );
+      buffer[1] = m256_const1_i128( 0x0000000080000000 );
    }
    return 0;
 }
@@ -1104,7 +1114,7 @@ int luffa_2way_close( luffa_2way_context *state, void *hashval )
      rnd512_2way( state, buffer );
    else
    {     // empty pad block, constant data
-      msg[0] = m256_const2_64( 0, 0x0000000080000000 );
+      msg[0] = m256_const1_i128( 0x0000000080000000 );
      msg[1] = m256_zero;
      rnd512_2way( state, msg );
    }
@@ -1159,13 +1169,13 @@ int luffa512_2way_full( luffa_2way_context *state, void *output,
    {
       // padding of partial block
       msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
-       msg[1] = m256_const2_64( 0, 0x0000000080000000 );
+       msg[1] = m256_const1_i128( 0x0000000080000000 );
       rnd512_2way( state, msg );
    }
    else
    {
       // empty pad block
-       msg[0] = m256_const2_64( 0, 0x0000000080000000 );
+       msg[0] = m256_const1_i128( 0x0000000080000000 );
       msg[1] = m256_zero;
       rnd512_2way( state, msg );
    }
@@ -1206,13 +1216,13 @@ int luffa_2way_update_close( luffa_2way_context *state,
    {
       // padding of partial block
       msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
-       msg[1] = m256_const2_64( 0, 0x0000000080000000 );
+       msg[1] = m256_const1_i128( 0x0000000080000000 );
       rnd512_2way( state, msg );
    }
    else
    {
       // empty pad block
-       msg[0] = m256_const2_64( 0, 0x0000000080000000 );
+       msg[0] = m256_const1_i128( 0x0000000080000000 );
       msg[1] = m256_zero;
       rnd512_2way( state, msg );
    }
--- a/algo/luffa/luffa_for_sse2.c
+++ b/algo/luffa/luffa_for_sse2.c
@@ -23,7 +23,7 @@
 #include "simd-utils.h"
 #include "luffa_for_sse2.h"

-#define MULT2(a0,a1) do \
+#define MULT2( a0, a1 ) do \
 { \
  __m128i b =  _mm_xor_si128( a0, _mm_shuffle_epi32( _mm_and_si128(a1,MASK), 16 ) ); \
  a0 = _mm_or_si128( _mm_srli_si128(b,4), _mm_slli_si128(a1,12) ); \
@@ -345,11 +345,11 @@ HashReturn update_and_final_luffa( hashState_luffa *state, BitSequence* output,
    // 16 byte partial block exists for 80 byte len
    if ( state->rembytes  )
       // padding of partial block
-       rnd512( state, m128_const_64( 0, 0x80000000 ),
+       rnd512( state, m128_const_i128(  0x80000000 ),
                      mm128_bswap_32( cast_m128i( data ) ) );
    else
       // empty pad block
-       rnd512( state, m128_zero, m128_const_64( 0, 0x80000000 ) );
+       rnd512( state, m128_zero, m128_const_i128( 0x80000000 ) );

    finalization512( state, (uint32*) output );
    if ( state->hashbitlen > 512 )
@@ -394,11 +394,11 @@ int luffa_full( hashState_luffa *state, BitSequence* output, int hashbitlen,
    // 16 byte partial block exists for 80 byte len
    if ( state->rembytes  )
       // padding of partial block
-       rnd512( state, m128_const_64( 0, 0x80000000 ),
+       rnd512( state, m128_const_i128( 0x80000000 ),
                      mm128_bswap_32( cast_m128i( data ) ) );
    else
       // empty pad block
-       rnd512( state, m128_zero, m128_const_64( 0, 0x80000000 ) );
+       rnd512( state, m128_zero, m128_const_i128( 0x80000000 ) );

    finalization512( state, (uint32*) output );
    if ( state->hashbitlen > 512 )
@@ -606,7 +606,6 @@ static void finalization512( hashState_luffa *state, uint32 *b )

    casti_m256i( b, 0 ) = _mm256_shuffle_epi8(
                                 casti_m256i( hash, 0 ), shuff_bswap32 );
-//    casti_m256i( b, 0 ) = mm256_bswap_32( casti_m256i( hash, 0 ) );

    rnd512( state, zero, zero );

@@ -621,7 +620,6 @@ static void finalization512( hashState_luffa *state, uint32 *b )

    casti_m256i( b, 1 ) = _mm256_shuffle_epi8( 
                                 casti_m256i( hash, 0 ), shuff_bswap32 );
-//    casti_m256i( b, 1 ) = mm256_bswap_32( casti_m256i( hash, 0 ) );
 }

 #else
--- a/algo/lyra2/allium-4way.c
+++ b/algo/lyra2/allium-4way.c
@@ -174,24 +174,19 @@ void allium_16way_hash( void *state, const void *input )
 #if defined(__VAES__)

   intrlv_4x128( vhash, hash0, hash1, hash2, hash3, 256 );
-
-   groestl256_4way_full( &ctx.groestl, vhash, vhash, 256 );
-
+   groestl256_4way_full( &ctx.groestl, vhash, vhash, 32 );
   dintrlv_4x128( state, state+32, state+64, state+96, vhash, 256 );
+
   intrlv_4x128( vhash, hash4, hash5, hash6, hash7, 256 );
-
-   groestl256_4way_full( &ctx.groestl, vhash, vhash, 256 );
-   
+   groestl256_4way_full( &ctx.groestl, vhash, vhash, 32 );
   dintrlv_4x128( state+128, state+160, state+192, state+224, vhash, 256 );
+
   intrlv_4x128( vhash, hash8, hash9, hash10, hash11, 256 );
-
-   groestl256_4way_full( &ctx.groestl, vhash, vhash, 256 );
-
+   groestl256_4way_full( &ctx.groestl, vhash, vhash, 32 );
   dintrlv_4x128( state+256, state+288, state+320, state+352, vhash, 256 );
-   intrlv_4x128( vhash, hash12, hash13, hash14, hash15, 256 );

-   groestl256_4way_full( &ctx.groestl, vhash, vhash, 256 );
- 
+   intrlv_4x128( vhash, hash12, hash13, hash14, hash15, 256 );
+   groestl256_4way_full( &ctx.groestl, vhash, vhash, 32 );
   dintrlv_4x128( state+384, state+416, state+448, state+480, vhash, 256 );
   
 #else
@@ -262,8 +257,11 @@ typedef struct {
   keccak256_4way_context    keccak;
   cubehashParam             cube;
   skein256_4way_context     skein;
+#if defined(__VAES__)
+   groestl256_2way_context   groestl;
+#else
   hashState_groestl256      groestl;
-
+#endif
 } allium_8way_ctx_holder;

 static __thread allium_8way_ctx_holder allium_8way_ctx;
@@ -273,7 +271,11 @@ bool init_allium_8way_ctx()
   keccak256_4way_init( &allium_8way_ctx.keccak );
   cubehashInit( &allium_8way_ctx.cube, 256, 16, 32 );
   skein256_4way_init( &allium_8way_ctx.skein );
+#if defined(__VAES__)
+   groestl256_2way_init( &allium_8way_ctx.groestl, 32 );
+#else
   init_groestl256( &allium_8way_ctx.groestl, 32 );
+#endif
   return true;
 }

@@ -352,9 +354,28 @@ void allium_8way_hash( void *hash, const void *input )
   skein256_4way_update( &ctx.skein, vhashB, 32 );
   skein256_4way_close( &ctx.skein, vhashB );

+#if defined(__VAES__)
+
+   uint64_t vhashC[4*2] __attribute__ ((aligned (64)));
+   uint64_t vhashD[4*2] __attribute__ ((aligned (64)));
+   
+   rintrlv_4x64_2x128( vhashC, vhashD, vhashA, 256 );
+   groestl256_2way_full( &ctx.groestl, vhashC, vhashC, 32 );
+   groestl256_2way_full( &ctx.groestl, vhashD, vhashD, 32 );
+   dintrlv_2x128( hash0, hash1, vhashC, 256 );
+   dintrlv_2x128( hash2, hash3, vhashD, 256 );
+
+   rintrlv_4x64_2x128( vhashC, vhashD, vhashB, 256 );
+   groestl256_2way_full( &ctx.groestl, vhashC, vhashC, 32 );
+   groestl256_2way_full( &ctx.groestl, vhashD, vhashD, 32 );
+   dintrlv_2x128( hash4, hash5, vhashC, 256 );
+   dintrlv_2x128( hash6, hash7, vhashD, 256 );
+
+#else
+
   dintrlv_4x64( hash0, hash1, hash2, hash3, vhashA, 256 );
   dintrlv_4x64( hash4, hash5, hash6, hash7, vhashB, 256 );
-
+   
   groestl256_full( &ctx.groestl, hash0, hash0, 256 );
   groestl256_full( &ctx.groestl, hash1, hash1, 256 );
   groestl256_full( &ctx.groestl, hash2, hash2, 256 );
@@ -363,6 +384,8 @@ void allium_8way_hash( void *hash, const void *input )
   groestl256_full( &ctx.groestl, hash5, hash5, 256 );
   groestl256_full( &ctx.groestl, hash6, hash6, 256 );
   groestl256_full( &ctx.groestl, hash7, hash7, 256 );
+
+#endif
 }

 int scanhash_allium_8way( struct work *work, uint32_t max_nonce,
--- a/algo/lyra2/lyra2-gate.c
+++ b/algo/lyra2/lyra2-gate.c
@@ -187,7 +187,8 @@ bool register_allium_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_allium;
  gate->hash      = (void*)&allium_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT
+	              | VAES_OPT | VAES256_OPT;
  opt_target_factor = 256.0;
  return true;
 };
--- a/algo/lyra2/phi2-4way.c
+++ b/algo/lyra2/phi2-4way.c
@@ -4,7 +4,7 @@
 #include "algo/gost/sph_gost.h"
 #include "algo/cubehash/cubehash_sse2.h"
 #include "lyra2.h"
-#if defined(__VAES__)
+#if defined(__VAES__) && defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
  #include "algo/echo/echo-hash-4way.h"
 #elif defined(__AES__)
  #include "algo/echo/aes_ni/hash_api.h"
--- a/algo/m7m/m7m.c
+++ b/algo/m7m/m7m.c
@@ -12,8 +12,7 @@
 #include "algo/tiger/sph_tiger.h"
 #include "algo/whirlpool/sph_whirlpool.h"
 #include "algo/ripemd/sph_ripemd.h"
-#include <openssl/sha.h>
-
+#include "algo/sha/sph_sha2.h"

 #define EPSa DBL_EPSILON
 #define EPS1 DBL_EPSILON
@@ -105,8 +104,8 @@ uint32_t sw2_( int nnounce )
 }

 typedef struct {
-    SHA256_CTX               sha256;
-    SHA512_CTX               sha512;
+    sph_sha256_context      sha256;
+    sph_sha512_context      sha512;
    sph_keccak512_context    keccak;
    sph_whirlpool_context    whirlpool;
    sph_haval256_5_context   haval;
@@ -118,8 +117,8 @@ m7m_ctx_holder m7m_ctx;

 void init_m7m_ctx()
 {
-    SHA256_Init( &m7m_ctx.sha256 );
-    SHA512_Init( &m7m_ctx.sha512 );
+    sph_sha256_init( &m7m_ctx );
+    sph_sha512_init( &m7m_ctx.sha512 );
    sph_keccak512_init( &m7m_ctx.keccak );
    sph_whirlpool_init( &m7m_ctx.whirlpool );
    sph_haval256_5_init( &m7m_ctx.haval );
@@ -143,11 +142,10 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,
    uint32_t hash[8] __attribute__((aligned(64)));
    uint8_t bhash[7][64] __attribute__((aligned(64)));
    uint32_t n = pdata[19] - 1;
-    int thr_id = mythr->id;  // thr_id arg is deprecated
+    int thr_id = mythr->id;
    uint32_t usw_, mpzscale;
    const uint32_t first_nonce = pdata[19];
    char data_str[161], hash_str[65], target_str[65];
-    //uint8_t *bdata = 0;
    uint8_t bdata[8192] __attribute__ ((aligned (64)));
    int i, digits;
    int bytes;
@@ -155,12 +153,12 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,

    m7m_ctx_holder ctx1, ctx2 __attribute__ ((aligned (64)));
    memcpy( &ctx1, &m7m_ctx, sizeof(m7m_ctx) );
-    SHA256_CTX         ctxf_sha256;
+    sph_sha256_context ctxf_sha256;

    memcpy(data, pdata, 80);

-    SHA256_Update(  &ctx1.sha256,    data, M7_MIDSTATE_LEN );
-    SHA512_Update(  &ctx1.sha512,    data, M7_MIDSTATE_LEN );
+    sph_sha256(     &ctx1.sha256,    data, M7_MIDSTATE_LEN );
+    sph_sha512(     &ctx1.sha512,    data, M7_MIDSTATE_LEN );
    sph_keccak512(  &ctx1.keccak,    data, M7_MIDSTATE_LEN );
    sph_whirlpool(  &ctx1.whirlpool, data, M7_MIDSTATE_LEN );
    sph_haval256_5( &ctx1.haval,     data, M7_MIDSTATE_LEN );
@@ -191,11 +189,11 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,

        memcpy( &ctx2, &ctx1, sizeof(m7m_ctx) );

-        SHA256_Update(  &ctx2.sha256, data_p64, 80 - M7_MIDSTATE_LEN );
-        SHA256_Final( (unsigned char*) (bhash[0]), &ctx2.sha256 );
+        sph_sha256( &ctx2.sha256, data_p64, 80 - M7_MIDSTATE_LEN );
+        sph_sha256_close( &ctx2.sha256, bhash[0] );

-        SHA512_Update(  &ctx2.sha512, data_p64, 80 - M7_MIDSTATE_LEN );
-        SHA512_Final( (unsigned char*) (bhash[1]), &ctx2.sha512 );
+        sph_sha512(  &ctx2.sha512, data_p64, 80 - M7_MIDSTATE_LEN );
+        sph_sha512_close( &ctx2.sha512, bhash[1] );

        sph_keccak512( &ctx2.keccak, data_p64, 80 - M7_MIDSTATE_LEN );
        sph_keccak512_close( &ctx2.keccak, (void*)(bhash[2]) );
@@ -227,9 +225,9 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,
        bytes = mpz_sizeinbase(product, 256);
        mpz_export((void *)bdata, NULL, -1, 1, 0, 0, product);

-        SHA256_Init( &ctxf_sha256 );
-        SHA256_Update(  &ctxf_sha256, bdata, bytes );
-        SHA256_Final( (unsigned char*) hash, &ctxf_sha256 );
+        sph_sha256_init( &ctxf_sha256 );
+        sph_sha256( &ctxf_sha256, bdata, bytes );
+        sph_sha256_close( &ctxf_sha256, hash );

        digits=(int)((sqrt((double)(n/2))*(1.+EPS))/9000+75);
        mp_bitcnt_t prec = (long int)(digits*BITS_PER_DIGIT+16);
@@ -262,18 +260,13 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,
            mpzscale=bytes;
            mpz_export(bdata, NULL, -1, 1, 0, 0, product);

-            SHA256_Init( &ctxf_sha256 );
-            SHA256_Update(  &ctxf_sha256, bdata, bytes );
-            SHA256_Final( (unsigned char*) hash, &ctxf_sha256 );
-        }
-
+            sph_sha256_init( &ctxf_sha256 );
+            sph_sha256( &ctxf_sha256, bdata, bytes );
+            sph_sha256_close( &ctxf_sha256, hash );
+	}

        if ( unlikely( valid_hash( (uint64_t*)hash, (uint64_t*)ptarget ) 
             && !opt_benchmark ) )
-
-
-//        if ( unlikely( hash[7] <= ptarget[7] ) )
-//        if ( likely( fulltest( hash, ptarget ) && !opt_benchmark ) )        
        {
           if ( opt_debug )
           {
--- a/algo/nist5/zr5.c
+++ b/algo/nist5/zr5.c
@@ -156,7 +156,7 @@ int scanhash_zr5( struct work *work, uint32_t max_nonce,
 void zr5_get_new_work( struct work* work, struct work* g_work, int thr_id,
                       uint32_t* end_nonce_ptr )
 {
-   pthread_rwlock_rdlock( &g_work_lock );
+//   pthread_rwlock_rdlock( &g_work_lock );

   // ignore POK in first word
   const int wkcmp_sz = 72;  // (19-1) * sizeof(uint32_t)
@@ -174,7 +174,7 @@ void zr5_get_new_work( struct work* work, struct work* g_work, int thr_id,
   else
       ++(*nonceptr);

-   pthread_rwlock_unlock( &g_work_lock );
+//   pthread_rwlock_unlock( &g_work_lock );
 }

 void zr5_display_pok( struct work* work )
--- a/algo/quark/hmq1725-4way.c
+++ b/algo/quark/hmq1725-4way.c
@@ -912,7 +912,7 @@ extern void hmq1725_4way_hash(void *state, const void *input)
   sph_whirlpool512_full( &ctx.whirlpool, hash2, hash2, 64 );
   sph_whirlpool512_full( &ctx.whirlpool, hash3, hash3, 64 );

-// A = fugue serial, B = sha512 prarallel
+// A = fugue serial, B = sha512 parallel
   
   intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );

--- a/algo/quark/hmq1725.c
+++ b/algo/quark/hmq1725.c
@@ -17,7 +17,7 @@
 #include "algo/shabal/sph_shabal.h"
 #include "algo/whirlpool/sph_whirlpool.h"
 #include "algo/haval/sph-haval.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/groestl/aes_ni/hash-groestl.h"
  #include "algo/echo/aes_ni/hash_api.h"
@@ -44,7 +44,7 @@ typedef struct {
  sph_hamsi512_context    hamsi1;
  sph_shabal512_context   shabal1;
  sph_whirlpool_context   whirlpool1, whirlpool2, whirlpool3, whirlpool4;
-  SHA512_CTX              sha1, sha2;
+  sph_sha512_context      sha1, sha2;
  sph_haval256_5_context  haval1, haval2;
 #if defined(__AES__)
  hashState_echo          echo1, echo2;
@@ -106,8 +106,8 @@ void init_hmq1725_ctx()
    sph_whirlpool_init(&hmq1725_ctx.whirlpool3);
    sph_whirlpool_init(&hmq1725_ctx.whirlpool4);

-    SHA512_Init( &hmq1725_ctx.sha1 );
-    SHA512_Init( &hmq1725_ctx.sha2 );
+    sph_sha512_init( &hmq1725_ctx.sha1 );
+    sph_sha512_init( &hmq1725_ctx.sha2 );

    sph_haval256_5_init(&hmq1725_ctx.haval1);
    sph_haval256_5_init(&hmq1725_ctx.haval2);
@@ -285,8 +285,8 @@ extern void hmq1725hash(void *state, const void *input)
    }
    else
    {
-        SHA512_Update( &h_ctx.sha1, hashB, 64 );
-        SHA512_Final( (unsigned char*) hashA, &h_ctx.sha1 );
+        sph_sha512( &h_ctx.sha1, hashB, 64 );
+        sph_sha512_close( &h_ctx.sha1, hashA );
    }

 #if defined(__AES__)
@@ -297,8 +297,8 @@ extern void hmq1725hash(void *state, const void *input)
    sph_groestl512_close(&h_ctx.groestl2, hashB); //4
 #endif

-    SHA512_Update( &h_ctx.sha2, hashB, 64 );
-    SHA512_Final( (unsigned char*) hashA, &h_ctx.sha2 );
+    sph_sha512( &h_ctx.sha2, hashB, 64 );
+    sph_sha512_close( &h_ctx.sha2, hashA );

    if ( hashA[0] & mask ) //4
    {
--- a/algo/quark/quark-4way.c
+++ b/algo/quark/quark-4way.c
@@ -127,10 +127,8 @@ void quark_8way_hash( void *state, const void *input )

     rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );

-     if ( ( vh_mask & 0x0f ) != 0x0f )
-       groestl512_4way_full( &ctx.groestl, vhashA, vhashA, 64 );
-     if ( ( vh_mask & 0xf0 ) != 0xf0 )
-       groestl512_4way_full( &ctx.groestl, vhashB, vhashB, 64 );
+     groestl512_4way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_4way_full( &ctx.groestl, vhashB, vhashB, 64 );

     rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );

@@ -139,22 +137,14 @@ void quark_8way_hash( void *state, const void *input )
    dintrlv_8x64( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
                  vhash, 512 );

-    if ( hash0[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
-    if ( hash1[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
-    if ( hash2[0] & 8)
-       groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
-    if ( hash3[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
-    if ( hash4[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash4, (char*)hash4, 512 );
-    if ( hash5[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash5, (char*)hash5, 512 );
-    if ( hash6[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash6, (char*)hash6, 512 );
-    if ( hash7[0] & 8 )
-       groestl512_full( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash4, (char*)hash4, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash5, (char*)hash5, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash6, (char*)hash6, 512 );
+    groestl512_full( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );

    intrlv_8x64( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
                 512 );
--- a/algo/ripemd/lbry.c
+++ b/algo/ripemd/lbry.c
@@ -7,28 +7,28 @@
 #include <string.h>
 #include <stdio.h>
 #include "sph_ripemd.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"

 void lbry_hash(void* output, const void* input)
 {
-   SHA256_CTX              ctx_sha256 __attribute__ ((aligned (64)));
-   SHA512_CTX              ctx_sha512 __attribute__ ((aligned (64)));
-   sph_ripemd160_context   ctx_ripemd __attribute__ ((aligned (64)));
+   sph_sha256_context    ctx_sha256 __attribute__ ((aligned (64)));
+   sph_sha512_context    ctx_sha512 __attribute__ ((aligned (64)));
+   sph_ripemd160_context ctx_ripemd __attribute__ ((aligned (64)));
   uint32_t _ALIGN(64) hashA[16];
   uint32_t _ALIGN(64) hashB[16];
   uint32_t _ALIGN(64) hashC[16];

-   SHA256_Init( &ctx_sha256 );
-   SHA256_Update( &ctx_sha256, input, 112 );
-   SHA256_Final( (unsigned char*) hashA, &ctx_sha256 );
+   sph_sha256_init( &ctx_sha256 );
+   sph_sha256( &ctx_sha256, input, 112 );
+   sph_sha256_close( &ctx_sha256, hashA );

-   SHA256_Init( &ctx_sha256 );
-   SHA256_Update( &ctx_sha256, hashA, 32 );
-   SHA256_Final( (unsigned char*) hashA, &ctx_sha256 );
+   sph_sha256_init( &ctx_sha256 );
+   sph_sha256( &ctx_sha256, hashA, 32 );
+   sph_sha256_close( &ctx_sha256, hashA );

-   SHA512_Init( &ctx_sha512 );
-   SHA512_Update( &ctx_sha512, hashA, 32 );
-   SHA512_Final( (unsigned char*) hashA, &ctx_sha512 );
+   sph_sha512_init( &ctx_sha512 );
+   sph_sha512( &ctx_sha512, hashA, 32 );
+   sph_sha512_close( &ctx_sha512, hashA );

   sph_ripemd160_init( &ctx_ripemd );
   sph_ripemd160 ( &ctx_ripemd, hashA, 32 );
@@ -38,14 +38,14 @@ void lbry_hash(void* output, const void* input)
   sph_ripemd160 ( &ctx_ripemd, hashA+8, 32 );
   sph_ripemd160_close( &ctx_ripemd, hashC );

-   SHA256_Init( &ctx_sha256 );
-   SHA256_Update( &ctx_sha256, hashB, 20 );
-   SHA256_Update( &ctx_sha256, hashC, 20 );
-   SHA256_Final( (unsigned char*) hashA, &ctx_sha256 );
+   sph_sha256_init( &ctx_sha256 );
+   sph_sha256( &ctx_sha256, hashB, 20 );
+   sph_sha256( &ctx_sha256, hashC, 20 );
+   sph_sha256_close( &ctx_sha256, hashA );

-   SHA256_Init( &ctx_sha256 );
-   SHA256_Update( &ctx_sha256, hashA, 32 );
-   SHA256_Final( (unsigned char*) hashA, &ctx_sha256 );
+   sph_sha256_init( &ctx_sha256 );
+   sph_sha256( &ctx_sha256, hashA, 32 );
+   sph_sha256_close( &ctx_sha256, hashA );

   memcpy( output, hashA, 32 );
 }
--- a/algo/sha/hmac-sha256-hash.c
+++ b/algo/sha/hmac-sha256-hash.c
@@ -39,10 +39,17 @@
 void
 SHA256_Buf( const void * in, size_t len, uint8_t digest[32] )
 {
-	SHA256_CTX ctx;
+#if defined(HMAC_SPH_SHA)
+   sph_sha256_context ctx;
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, in, len );
+   sph_sha256_close( &ctx, digest );
+#else
+   SHA256_CTX ctx;
   SHA256_Init( &ctx );
   SHA256_Update( &ctx, in, len );
   SHA256_Final( digest, &ctx );
+#endif
 }

 /**
@@ -64,35 +71,59 @@ HMAC_SHA256_Buf( const void *K, size_t Klen, const void *in, size_t len,
 void
 HMAC_SHA256_Init( HMAC_SHA256_CTX *ctx, const void *_K, size_t Klen )
 {
-	unsigned char pad[64];
-	unsigned char khash[32];
-	const unsigned char * K = _K;
-	size_t i;
+   unsigned char pad[64];
+   unsigned char khash[32];
+   const unsigned char * K = _K;
+   size_t i;

-	/* If Klen > 64, the key is really SHA256(K). */
-	if ( Klen > 64 )
+   /* If Klen > 64, the key is really SHA256(K). */
+   if ( Klen > 64 )
   {
-		SHA256_Init( &ctx->ictx );
-		SHA256_Update( &ctx->ictx, K, Klen );
-		SHA256_Final( khash, &ctx->ictx );
-		K = khash;
-		Klen = 32;
-	}
+	   
+#if defined(HMAC_SPH_SHA)
+      sph_sha256_init( &ctx->ictx );
+      sph_sha256( &ctx->ictx, K, Klen );
+      sph_sha256_close( &ctx->ictx, khash );
+#else
+      SHA256_Init( &ctx->ictx );
+      SHA256_Update( &ctx->ictx, K, Klen );
+      SHA256_Final( khash, &ctx->ictx );
+#endif
+       K = khash;
+       Klen = 32;
+   }

-	/* Inner SHA256 operation is SHA256(K xor [block of 0x36] || data). */
+   /* Inner SHA256 operation is SHA256(K xor [block of 0x36] || data). */
+#if defined(HMAC_SPH_SHA)
+   sph_sha256_init( &ctx->ictx );
+#else
   SHA256_Init( &ctx->ictx );
-
+#endif

   for ( i = 0; i < Klen; i++ )  pad[i] = K[i] ^ 0x36;
-   memset( pad + Klen, 0x36, 64 - Klen );
-	SHA256_Update( &ctx->ictx, pad, 64 );

-	/* Outer SHA256 operation is SHA256(K xor [block of 0x5c] || hash). */
-	SHA256_Init( &ctx->octx );
+   memset( pad + Klen, 0x36, 64 - Klen );
+#if defined(HMAC_SPH_SHA)
+   sph_sha256( &ctx->ictx, pad, 64 );
+#else
+   SHA256_Update( &ctx->ictx, pad, 64 );
+#endif
+
+   /* Outer SHA256 operation is SHA256(K xor [block of 0x5c] || hash). */
+#if defined(HMAC_SPH_SHA)
+   sph_sha256_init( &ctx->octx );
+#else   
+   SHA256_Init( &ctx->octx );
+#endif

   for ( i = 0; i < Klen; i++ )  pad[i] = K[i] ^ 0x5c;
+
   memset( pad + Klen, 0x5c, 64 - Klen );
-	SHA256_Update( &ctx->octx, pad, 64 );
+#if defined(HMAC_SPH_SHA)
+   sph_sha256( &ctx->octx, pad, 64 );
+#else
+   SHA256_Update( &ctx->octx, pad, 64 );
+#endif
 }

 /* Add bytes to the HMAC-SHA256 operation. */
@@ -100,23 +131,33 @@ void
 HMAC_SHA256_Update( HMAC_SHA256_CTX *ctx, const void *in, size_t len )
 {
 	/* Feed data to the inner SHA256 operation. */
-	SHA256_Update( &ctx->ictx, in, len );
+#if defined(HMAC_SPH_SHA)
+   sph_sha256( &ctx->ictx, in, len );
+#else
+   SHA256_Update( &ctx->ictx, in, len );
+#endif
 }

 /* Finish an HMAC-SHA256 operation. */
 void
 HMAC_SHA256_Final( unsigned char digest[32], HMAC_SHA256_CTX *ctx )
 {
-	unsigned char ihash[32];
+   unsigned char ihash[32];

-	/* Finish the inner SHA256 operation. */
-	SHA256_Final( ihash, &ctx->ictx );
+#if defined(HMAC_SPH_SHA)
+   sph_sha256_close( &ctx->ictx, ihash );
+   sph_sha256( &ctx->octx, ihash, 32 );
+   sph_sha256_close( &ctx->octx, digest );
+#else
+   /* Finish the inner SHA256 operation. */
+   SHA256_Final( ihash, &ctx->ictx );

-	/* Feed the inner hash to the outer SHA256 operation. */
-	SHA256_Update( &ctx->octx, ihash, 32 );
+   /* Feed the inner hash to the outer SHA256 operation. */
+   SHA256_Update( &ctx->octx, ihash, 32 );

-	/* Finish the outer SHA256 operation. */
-	SHA256_Final( digest, &ctx->octx );
+   /* Finish the outer SHA256 operation. */
+   SHA256_Final( digest, &ctx->octx );
+#endif
 }

 /**
--- a/algo/sha/hmac-sha256-hash.h
+++ b/algo/sha/hmac-sha256-hash.h
@@ -29,14 +29,24 @@
 #ifndef HMAC_SHA256_H__
 #define HMAC_SHA256_H__

+//#define HMAC_SSL_SHA 1
+#define HMAC_SPH_SHA 1
+
 #include <sys/types.h>
 #include <stdint.h>
+#include "sph_sha2.h"
 #include <openssl/sha.h>

+
 typedef struct HMAC_SHA256Context
 {
-     SHA256_CTX ictx;
-     SHA256_CTX octx;
+#if defined(HMAC_SPH_SHA)
+   sph_sha256_context ictx;
+   sph_sha256_context octx;
+#else
+   SHA256_CTX ictx;
+   SHA256_CTX octx;
+#endif
 } HMAC_SHA256_CTX;

 void SHA256_Buf( const void *, size_t len, uint8_t digest[32] );
--- a/algo/sha/sha2.c
+++ b/algo/sha/sha2.c
@@ -12,7 +12,6 @@

 #include <string.h>
 #include <inttypes.h>
-#include <openssl/sha.h>

 #if defined(USE_ASM) && defined(__arm__) && defined(__APCS_32__)
 #define EXTERN_SHA256
@@ -198,16 +197,6 @@ static void sha256d_80_swap(uint32_t *hash, const uint32_t *data)

 extern void sha256d(unsigned char *hash, const unsigned char *data, int len)
 {
-#if defined(__SHA__)
-   SHA256_CTX ctx;
-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, data, len );
-   SHA256_Final( (unsigned char*)hash, &ctx );
-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
-#else
-
   uint32_t S[16], T[16];
 	int i, r;

@@ -229,7 +218,6 @@ extern void sha256d(unsigned char *hash, const unsigned char *data, int len)
 	sha256_transform(T, S, 0);
 	for (i = 0; i < 8; i++)
 		be32enc((uint32_t *)hash + i, T[i]);
-#endif
 }

 static inline void sha256d_preextend(uint32_t *W)
@@ -676,14 +664,9 @@ int scanhash_SHA256d( struct work *work, const uint32_t max_nonce,

 bool register_sha256d_algo( algo_gate_t* gate )
 {
-#if defined(__SHA__)
-   gate->optimizations = SHA_OPT;
-   gate->scanhash = (void*)&scanhash_SHA256d;
-#else
   gate->optimizations = SSE2_OPT | AVX2_OPT;
   gate->scanhash = (void*)&scanhash_sha256d;
-#endif
-    gate->hash     = (void*)&sha256d;
-    return true;
+   gate->hash     = (void*)&sha256d;
+   return true;
 };

--- a/algo/sha/sha256-hash-opt.c
+++ b/algo/sha/sha256-hash-opt.c
@@ -0,0 +1,200 @@
+/*   Intel SHA extensions using C intrinsics               */
+/*   Written and place in public domain by Jeffrey Walton  */
+/*   Based on code from Intel, and by Sean Gulley for      */
+/*   the miTLS project.                                    */
+
+// A drop in replacement for the function of the same name in sph_sha2.c.
+
+#if defined(__SHA__)
+
+#include "simd-utils.h"
+
+static void sha2_round( const uint8_t input[], uint32_t state[8] )
+{
+    __m128i STATE0, STATE1;
+    __m128i MSG, TMP, MASK;
+    __m128i TMSG0, TMSG1, TMSG2, TMSG3;
+    __m128i ABEF_SAVE, CDGH_SAVE;
+
+    // Load initial values
+    TMP = _mm_load_si128((__m128i*) &state[0]);
+    STATE1 = _mm_load_si128((__m128i*) &state[4]);
+    MASK = _mm_set_epi64x(0x0c0d0e0f08090a0bULL, 0x0405060700010203ULL);
+
+    TMP = _mm_shuffle_epi32(TMP, 0xB1); // CDAB
+    STATE1 = _mm_shuffle_epi32(STATE1, 0x1B); // EFGH
+    STATE0 = _mm_alignr_epi8(TMP, STATE1, 8); // ABEF
+    STATE1 = _mm_blend_epi16(STATE1, TMP, 0xF0); // CDGH
+
+    // Save current hash
+    ABEF_SAVE = STATE0;
+    CDGH_SAVE = STATE1;
+
+    // Rounds 0-3
+    MSG = _mm_load_si128((const __m128i*) (input+0));
+    TMSG0 = _mm_shuffle_epi8(MSG, MASK);
+    MSG = _mm_add_epi32(TMSG0, _mm_set_epi64x(0xE9B5DBA5B5C0FBCFULL, 0x71374491428A2F98ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+
+    // Rounds 4-7
+    TMSG1 = _mm_load_si128((const __m128i*) (input+16));
+    TMSG1 = _mm_shuffle_epi8(TMSG1, MASK);
+    MSG = _mm_add_epi32(TMSG1, _mm_set_epi64x(0xAB1C5ED5923F82A4ULL, 0x59F111F13956C25BULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG0 = _mm_sha256msg1_epu32(TMSG0, TMSG1);
+
+    // Rounds 8-11
+    TMSG2 = _mm_load_si128((const __m128i*) (input+32));
+    TMSG2 = _mm_shuffle_epi8(TMSG2, MASK);
+    MSG = _mm_add_epi32(TMSG2, _mm_set_epi64x(0x550C7DC3243185BEULL, 0x12835B01D807AA98ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG1 = _mm_sha256msg1_epu32(TMSG1, TMSG2);
+
+    // Rounds 12-15
+    TMSG3 = _mm_load_si128((const __m128i*) (input+48));
+    TMSG3 = _mm_shuffle_epi8(TMSG3, MASK);
+    MSG = _mm_add_epi32(TMSG3, _mm_set_epi64x(0xC19BF1749BDC06A7ULL, 0x80DEB1FE72BE5D74ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG3, TMSG2, 4);
+    TMSG0 = _mm_add_epi32(TMSG0, TMP);
+    TMSG0 = _mm_sha256msg2_epu32(TMSG0, TMSG3);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG2 = _mm_sha256msg1_epu32(TMSG2, TMSG3);
+
+    // Rounds 16-19
+    MSG = _mm_add_epi32(TMSG0, _mm_set_epi64x(0x240CA1CC0FC19DC6ULL, 0xEFBE4786E49B69C1ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG0, TMSG3, 4);
+    TMSG1 = _mm_add_epi32(TMSG1, TMP);
+    TMSG1 = _mm_sha256msg2_epu32(TMSG1, TMSG0);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG3 = _mm_sha256msg1_epu32(TMSG3, TMSG0);
+
+    // Rounds 20-23
+    MSG = _mm_add_epi32(TMSG1, _mm_set_epi64x(0x76F988DA5CB0A9DCULL, 0x4A7484AA2DE92C6FULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG1, TMSG0, 4);
+    TMSG2 = _mm_add_epi32(TMSG2, TMP);
+    TMSG2 = _mm_sha256msg2_epu32(TMSG2, TMSG1);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG0 = _mm_sha256msg1_epu32(TMSG0, TMSG1);
+
+    // Rounds 24-27
+    MSG = _mm_add_epi32(TMSG2, _mm_set_epi64x(0xBF597FC7B00327C8ULL, 0xA831C66D983E5152ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG2, TMSG1, 4);
+    TMSG3 = _mm_add_epi32(TMSG3, TMP);
+    TMSG3 = _mm_sha256msg2_epu32(TMSG3, TMSG2);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG1 = _mm_sha256msg1_epu32(TMSG1, TMSG2);
+
+    // Rounds 28-31
+    MSG = _mm_add_epi32(TMSG3, _mm_set_epi64x(0x1429296706CA6351ULL,  0xD5A79147C6E00BF3ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG3, TMSG2, 4);
+    TMSG0 = _mm_add_epi32(TMSG0, TMP);
+    TMSG0 = _mm_sha256msg2_epu32(TMSG0, TMSG3);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG2 = _mm_sha256msg1_epu32(TMSG2, TMSG3);
+
+    // Rounds 32-35
+    MSG = _mm_add_epi32(TMSG0, _mm_set_epi64x(0x53380D134D2C6DFCULL, 0x2E1B213827B70A85ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG0, TMSG3, 4);
+    TMSG1 = _mm_add_epi32(TMSG1, TMP);
+    TMSG1 = _mm_sha256msg2_epu32(TMSG1, TMSG0);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG3 = _mm_sha256msg1_epu32(TMSG3, TMSG0);
+
+    // Rounds 36-39
+    MSG = _mm_add_epi32(TMSG1, _mm_set_epi64x(0x92722C8581C2C92EULL, 0x766A0ABB650A7354ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG1, TMSG0, 4);
+    TMSG2 = _mm_add_epi32(TMSG2, TMP);
+    TMSG2 = _mm_sha256msg2_epu32(TMSG2, TMSG1);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG0 = _mm_sha256msg1_epu32(TMSG0, TMSG1);
+
+    // Rounds 40-43
+    MSG = _mm_add_epi32(TMSG2, _mm_set_epi64x(0xC76C51A3C24B8B70ULL, 0xA81A664BA2BFE8A1ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG2, TMSG1, 4);
+    TMSG3 = _mm_add_epi32(TMSG3, TMP);
+    TMSG3 = _mm_sha256msg2_epu32(TMSG3, TMSG2);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG1 = _mm_sha256msg1_epu32(TMSG1, TMSG2);
+
+    // Rounds 44-47
+    MSG = _mm_add_epi32(TMSG3, _mm_set_epi64x(0x106AA070F40E3585ULL, 0xD6990624D192E819ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG3, TMSG2, 4);
+    TMSG0 = _mm_add_epi32(TMSG0, TMP);
+    TMSG0 = _mm_sha256msg2_epu32(TMSG0, TMSG3);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG2 = _mm_sha256msg1_epu32(TMSG2, TMSG3);
+
+    // Rounds 48-51
+    MSG = _mm_add_epi32(TMSG0, _mm_set_epi64x(0x34B0BCB52748774CULL, 0x1E376C0819A4C116ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG0, TMSG3, 4);
+    TMSG1 = _mm_add_epi32(TMSG1, TMP);
+    TMSG1 = _mm_sha256msg2_epu32(TMSG1, TMSG0);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+    TMSG3 = _mm_sha256msg1_epu32(TMSG3, TMSG0);
+
+    // Rounds 52-55
+    MSG = _mm_add_epi32(TMSG1, _mm_set_epi64x(0x682E6FF35B9CCA4FULL, 0x4ED8AA4A391C0CB3ULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG1, TMSG0, 4);
+    TMSG2 = _mm_add_epi32(TMSG2, TMP);
+    TMSG2 = _mm_sha256msg2_epu32(TMSG2, TMSG1);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+
+    // Rounds 56-59
+    MSG = _mm_add_epi32(TMSG2, _mm_set_epi64x(0x8CC7020884C87814ULL, 0x78A5636F748F82EEULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    TMP = _mm_alignr_epi8(TMSG2, TMSG1, 4);
+    TMSG3 = _mm_add_epi32(TMSG3, TMP);
+    TMSG3 = _mm_sha256msg2_epu32(TMSG3, TMSG2);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+
+    // Rounds 60-63
+    MSG = _mm_add_epi32(TMSG3, _mm_set_epi64x(0xC67178F2BEF9A3F7ULL, 0xA4506CEB90BEFFFAULL));
+    STATE1 = _mm_sha256rnds2_epu32(STATE1, STATE0, MSG);
+    MSG = _mm_shuffle_epi32(MSG, 0x0E);
+    STATE0 = _mm_sha256rnds2_epu32(STATE0, STATE1, MSG);
+
+    // Add values back to state
+    STATE0 = _mm_add_epi32(STATE0, ABEF_SAVE);
+    STATE1 = _mm_add_epi32(STATE1, CDGH_SAVE);
+
+    TMP = _mm_shuffle_epi32(STATE0, 0x1B); // FEBA
+    STATE1 = _mm_shuffle_epi32(STATE1, 0xB1); // DCHG
+    STATE0 = _mm_blend_epi16(TMP, STATE1, 0xF0); // DCBA
+    STATE1 = _mm_alignr_epi8(STATE1, TMP, 8); // ABEF
+
+    // Save state
+    _mm_store_si128((__m128i*) &state[0], STATE0);
+    _mm_store_si128((__m128i*) &state[4], STATE1);
+}
+
+
+#endif
--- a/algo/sha/sha256q-4way.c
+++ b/algo/sha/sha256q-4way.c
@@ -5,6 +5,79 @@
 #include <stdio.h>
 #include "sha-hash-4way.h"

+#if defined(SHA256T_16WAY)
+
+static __thread sha256_16way_context sha256_ctx16 __attribute__ ((aligned (64)));
+
+void sha256q_16way_hash( void* output, const void* input )
+{
+   uint32_t vhash[8*16] __attribute__ ((aligned (64)));
+   sha256_16way_context ctx;
+   memcpy( &ctx, &sha256_ctx16, sizeof ctx );
+
+   sha256_16way_update( &ctx, input + (64<<4), 16 );
+   sha256_16way_close( &ctx, vhash );
+
+   sha256_16way_init( &ctx );
+   sha256_16way_update( &ctx, vhash, 32 );
+   sha256_16way_close( &ctx, vhash );
+
+   sha256_16way_init( &ctx );
+   sha256_16way_update( &ctx, vhash, 32 );
+   sha256_16way_close( &ctx, vhash );
+
+   sha256_16way_init( &ctx );
+   sha256_16way_update( &ctx, vhash, 32 );
+   sha256_16way_close( &ctx, output );
+}
+
+int scanhash_sha256q_16way( struct work *work, const uint32_t max_nonce,
+                           uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t vdata[20*16] __attribute__ ((aligned (64)));
+   uint32_t hash32[8*16] __attribute__ ((aligned (32)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 = &(hash32[7<<4]);
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t targ32_d7 = ptarget[7];
+   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 16;
+   uint32_t n = first_nonce;
+   __m512i  *noncev = (__m512i*)vdata + 19;   // aligned
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+
+   mm512_bswap32_intrlv80_16x32( vdata, pdata );
+   *noncev = _mm512_set_epi32( n+15, n+14, n+13, n+12, n+11, n+10, n+9, n+8,
+                               n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n+1, n );
+   sha256_16way_init( &sha256_ctx16 );
+   sha256_16way_update( &sha256_ctx16, vdata, 64 );
+
+   do
+   {
+     pdata[19] = n;
+     sha256q_16way_hash( hash32, vdata );
+     for ( int lane = 0; lane < 16; lane++ )
+     if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
+     {
+        extr_lane_16x32( lane_hash, hash32, lane, 256 );
+        if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
+        {
+           pdata[19] = bswap_32( n + lane );
+           submit_solution( work, lane_hash, mythr );
+        }
+      }
+      *noncev = _mm512_add_epi32( *noncev, m512_const1_32( 16 ) );
+      n += 16;
+   } while ( (n < last_nonce) && !work_restart[thr_id].restart );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+
+#endif
+
 #if defined(SHA256T_8WAY)

 static __thread sha256_8way_context sha256_ctx8 __attribute__ ((aligned (64)));
@@ -31,68 +104,47 @@ void sha256q_8way_hash( void* output, const void* input )
   sha256_8way_close( &ctx, output );
 }

-int scanhash_sha256q_8way( struct work *work, uint32_t max_nonce,
-	                   uint64_t *hashes_done, struct thr_info *mythr )
+int scanhash_sha256q_8way( struct work *work, const uint32_t max_nonce,
+                           uint64_t *hashes_done, struct thr_info *mythr )
 {
-   uint32_t vdata[20*8] __attribute__ ((aligned (64)));
-   uint32_t hash[8*8] __attribute__ ((aligned (32)));
+   uint32_t vdata[20*8]  __attribute__ ((aligned (64)));
+   uint32_t hash32[8*8]    __attribute__ ((aligned (32)));
   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 = &(hash32[7<<3]);
   uint32_t *pdata = work->data;
-   uint32_t *ptarget = work->target;
-   const uint32_t Htarg = ptarget[7];
+   const uint32_t *ptarget = work->target;
+   const uint32_t targ32_d7 = ptarget[7];
   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 8;
   uint32_t n = first_nonce;
   __m256i  *noncev = (__m256i*)vdata + 19;   // aligned
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;

-   const uint64_t htmax[] = {          0,
-                                     0xF,
-                                    0xFF,
-                                   0xFFF,
-                                  0xFFFF,
-                              0x10000000 };
-   const uint32_t masks[] = {  0xFFFFFFFF,
-                               0xFFFFFFF0,
-                               0xFFFFFF00,
-                               0xFFFFF000,
-                               0xFFFF0000,
-                                        0 };
-
-   // Need big endian data
   mm256_bswap32_intrlv80_8x32( vdata, pdata );
+   *noncev = _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n );
   sha256_8way_init( &sha256_ctx8 );
   sha256_8way_update( &sha256_ctx8, vdata, 64 );

-   for ( int m = 0; m < 6; m++ ) if ( Htarg <= htmax[m] )
+   do
   {
-      uint32_t mask = masks[m];
-      do
-      {
-         *noncev = mm256_bswap_32(
-		            _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n ) );
-
-	      pdata[19] = n;
-         sha256q_8way_hash( hash, vdata );
-
-         uint32_t *hash7 = &(hash[7<<3]); 
-	 
-         for ( int lane = 0; lane < 8; lane++ )
-         if ( !( hash7[ lane ] & mask ) )
-         { 
-            // deinterleave hash for lane
-	         extr_lane_8x32( lane_hash, hash, lane, 256 );
-
-	         if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
-            {
-	           pdata[19] = n + lane;
-              submit_solution( work, lane_hash, mythr );
-            }
-	      }
-         n += 8;
-      } while ( (n < max_nonce-10) && !work_restart[thr_id].restart );
-      break;
-   }
-   *hashes_done = n - first_nonce + 1;
+     pdata[19] = n;
+     sha256q_8way_hash( hash32, vdata );
+     for ( int lane = 0; lane < 8; lane++ )
+     if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
+     {
+        extr_lane_8x32( lane_hash, hash32, lane, 256 );
+        if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
+        {
+           pdata[19] = bswap_32( n + lane );
+           submit_solution( work, lane_hash, mythr );
+        }
+      }
+      *noncev = _mm256_add_epi32( *noncev, m256_const1_32( 8 ) );
+      n += 8;
+   } while ( (n < last_nonce) && !work_restart[thr_id].restart );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
   return 0;
 }

--- a/algo/sha/sha256q.c
+++ b/algo/sha/sha256q.c
@@ -1,108 +1,74 @@
 #include "sha256t-gate.h"
-
-#if !defined(SHA256T_16WAY) && !defined(SHA256T_8WAY) && !defined(SHA256T_4WAY)
-
 #include <stdlib.h>
 #include <stdint.h>
 #include <string.h>
 #include <stdio.h>
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"

-static __thread SHA256_CTX sha256q_ctx __attribute__ ((aligned (64)));
+static __thread sph_sha256_context sha256q_ctx __attribute__ ((aligned (64)));

 void sha256q_midstate( const void* input )
 {
-    SHA256_Init( &sha256q_ctx );
-    SHA256_Update( &sha256q_ctx, input, 64 );
+   sph_sha256_init( &sha256q_ctx );
+   sph_sha256( &sha256q_ctx, input, 64 );
 }

-void sha256q_hash( void* output, const void* input )
+int sha256q_hash( void* output, const void* input )
 {
   uint32_t _ALIGN(64) hash[16];
   const int midlen = 64;            // bytes
   const int tail   = 80 - midlen;   // 16

-   SHA256_CTX ctx __attribute__ ((aligned (64)));
+   sph_sha256_context ctx __attribute__ ((aligned (64)));
   memcpy( &ctx, &sha256q_ctx, sizeof sha256q_ctx );

-   SHA256_Update( &ctx, input + midlen, tail );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256( &ctx, input + midlen, tail );
+   sph_sha256_close( &ctx, hash );

-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, hash, 32 );
+   sph_sha256_close( &ctx, hash );

-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, hash, 32 );
+   sph_sha256_close( &ctx, hash );

-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, hash, 32 );
+   sph_sha256_close( &ctx, output );

-   memcpy( output, hash, 32 );
+   return 1;
 }

 int scanhash_sha256q( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr )
 {
+   uint32_t edata[20] __attribute__((aligned(64)));
+   uint32_t hash[8] __attribute__((aligned(64)));
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
-   uint32_t n = pdata[19] - 1;
   const uint32_t first_nonce = pdata[19];
-   const uint32_t Htarg = ptarget[7];
-#ifdef _MSC_VER
-   uint32_t __declspec(align(32)) hash64[8];
-#else
-   uint32_t hash64[8] __attribute__((aligned(32)));
-#endif
-   uint32_t endiandata[32];
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   const uint32_t last_nonce = max_nonce - 1;
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;

-   uint64_t htmax[] = {
-		0,
-		0xF,
-		0xFF,
-		0xFFF,
-		0xFFFF,
-		0x10000000
-	};
-   uint32_t masks[] = {
-		0xFFFFFFFF,
-		0xFFFFFFF0,
-		0xFFFFFF00,
-		0xFFFFF000,
-		0xFFFF0000,
-		0
-	};
+   mm128_bswap32_80( edata, pdata );
+   sha256q_midstate( edata );

-   // we need bigendian data...
-   casti_m128i( endiandata, 0 ) = mm128_bswap_32( casti_m128i( pdata, 0 ) );
-   casti_m128i( endiandata, 1 ) = mm128_bswap_32( casti_m128i( pdata, 1 ) );
-   casti_m128i( endiandata, 2 ) = mm128_bswap_32( casti_m128i( pdata, 2 ) );
-   casti_m128i( endiandata, 3 ) = mm128_bswap_32( casti_m128i( pdata, 3 ) );
-   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
-
-   sha256q_midstate( endiandata );
-
-   for ( int m = 0; m < 6; m++ )
+   do
   {
-      if ( Htarg <= htmax[m] )
+      edata[19] = n;
+      if ( likely( sha256q_hash( hash, edata ) ) )
+      if ( unlikely( valid_hash( hash, ptarget ) && !bench ) )
      {
-         uint32_t mask = masks[m];
-         do {
-            pdata[19] = ++n;
-            be32enc(&endiandata[19], n);
-            sha256q_hash( hash64, endiandata );
-            if ( !( hash64[7] & mask ) )
-            if ( fulltest( hash64, ptarget ) && !opt_benchmark )
-               submit_solution( work, hash64, mythr );
-         } while ( n < max_nonce && !work_restart[thr_id].restart );
-         break;
+         pdata[19] = bswap_32( n );
+         submit_solution( work, hash, mythr );
      }
-   }
-   *hashes_done = n - first_nonce + 1;
+      n++;
+   } while ( n < last_nonce && !work_restart[thr_id].restart );
+   *hashes_done = n - first_nonce;
   pdata[19] = n;
   return 0;
 }
-#endif
+
--- a/algo/sha/sha256t-4way.c
+++ b/algo/sha/sha256t-4way.c
@@ -5,6 +5,75 @@
 #include <stdio.h>
 #include "sha-hash-4way.h"

+#if defined(SHA256T_16WAY)
+
+static __thread sha256_16way_context sha256_ctx16 __attribute__ ((aligned (64)));
+
+void sha256t_16way_hash( void* output, const void* input )
+{
+   uint32_t vhash[8*16] __attribute__ ((aligned (64)));
+   sha256_16way_context ctx;
+   memcpy( &ctx, &sha256_ctx16, sizeof ctx );
+
+   sha256_16way_update( &ctx, input + (64<<4), 16 );
+   sha256_16way_close( &ctx, vhash );
+
+   sha256_16way_init( &ctx );
+   sha256_16way_update( &ctx, vhash, 32 );
+   sha256_16way_close( &ctx, vhash );
+
+   sha256_16way_init( &ctx );
+   sha256_16way_update( &ctx, vhash, 32 );
+   sha256_16way_close( &ctx, output );
+}
+
+int scanhash_sha256t_16way( struct work *work, const uint32_t max_nonce,
+                           uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t vdata[20*16] __attribute__ ((aligned (64)));
+   uint32_t hash32[8*16] __attribute__ ((aligned (32)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 = &(hash32[7<<4]);
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t targ32_d7 = ptarget[7];
+   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 16;
+   uint32_t n = first_nonce;
+   __m512i  *noncev = (__m512i*)vdata + 19;   // aligned
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+
+   mm512_bswap32_intrlv80_16x32( vdata, pdata );
+   *noncev = _mm512_set_epi32( n+15, n+14, n+13, n+12, n+11, n+10, n+9, n+8,
+                               n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n+1, n );
+   sha256_16way_init( &sha256_ctx16 );
+   sha256_16way_update( &sha256_ctx16, vdata, 64 );
+
+   do
+   {
+     pdata[19] = n;
+     sha256t_16way_hash( hash32, vdata );
+     for ( int lane = 0; lane < 16; lane++ )
+     if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
+     {
+        extr_lane_16x32( lane_hash, hash32, lane, 256 );
+        if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
+        {
+           pdata[19] = bswap_32( n + lane );
+           submit_solution( work, lane_hash, mythr );
+        }
+      }
+      *noncev = _mm512_add_epi32( *noncev, m512_const1_32( 16 ) );
+      n += 16;
+   } while ( (n < last_nonce) && !work_restart[thr_id].restart );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+
+#endif
+
 #if defined(SHA256T_8WAY)

 static __thread sha256_8way_context sha256_ctx8 __attribute__ ((aligned (64)));
@@ -31,61 +100,43 @@ int scanhash_sha256t_8way( struct work *work, const uint32_t max_nonce,
                           uint64_t *hashes_done, struct thr_info *mythr )
 {
   uint32_t vdata[20*8]  __attribute__ ((aligned (64)));
-   uint32_t hash[8*8]    __attribute__ ((aligned (32)));
+   uint32_t hash32[8*8]    __attribute__ ((aligned (32)));
   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
-   uint32_t *hash7 = &(hash[7<<3]);
+   uint32_t *hash32_d7 = &(hash32[7<<3]);
   uint32_t *pdata = work->data;
   const uint32_t *ptarget = work->target;
-   const uint32_t Htarg = ptarget[7];
+   const uint32_t targ32_d7 = ptarget[7];
   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 8;
   uint32_t n = first_nonce;
   __m256i  *noncev = (__m256i*)vdata + 19;   // aligned
   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;

-   const uint64_t htmax[] = {          0,
-                                     0xF,
-                                    0xFF,
-                                   0xFFF,
-                                  0xFFFF,
-                              0x10000000 };
-   const uint32_t masks[] = {  0xFFFFFFFF,
-                               0xFFFFFFF0,
-                               0xFFFFFF00,
-                               0xFFFFF000,
-                               0xFFFF0000,
-                                        0 };
-
-
-   // Need big endian data
   mm256_bswap32_intrlv80_8x32( vdata, pdata );
+   *noncev = _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n );
   sha256_8way_init( &sha256_ctx8 );
   sha256_8way_update( &sha256_ctx8, vdata, 64 );

-   for ( int m = 0; m < 6; m++ ) if ( Htarg <= htmax[m] )
+   do
   {
-      const uint32_t mask = masks[m];
-      do
-      {
-        *noncev = mm256_bswap_32( _mm256_set_epi32(
-                                          n+7,n+6,n+5,n+4,n+3,n+2,n+1,n ) );
-         pdata[19] = n;
-         sha256t_8way_hash( hash, vdata );
-         for ( int lane = 0; lane < 8; lane++ )
-         if ( !( hash7[ lane ] & mask ) )
-         {
-            // deinterleave hash for lane
-            extr_lane_8x32( lane_hash, hash, lane, 256 );
-            if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
-            {
-              pdata[19] = n + lane;
-              submit_solution( work, lane_hash, mythr );
-	         }
-         }
-         n += 8;
-      } while ( (n < max_nonce-10) && !work_restart[thr_id].restart );
-      break;
-   }
-   *hashes_done = n - first_nonce + 1;
+     pdata[19] = n;
+     sha256t_8way_hash( hash32, vdata );
+     for ( int lane = 0; lane < 8; lane++ )
+     if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
+     {
+        extr_lane_8x32( lane_hash, hash32, lane, 256 );
+        if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
+        {
+           pdata[19] = bswap_32( n + lane );
+           submit_solution( work, lane_hash, mythr );
+        }
+      }
+      *noncev = _mm256_add_epi32( *noncev, m256_const1_32( 8 ) );
+      n += 8;
+   } while ( (n < last_nonce) && !work_restart[thr_id].restart );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
   return 0;
 }

--- a/algo/sha/sha256t-gate.c
+++ b/algo/sha/sha256t-gate.c
@@ -2,38 +2,41 @@

 bool register_sha256t_algo( algo_gate_t* gate )
 {
-#if defined(SHA256T_8WAY)
-    gate->optimizations = SSE2_OPT | AVX2_OPT | SHA_OPT;
-    gate->scanhash   = (void*)&scanhash_sha256t_8way;
-    gate->hash       = (void*)&sha256t_8way_hash;
-#elif defined(SHA256T_4WAY)
-    gate->optimizations = SSE2_OPT | AVX2_OPT | SHA_OPT;
-    gate->scanhash   = (void*)&scanhash_sha256t_4way;
-    gate->hash       = (void*)&sha256t_4way_hash;
-#else
+    gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
+#if defined(SHA256T_16WAY)
+    gate->scanhash   = (void*)&scanhash_sha256t_16way;
+    gate->hash       = (void*)&sha256t_16way_hash;
+#elif defined(__SHA__)
    gate->optimizations = SHA_OPT;
    gate->scanhash   = (void*)&scanhash_sha256t;
    gate->hash       = (void*)&sha256t_hash;
+#elif defined(SHA256T_8WAY)
+    gate->scanhash   = (void*)&scanhash_sha256t_8way;
+    gate->hash       = (void*)&sha256t_8way_hash;
+#else
+    gate->scanhash   = (void*)&scanhash_sha256t_4way;
+    gate->hash       = (void*)&sha256t_4way_hash;
 #endif
    return true;
 }

 bool register_sha256q_algo( algo_gate_t* gate )
 {
-#if defined(SHA256T_8WAY)
-    gate->optimizations = SSE2_OPT | AVX2_OPT | SHA_OPT;
-    gate->scanhash   = (void*)&scanhash_sha256q_8way;
-    gate->hash       = (void*)&sha256q_8way_hash;
-#elif defined(SHA256T_4WAY)
-    gate->optimizations = SSE2_OPT | AVX2_OPT | SHA_OPT;
-    gate->scanhash   = (void*)&scanhash_sha256q_4way;
-    gate->hash       = (void*)&sha256q_4way_hash;
-#else
+    gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
+#if defined(SHA256T_16WAY)
+    gate->scanhash   = (void*)&scanhash_sha256q_16way;
+    gate->hash       = (void*)&sha256q_16way_hash;
+#elif defined(__SHA__)
    gate->optimizations = SHA_OPT;
    gate->scanhash   = (void*)&scanhash_sha256q;
    gate->hash       = (void*)&sha256q_hash;
+#elif defined(SHA256T_8WAY)
+    gate->scanhash   = (void*)&scanhash_sha256q_8way;
+    gate->hash       = (void*)&sha256q_8way_hash;
+#else
+    gate->scanhash   = (void*)&scanhash_sha256q_4way;
+    gate->hash       = (void*)&sha256q_4way_hash;
 #endif
    return true;
-
 }

--- a/algo/sha/sha256t-gate.h
+++ b/algo/sha/sha256t-gate.h
@@ -4,18 +4,27 @@
 #include <stdint.h>
 #include "algo-gate-api.h"

-// Override multi way on ryzen, SHA is better.
-#if !defined(__SHA__)
- #if defined(__AVX2__)
-  #define SHA256T_8WAY
- #elif defined(__SSE2__)
-  #define SHA256T_4WAY
- #endif
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+  #define SHA256T_16WAY 1
+#elif defined(__AVX2__)
+  #define SHA256T_8WAY 1
+#else
+  #define SHA256T_4WAY 1
 #endif

 bool register_sha256t_algo( algo_gate_t* gate );
 bool register_sha256q_algo( algo_gate_t* gate );

+#if defined(SHA256T_16WAY)
+
+void sha256t_16way_hash( void *output, const void *input );
+int scanhash_sha256t_16way( struct work *work, uint32_t max_nonce,
+                           uint64_t *hashes_done, struct thr_info *mythr );
+void sha256q_16way_hash( void *output, const void *input );
+int scanhash_sha256q_16way( struct work *work, uint32_t max_nonce,
+                           uint64_t *hashes_done, struct thr_info *mythr );
+#endif
+
 #if defined(SHA256T_8WAY)

 void sha256t_8way_hash( void *output, const void *input );
@@ -36,10 +45,11 @@ int scanhash_sha256q_4way( struct work *work, uint32_t max_nonce,
                           uint64_t *hashes_done, struct thr_info *mythr );
 #endif

-void sha256t_hash( void *output, const void *input );
+
+int sha256t_hash( void *output, const void *input );
 int scanhash_sha256t( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr );
-void sha256q_hash( void *output, const void *input );
+int sha256q_hash( void *output, const void *input );
 int scanhash_sha256q( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr );

--- a/algo/sha/sha256t.c
+++ b/algo/sha/sha256t.c
@@ -1,105 +1,73 @@
 #include "sha256t-gate.h"
-
-#if !defined(SHA256T_16WAY) && !defined(SHA256T_8WAY) && !defined(SHA256T_4WAY)
-
 #include <stdlib.h>
 #include <stdint.h>
 #include <string.h>
 #include <stdio.h>
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"

-static __thread SHA256_CTX sha256t_ctx __attribute__ ((aligned (64)));
+// Only used on CPUs with SHA
+
+static __thread sph_sha256_context sha256t_ctx __attribute__ ((aligned (64)));

 void sha256t_midstate( const void* input )
 {
-    SHA256_Init( &sha256t_ctx );
-    SHA256_Update( &sha256t_ctx, input, 64 );
+   sph_sha256_init( &sha256t_ctx );
+   sph_sha256( &sha256t_ctx, input, 64 );
 }

-void sha256t_hash( void* output, const void* input )
+int sha256t_hash( void* output, const void* input )
 {
   uint32_t _ALIGN(64) hash[16];
   const int midlen = 64;            // bytes
   const int tail   = 80 - midlen;   // 16

-   SHA256_CTX ctx __attribute__ ((aligned (64)));
+   sph_sha256_context ctx __attribute__ ((aligned (64)));
   memcpy( &ctx, &sha256t_ctx, sizeof sha256t_ctx );

-   SHA256_Update( &ctx, input + midlen, tail );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256( &ctx, input + midlen, tail );
+   sph_sha256_close( &ctx, hash );

-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, hash, 32 );
+   sph_sha256_close( &ctx, hash );

-   SHA256_Init( &ctx );
-   SHA256_Update( &ctx, hash, 32 );
-   SHA256_Final( (unsigned char*)hash, &ctx );
+   sph_sha256_init( &ctx );
+   sph_sha256( &ctx, hash, 32 );
+   sph_sha256_close( &ctx, output );

-   memcpy( output, hash, 32 );
+   return 1;
 }

 int scanhash_sha256t( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr )
 {
+   uint32_t edata[20] __attribute__((aligned(64)));
+   uint32_t hash[8] __attribute__((aligned(64)));
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
-   uint32_t n = pdata[19] - 1;
   const uint32_t first_nonce = pdata[19];
-   const uint32_t Htarg = ptarget[7];
-#ifdef _MSC_VER
-   uint32_t __declspec(align(32)) hash64[8];
-#else
-   uint32_t hash64[8] __attribute__((aligned(32)));
-#endif
-   uint32_t endiandata[32];
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   const uint32_t last_nonce = max_nonce - 1;
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;

-   uint64_t htmax[] = {
-		0,
-		0xF,
-		0xFF,
-		0xFFF,
-		0xFFFF,
-		0x10000000
-	};
-   uint32_t masks[] = {
-		0xFFFFFFFF,
-		0xFFFFFFF0,
-		0xFFFFFF00,
-		0xFFFFF000,
-		0xFFFF0000,
-		0
-	};
+   mm128_bswap32_80( edata, pdata );
+   sha256t_midstate( edata );

-   // we need bigendian data...
-   casti_m128i( endiandata, 0 ) = mm128_bswap_32( casti_m128i( pdata, 0 ) );
-   casti_m128i( endiandata, 1 ) = mm128_bswap_32( casti_m128i( pdata, 1 ) );
-   casti_m128i( endiandata, 2 ) = mm128_bswap_32( casti_m128i( pdata, 2 ) );
-   casti_m128i( endiandata, 3 ) = mm128_bswap_32( casti_m128i( pdata, 3 ) );
-   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
-
-   sha256t_midstate( endiandata );
-
-   for ( int m = 0; m < 6; m++ )
+   do
   {
-      if ( Htarg <= htmax[m] )
+      edata[19] = n;
+      if ( likely( sha256t_hash( hash, edata ) ) )
+      if ( unlikely( valid_hash( hash, ptarget ) && !bench ) )
      {
-         uint32_t mask = masks[m];
-         do {
-            pdata[19] = ++n;
-            be32enc(&endiandata[19], n);
-            sha256t_hash( hash64, endiandata );
-            if ( !(hash64[7] & mask) )
-            if ( fulltest( hash64, ptarget ) && !opt_benchmark )
-               submit_solution( work, hash64, mythr );
-         } while ( n < max_nonce && !work_restart[thr_id].restart );
-         break;
+         pdata[19] = bswap_32( n );
+         submit_solution( work, hash, mythr );
      }
-   }
-   *hashes_done = n - first_nonce + 1;
+      n++;
+   } while ( n < last_nonce && !work_restart[thr_id].restart );
+   *hashes_done = n - first_nonce;
   pdata[19] = n;
   return 0;
 }
-#endif
+

--- a/algo/sha/sph_sha2.c
+++ b/algo/sha/sph_sha2.c
@@ -71,8 +71,13 @@ static const sph_u32 H256[8] = {
 * of the compression function.
 */

-#if SPH_SMALL_FOOTPRINT_SHA2
+#if defined(__SHA__)

+#include "sha256-hash-opt.c"
+
+#else   // no SHA
+
+/*
 static const sph_u32 K[64] = {
 	SPH_C32(0x428A2F98), SPH_C32(0x71374491),
 	SPH_C32(0xB5C0FBCF), SPH_C32(0xE9B5DBA5),
@@ -107,6 +112,9 @@ static const sph_u32 K[64] = {
 	SPH_C32(0x90BEFFFA), SPH_C32(0xA4506CEB),
 	SPH_C32(0xBEF9A3F7), SPH_C32(0xC67178F2)
 };
+*/
+
+#if SPH_SMALL_FOOTPRINT_SHA2

 #define SHA2_MEXP1(in, pc)   do { \
 		W[pc] = in(pc); \
@@ -191,7 +199,7 @@ static const sph_u32 K[64] = {
 		(r)[7] = SPH_T32((r)[7] + H); \
 	} while (0)

-#else
+#else  // large footprint (default)

 #define SHA2_ROUND_BODY(in, r)   do { \
 		sph_u32 A, B, C, D, E, F, G, H, T1, T2; \
@@ -600,7 +608,7 @@ static const sph_u32 K[64] = {
 		(r)[7] = SPH_T32((r)[7] + H); \
 	} while (0)

-#endif
+#endif  // small footprint else

 /*
 * One round of SHA-224 / SHA-256. The data must be aligned for 32-bit access.
@@ -613,6 +621,9 @@ sha2_round(const unsigned char *data, sph_u32 r[8])
 #undef SHA2_IN
 }

+#endif   // SHA else
+
+
 /* see sph_sha2.h */
 void
 sph_sha224_init(void *cc)
@@ -653,7 +664,7 @@ void
 sph_sha224_close(void *cc, void *dst)
 {
 	sha224_close(cc, dst, 7);
-	sph_sha224_init(cc);
+//	sph_sha224_init(cc);
 }

 /* see sph_sha2.h */
@@ -661,7 +672,7 @@ void
 sph_sha224_addbits_and_close(void *cc, unsigned ub, unsigned n, void *dst)
 {
 	sha224_addbits_and_close(cc, ub, n, dst, 7);
-	sph_sha224_init(cc);
+//	sph_sha224_init(cc);
 }

 /* see sph_sha2.h */
@@ -677,14 +688,22 @@ void
 sph_sha256_addbits_and_close(void *cc, unsigned ub, unsigned n, void *dst)
 {
 	sha224_addbits_and_close(cc, ub, n, dst, 8);
-	sph_sha256_init(cc);
+//	sph_sha256_init(cc);
 }

-/* see sph_sha2.h */
-void
-sph_sha224_comp(const sph_u32 msg[16], sph_u32 val[8])
+void sph_sha256_full( void *dst, const void *data, size_t len )
 {
-#define SHA2_IN(x)   msg[x]
-	SHA2_ROUND_BODY(SHA2_IN, val);
-#undef SHA2_IN
-}
+   sph_sha256_context cc;
+   sph_sha256_init( &cc );
+   sph_sha256( &cc, data, len );
+   sph_sha256_close( &cc, dst );
+}   
+
+/* see sph_sha2.h */
+//void
+//sph_sha224_comp(const sph_u32 msg[16], sph_u32 val[8])
+//{
+//#define SHA2_IN(x)   msg[x]
+//	SHA2_ROUND_BODY(SHA2_IN, val);
+//#undef SHA2_IN
+//}
--- a/algo/sha/sph_sha2.h
+++ b/algo/sha/sph_sha2.h
@@ -73,7 +73,7 @@ typedef struct {
 	sph_u32 count_high, count_low;
 #endif
 #endif
-} sph_sha224_context;
+} sph_sha224_context __attribute__((aligned(64)));

 /**
 * This structure is a context for SHA-256 computations. It is identical
@@ -205,6 +205,10 @@ void sph_sha256_comp(const sph_u32 msg[16], sph_u32 val[8]);
 #define sph_sha256_comp   sph_sha224_comp
 #endif

+void sph_sha256_full( void *dst, const void *data, size_t len );
+
+
+
 #if SPH_64

 /**
--- a/algo/shavite/shavite-hash-2way.c
+++ b/algo/shavite/shavite-hash-2way.c
@@ -23,6 +23,19 @@ static const uint32_t IV512[] =
   _mm256_blend_epi32( mm256_ror128_32( a ), \
                       mm256_ror128_32( b ), 0x88 )

+#if defined(__VAES__)
+
+#define mm256_aesenc_2x128( x, k ) \
+   _mm256_aesenc_epi128( x, _mm256_castsi128_si256( k ) )
+
+#else
+
+#define mm256_aesenc_2x128( x, k ) \
+   mm256_concat_128( _mm_aesenc_si128( mm128_extr_hi128_256( x ), k ), \
+                     _mm_aesenc_si128( mm128_extr_lo128_256( x ), k ) )
+
+#endif
+
 static void
 c512_2way( shavite512_2way_context *ctx, const void *msg )
 {
@@ -304,7 +317,7 @@ void shavite512_2way_close( shavite512_2way_context *ctx, void *dst )
    uint32_t vp = ctx->ptr>>5;

    // Terminating byte then zero pad
-    casti_m256i( buf, vp++ ) = m256_const2_64( 0, 0x0000000000000080 );
+    casti_m256i( buf, vp++ ) = m256_const1_i128( 0x0000000000000080 );

    // Zero pad full vectors up to count
    for ( ; vp < 6; vp++ )      
@@ -384,13 +397,13 @@ void shavite512_2way_update_close( shavite512_2way_context *ctx, void *dst,

   if ( vp == 0 )    // empty buf, xevan.
   { 
-      casti_m256i( buf, 0 ) = m256_const2_64( 0, 0x0000000000000080 );
+      casti_m256i( buf, 0 ) = m256_const1_i128( 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + 1, 5 );
      ctx->count0 = ctx->count1 = ctx->count2 = ctx->count3 = 0;
   }
   else     // half full buf, everyone else.
   {
-    casti_m256i( buf, vp++ ) = m256_const2_64( 0, 0x0000000000000080 );
+    casti_m256i( buf, vp++ ) = m256_const1_i128( 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + vp, 6 - vp );
   }

@@ -474,13 +487,13 @@ void shavite512_2way_full( shavite512_2way_context *ctx, void *dst,

   if ( vp == 0 )    // empty buf, xevan.
   {
-      casti_m256i( buf, 0 ) = m256_const2_64( 0, 0x0000000000000080 );
+      casti_m256i( buf, 0 ) = m256_const1_i128( 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + 1, 5 );
      ctx->count0 = ctx->count1 = ctx->count2 = ctx->count3 = 0;
   }
   else     // half full buf, everyone else.
   {
-    casti_m256i( buf, vp++ ) = m256_const2_64( 0, 0x0000000000000080 );
+    casti_m256i( buf, vp++ ) = m256_const1_i128( 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + vp, 6 - vp );
   }

--- a/algo/shavite/shavite-hash-4way.c
+++ b/algo/shavite/shavite-hash-4way.c
@@ -292,7 +292,7 @@ void shavite512_4way_close( shavite512_4way_context *ctx, void *dst )
    uint32_t vp = ctx->ptr>>6;

    // Terminating byte then zero pad
-    casti_m512i( buf, vp++ ) = m512_const2_64( 0, 0x0000000000000080 );
+    casti_m512i( buf, vp++ ) = m512_const1_i128( 0x0000000000000080 );

    // Zero pad full vectors up to count
    for ( ; vp < 6; vp++ )      
@@ -372,13 +372,13 @@ void shavite512_4way_update_close( shavite512_4way_context *ctx, void *dst,

   if ( vp == 0 )    // empty buf, xevan.
   { 
-      casti_m512i( buf, 0 ) = m512_const2_64( 0, 0x0000000000000080 );
+      casti_m512i( buf, 0 ) = m512_const1_i128( 0x0000000000000080 );
      memset_zero_512( (__m512i*)buf + 1, 5 );
      ctx->count0 = ctx->count1 = ctx->count2 = ctx->count3 = 0;
   }
   else     // half full buf, everyone else.
   {
-    casti_m512i( buf, vp++ ) = m512_const2_64( 0, 0x0000000000000080 );
+    casti_m512i( buf, vp++ ) = m512_const1_i128( 0x0000000000000080 );
      memset_zero_512( (__m512i*)buf + vp, 6 - vp );
   }

@@ -463,13 +463,13 @@ void shavite512_4way_full( shavite512_4way_context *ctx, void *dst,

   if ( vp == 0 )    // empty buf, xevan.
   {
-      casti_m512i( buf, 0 ) = m512_const2_64( 0, 0x0000000000000080 );
+      casti_m512i( buf, 0 ) = m512_const1_i128( 0x0000000000000080 );
      memset_zero_512( (__m512i*)buf + 1, 5 );
      ctx->count0 = ctx->count1 = ctx->count2 = ctx->count3 = 0;
   }
   else     // half full buf, everyone else.
   {
-    casti_m512i( buf, vp++ ) = m512_const2_64( 0, 0x0000000000000080 );
+    casti_m512i( buf, vp++ ) = m512_const1_i128( 0x0000000000000080 );
      memset_zero_512( (__m512i*)buf + vp, 6 - vp );
   }

--- a/algo/skein/skein-4way.c
+++ b/algo/skein/skein-4way.c
@@ -2,14 +2,8 @@
 #include <string.h>
 #include <stdint.h>
 #include "skein-hash-4way.h"
-
-// 8 way is faster than SHA on Icelake
-// SHA is faster than 4 way on Ryzen
-//
-#if defined(__SHA__)
-  #include <openssl/sha.h>
-#endif
 #include "algo/sha/sha-hash-4way.h"
+#include "algo/sha/sph_sha2.h"

 #if defined (SKEIN_8WAY)

@@ -93,7 +87,7 @@ void skeinhash_4way( void *state, const void *input )
     uint32_t hash1[16] __attribute__ ((aligned (64)));
     uint32_t hash2[16] __attribute__ ((aligned (64)));
     uint32_t hash3[16] __attribute__ ((aligned (64)));
-     SHA256_CTX ctx_sha256;
+     sph_sha256_context ctx_sha256;
 #else
     uint32_t vhash32[16*4] __attribute__ ((aligned (64)));
     sha256_4way_context ctx_sha256;
@@ -102,31 +96,29 @@ void skeinhash_4way( void *state, const void *input )
     skein512_4way_final16( &ctx_skein, vhash64, input + (64*4) );

 #if defined(__SHA__)      
+
     dintrlv_4x64( hash0, hash1, hash2, hash3, vhash64, 512 );
-
-     SHA256_Init( &ctx_sha256 );
-     SHA256_Update( &ctx_sha256, (unsigned char*)hash0, 64 );
-     SHA256_Final( (unsigned char*)hash0, &ctx_sha256 );
-
-     SHA256_Init( &ctx_sha256 );
-     SHA256_Update( &ctx_sha256, (unsigned char*)hash1, 64 );
-     SHA256_Final( (unsigned char*)hash1, &ctx_sha256 );
-
-     SHA256_Init( &ctx_sha256 );
-     SHA256_Update( &ctx_sha256, (unsigned char*)hash2, 64 );
-     SHA256_Final( (unsigned char*)hash2, &ctx_sha256 );
-
-     SHA256_Init( &ctx_sha256 );
-     SHA256_Update( &ctx_sha256, (unsigned char*)hash3, 64 );
-     SHA256_Final( (unsigned char*)hash3, &ctx_sha256 );
-
+     sph_sha256_init( &ctx_sha256 );
+     sph_sha256( &ctx_sha256, hash0, 64 );
+     sph_sha256_close( &ctx_sha256, hash0 );
+     sph_sha256_init( &ctx_sha256 );
+     sph_sha256( &ctx_sha256, hash1, 64 );
+     sph_sha256_close( &ctx_sha256, hash1 );
+     sph_sha256_init( &ctx_sha256 );
+     sph_sha256( &ctx_sha256, hash2, 64 );
+     sph_sha256_close( &ctx_sha256, hash2 );
+     sph_sha256_init( &ctx_sha256 );
+     sph_sha256( &ctx_sha256, hash3, 64 );
+     sph_sha256_close( &ctx_sha256, hash3 );
     intrlv_4x32( state, hash0, hash1, hash2, hash3, 256 );
-#else
-     rintrlv_4x64_4x32( vhash32, vhash64, 512 );

+#else
+
+     rintrlv_4x64_4x32( vhash32, vhash64, 512 );
     sha256_4way_init( &ctx_sha256 );
     sha256_4way_update( &ctx_sha256, vhash32, 64 );
     sha256_4way_close( &ctx_sha256, state );
+
 #endif
 }

--- a/algo/skein/skein.c
+++ b/algo/skein/skein.c
@@ -5,21 +5,21 @@
 #include <string.h>
 #include <stdint.h>
 #include "sph_skein.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"

 void skeinhash(void *state, const void *input)
 {
     uint32_t hash[16] __attribute__ ((aligned (64)));
     sph_skein512_context ctx_skein;
-     SHA256_CTX           ctx_sha256;
+     sph_sha256_context   ctx_sha256;

     sph_skein512_init( &ctx_skein );
     sph_skein512( &ctx_skein, input, 80 );
     sph_skein512_close( &ctx_skein, hash );

-     SHA256_Init( &ctx_sha256 );
-     SHA256_Update( &ctx_sha256, (unsigned char*)hash, 64 );
-     SHA256_Final( (unsigned char*) hash, &ctx_sha256 );
+     sph_sha256_init( &ctx_sha256 );
+     sph_sha256( &ctx_sha256, hash, 64 );
+     sph_sha256_close( &ctx_sha256, hash );

     memcpy(state, hash, 32);
 }
--- a/algo/swifftx/stdbool.h
+++ b/algo/swifftx/stdbool.h
@@ -1,47 +0,0 @@
-/*
- * Copyright (c) 2000 Jeroen Ruigrok van der Werven <asmodai@FreeBSD.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- * $FreeBSD: src/include/stdbool.h,v 1.6 2002/08/16 07:33:14 alfred Exp $
- */
-
-#ifndef _STDBOOL_H_
-#define _STDBOOL_H_     
-
-#define __bool_true_false_are_defined   1
-
-#ifndef __cplusplus
-
-#define false   0
-#define true    1
-
-//#define bool    _Bool
-//#if __STDC_VERSION__ < 199901L && __GNUC__ < 3
-//typedef int     _Bool;
-//#endif
-typedef int bool;
-
-#endif /* !__cplusplus */
-
-#endif /* !_STDBOOL_H_ */
--- a/algo/swifftx/swifftx.c.bak
+++ b/algo/swifftx/swifftx.c.bak
--- a/algo/verthash/Verthash.c
+++ b/algo/verthash/Verthash.c
@@ -0,0 +1,741 @@
+/*
+ * Copyright 2018-2021 CryptoGraphics
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version. See LICENSE for more details.
+ */
+
+#include "algo-gate-api.h"
+#include "Verthash.h"
+#include "mm_malloc.h"
+
+//-----------------------------------------------------------------------------
+// Verthash info management
+int verthash_info_init(verthash_info_t* info, const char* file_name)
+{
+    // init fields to 0
+    info->fileName = NULL;
+    info->data = NULL;
+    info->dataSize = 0;
+    info->bitmask = 0;
+    size_t fileNameLen;
+
+    if ( !file_name || !( fileNameLen = strlen( file_name ) ) ) 
+    { 
+       applog( LOG_ERR, "Invalid file specification" );
+       return -1; 
+    }
+    
+    info->fileName = (char*)malloc( fileNameLen + 1 );
+    if ( !info->fileName )
+    {
+        applog( LOG_ERR, "Failed to allocate memory for Verthash data" );
+        return -1;
+    }
+
+    memset( info->fileName, 0, fileNameLen + 1 );
+    memcpy( info->fileName, file_name, fileNameLen );
+
+    FILE *fileMiningData = fopen_utf8( info->fileName, "rb" );
+    if ( !fileMiningData )
+    {
+       if ( opt_data_file || !opt_verify ) 
+       {
+          if ( opt_data_file )
+             applog( LOG_ERR, "Verthash data file not found or invalid: %s",
+                     info->fileName );
+          else
+          {
+             applog( LOG_ERR,
+                     "No Verthash data file specified and default not found");
+             applog( LOG_NOTICE,
+                     "Add '--verify' to create default 'verthash.dat'");
+          }
+          return -1;
+       }
+       else
+       {
+          applog( LOG_NOTICE, "Creating default 'verthash.dat' in current directory, this will take several minutes");
+          if ( verthash_generate_data_file( info->fileName ) )
+             return -1;
+
+          fileMiningData = fopen_utf8( info->fileName, "rb" );
+          if ( !fileMiningData )
+          {
+              applog( LOG_ERR, "File system error opening %s", info->fileName );
+              return -1;
+          }
+
+          applog( LOG_NOTICE, "Verthash data file created successfully" );
+       }
+    }
+
+    // Get file size
+    fseek(fileMiningData, 0, SEEK_END);
+    int fileSize = ftell(fileMiningData);
+    fseek(fileMiningData, 0, SEEK_SET);
+
+    if ( fileSize < 0 ) 
+    {
+        fclose(fileMiningData);
+        return 1;
+    }
+
+    // Allocate data
+    info->data = (uint8_t *)_mm_malloc( fileSize, 64 );
+    if (!info->data)
+    {
+        fclose(fileMiningData);
+        // Memory allocation fatal error.
+        return 2;
+    }
+
+    // Load data
+    if ( !fread( info->data, fileSize, 1, fileMiningData ) )
+    {
+        applog( LOG_ERR, "File system error reading %s", info->fileName );
+        fclose(fileMiningData);
+        return -1;
+    }
+
+    fclose(fileMiningData);
+
+    // Update fields
+    info->bitmask = ((fileSize - VH_HASH_OUT_SIZE)/VH_BYTE_ALIGNMENT) + 1;
+    info->dataSize = fileSize;
+
+    applog( LOG_NOTICE, "Using Verthash data file '%s'", info->fileName );
+    return 0;
+}
+
+//-----------------------------------------------------------------------------
+void verthash_info_free(verthash_info_t* info)
+{
+    free(info->fileName);
+    free(info->data);
+    info->dataSize = 0;
+    info->bitmask = 0;
+}
+
+
+//-----------------------------------------------------------------------------
+// Verthash hash
+#define VH_P0_SIZE 64
+#define VH_N_ITER 8 
+#define VH_N_SUBSET VH_P0_SIZE*VH_N_ITER
+#define VH_N_ROT 32
+#define VH_N_INDEXES 4096
+#define VH_BYTE_ALIGNMENT 16
+
+static inline uint32_t fnv1a(const uint32_t a, const uint32_t b)
+{
+    return (a ^ b) * 0x1000193;
+}
+
+#if 0
+static void rotate_indexes( uint32_t *p )
+{
+#if defined(__AVX2__)
+
+   for ( size_t x = 0; x < VH_N_SUBSET / sizeof(__m256i); x += 8 )
+   {
+      __m256i *px = (__m256i*)p + x;
+
+      px[0] = mm256_rol_32( px[0], 1 );
+      px[1] = mm256_rol_32( px[1], 1 );
+      px[2] = mm256_rol_32( px[2], 1 );
+      px[3] = mm256_rol_32( px[3], 1 );
+      px[4] = mm256_rol_32( px[4], 1 );
+      px[5] = mm256_rol_32( px[5], 1 );
+      px[6] = mm256_rol_32( px[6], 1 );
+      px[7] = mm256_rol_32( px[7], 1 );
+   }
+
+#else
+
+   for ( size_t x = 0; x < VH_N_SUBSET / sizeof(__m128i); x += 8 )
+   {
+      __m128i *px = (__m128i*)p0_index + x;
+
+      px[0] = mm128_rol_32( px[0], 1 );
+      px[1] = mm128_rol_32( px[1], 1 );
+      px[2] = mm128_rol_32( px[2], 1 );
+      px[3] = mm128_rol_32( px[3], 1 );
+      px[4] = mm128_rol_32( px[4], 1 );
+      px[5] = mm128_rol_32( px[5], 1 );
+      px[6] = mm128_rol_32( px[6], 1 );
+      px[7] = mm128_rol_32( px[7], 1 );
+   }
+
+#endif
+/*   
+   for ( size_t x = 0; x < VH_N_SUBSET / sizeof(uint32_t); ++x )
+      p[x] = ( p[x] << 1 ) | ( p[x] >> 31 );
+*/
+}
+#endif
+
+static inline uint32_t rotl32( uint32_t a, size_t r )
+{
+   return ( a << r ) | ( a >> (32-r) );
+}
+
+// Vectorized and targetted version of fnv1a
+#if defined (__AVX2__)        
+
+#define MULXOR \
+   *(__m256i*)hash = _mm256_mullo_epi32( _mm256_xor_si256( \
+                                 *(__m256i*)hash, *(__m256i*)blob_off ), k );
+
+#elif defined(__SSE41__)
+
+#define MULXOR \
+   casti_m128i( hash, 0 ) = _mm_mullo_epi32( _mm_xor_si128( \
+                  casti_m128i( hash, 0 ), casti_m128i( blob_off, 0 ) ), k ); \
+   casti_m128i( hash, 1 ) = _mm_mullo_epi32( _mm_xor_si128( \
+                  casti_m128i( hash, 1 ), casti_m128i( blob_off, 1 ) ), k );
+
+#else
+
+#define MULXOR \
+   for ( size_t j = 0; j < VH_HASH_OUT_SIZE / sizeof(uint32_t); j++ ) \
+       hash[j] = fnv1a( hash[j], blob_off[j] ); \
+
+#endif
+
+#define UPDATE_ACCUMULATOR \
+   accumulator = fnv1a( accumulator, blob_off[0] ); \
+   accumulator = fnv1a( accumulator, blob_off[1] ); \
+   accumulator = fnv1a( accumulator, blob_off[2] ); \
+   accumulator = fnv1a( accumulator, blob_off[3] ); \
+   accumulator = fnv1a( accumulator, blob_off[4] ); \
+   accumulator = fnv1a( accumulator, blob_off[5] ); \
+   accumulator = fnv1a( accumulator, blob_off[6] ); \
+   accumulator = fnv1a( accumulator, blob_off[7] )
+
+
+// first pass no rotate
+#define ROUND_0 \
+for ( size_t i = 0; i < VH_N_SUBSET / sizeof(uint32_t); i++ ) \
+{ \
+   const uint32_t *blob_off = blob + \
+                         ( ( fnv1a( subset[i], accumulator ) % mdiv ) \
+                         * ( VH_BYTE_ALIGNMENT / sizeof(uint32_t) ) ); \
+   UPDATE_ACCUMULATOR; \
+   MULXOR; \
+}
+
+// subsequent passes rotate by r on demand, no need for mass rotate
+#define ROUND_r( r ) \
+for ( size_t i = 0; i < VH_N_SUBSET / sizeof(uint32_t); i++ ) \
+{ \
+   const uint32_t *blob_off = blob + \
+                 ( ( fnv1a( rotl32( subset[i], r ), accumulator ) % mdiv ) \
+                 * ( VH_BYTE_ALIGNMENT / sizeof(uint32_t) ) ); \
+   UPDATE_ACCUMULATOR; \
+   MULXOR; \
+}
+
+void verthash_hash( const void *blob_bytes, const size_t blob_size,
+                    const void *input, void *output )
+{
+    uint32_t hash[ VH_HASH_OUT_SIZE / 4 ] __attribute__ ((aligned (64)));
+    uint32_t subset[ VH_N_SUBSET / 4 ] __attribute__ ((aligned (64)));
+    const uint32_t *blob = (const uint32_t*)blob_bytes;
+    uint32_t accumulator = 0x811c9dc5;
+    const uint32_t mdiv = ( ( blob_size - VH_HASH_OUT_SIZE )
+                             / VH_BYTE_ALIGNMENT ) + 1;
+#if defined (__AVX2__)        
+    const __m256i k = _mm256_set1_epi32( 0x1000193 );
+#elif defined(__SSE41__)
+    const __m128i k = _mm_set1_epi32( 0x1000193 );
+#endif
+    
+    sha3( input, VH_HEADER_SIZE, hash, VH_HASH_OUT_SIZE );
+    verthash_sha3_512_final_8( subset, ( (uint64_t*)input )[ 9 ] );
+
+    ROUND_0;
+    for ( size_t r = 1; r < VH_N_ROT; ++r )
+       ROUND_r( r );
+
+    memcpy( output, hash, VH_HASH_OUT_SIZE );
+}
+
+//-----------------------------------------------------------------------------
+// Verthash data file generator
+
+#define NODE_SIZE 32
+
+struct Graph
+{
+    FILE *db;
+    int64_t log2;
+    int64_t pow2;
+    uint8_t *pk;
+    int64_t index;
+};
+
+int64_t Log2(int64_t x)
+{
+    int64_t r = 0;
+    for (; x > 1; x >>= 1)
+    {
+        r++;
+    }
+
+    return r;
+}
+
+int64_t bfsToPost(struct Graph *g, const int64_t node)
+{
+    return node & ~g->pow2;
+}
+
+int64_t numXi(int64_t index)
+{
+    return (1 << ((uint64_t)index)) * (index + 1) * index;
+}
+
+void WriteId(struct Graph *g, uint8_t *Node, const int64_t id)
+{
+    fseek(g->db, id * NODE_SIZE, SEEK_SET);
+    fwrite(Node, 1, NODE_SIZE, g->db);
+}
+
+void WriteNode(struct Graph *g, uint8_t *Node, const int64_t id)
+{
+    const int64_t idx = bfsToPost(g, id);
+    WriteId(g, Node, idx);
+}
+
+void NewNode(struct Graph *g, const int64_t id, uint8_t *hash)
+{
+    WriteNode(g, hash, id);
+}
+
+uint8_t *GetId(struct Graph *g, const int64_t id)
+{
+    fseek(g->db, id * NODE_SIZE, SEEK_SET);
+    uint8_t *node = (uint8_t *)malloc(NODE_SIZE);
+    const size_t bytes_read = fread(node, 1, NODE_SIZE, g->db);
+    if(bytes_read != NODE_SIZE) {
+        return NULL;
+    }
+    return node;
+}
+
+uint8_t *GetNode(struct Graph *g, const int64_t id)
+{
+    const int64_t idx = bfsToPost(g, id);
+    return GetId(g, idx);
+}
+
+uint32_t WriteVarInt(uint8_t *buffer, int64_t val)
+{
+    memset(buffer, 0, NODE_SIZE);
+    uint64_t uval = ((uint64_t)(val)) << 1;
+    if (val < 0)
+    {
+        uval = ~uval;
+    }
+    uint32_t i = 0;
+    while (uval >= 0x80)
+    {
+        buffer[i] = (uint8_t)uval | 0x80;
+        uval >>= 7;
+        i++;
+    }
+    buffer[i] = (uint8_t)uval;
+    return i;
+}
+
+void ButterflyGraph(struct Graph *g, int64_t index, int64_t *count)
+{
+    if (index == 0)
+    {
+        index = 1;
+    }
+
+    int64_t numLevel = 2 * index;
+    int64_t perLevel = (int64_t)(1 << (uint64_t)index);
+    int64_t begin = *count - perLevel;
+    int64_t level, i;
+
+    for (level = 1; level < numLevel; level++)
+    {
+        for (i = 0; i < perLevel; i++)
+        {
+            int64_t prev;
+            int64_t shift = index - level;
+            if (level > numLevel / 2)
+            {
+                shift = level - numLevel / 2;
+            }
+            if (((i >> (uint64_t)shift) & 1) == 0)
+            {
+                prev = i + (1 << (uint64_t)shift);
+            }
+            else
+            {
+                prev = i - (1 << (uint64_t)shift);
+            }
+
+            uint8_t *parent0 = GetNode(g, begin + (level - 1) * perLevel + prev);
+            uint8_t *parent1 = GetNode(g, *count - perLevel);
+            uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+            WriteVarInt(buf, *count);
+            uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 4);
+            memcpy(hashInput, g->pk, NODE_SIZE);
+            memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+            memcpy(hashInput + (NODE_SIZE * 2), parent0, NODE_SIZE);
+            memcpy(hashInput + (NODE_SIZE * 3), parent1, NODE_SIZE);
+
+            uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+            sha3(hashInput, NODE_SIZE * 4, hashOutput, NODE_SIZE);
+
+            NewNode(g, *count, hashOutput);
+            (*count)++;
+
+            free(hashOutput);
+            free(hashInput);
+            free(parent0);
+            free(parent1);
+            free(buf);
+        }
+    }
+}
+
+void XiGraphIter(struct Graph *g, int64_t index)
+{
+    int64_t count = g->pow2;
+
+    int8_t stackSize = 5;
+    int64_t *stack = (int64_t *)malloc(sizeof(int64_t) * stackSize);
+    for (int i = 0; i < 5; i++)
+        stack[i] = index;
+
+    int8_t graphStackSize = 5;
+    int32_t *graphStack = (int32_t *)malloc(sizeof(int32_t) * graphStackSize);
+    for (int i = 0; i < 5; i++)
+        graphStack[i] = graphStackSize - i - 1;
+
+    int64_t i = 0;
+    int64_t graph = 0;
+    int64_t pow2index = 1 << ((uint64_t)index);
+
+    for (i = 0; i < pow2index; i++)
+    {
+        uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+        WriteVarInt(buf, count);
+        uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 2);
+        memcpy(hashInput, g->pk, NODE_SIZE);
+        memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+        uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+
+        sha3(hashInput, NODE_SIZE * 2, hashOutput, NODE_SIZE);
+        NewNode(g, count, hashOutput);
+        count++;
+
+        free(hashOutput);
+        free(hashInput);
+        free(buf);
+    }
+
+    if (index == 1)
+    {
+        ButterflyGraph(g, index, &count);
+        return;
+    }
+
+    while (stackSize != 0 && graphStackSize != 0)
+    {
+
+        index = stack[stackSize - 1];
+        graph = graphStack[graphStackSize - 1];
+
+        stackSize--;
+        if (stackSize > 0)
+        {
+            int64_t *tempStack = (int64_t *)malloc(sizeof(int64_t) * (stackSize));
+            memcpy(tempStack, stack, sizeof(int64_t) * (stackSize));
+            free(stack);
+            stack = tempStack;
+        }
+
+        graphStackSize--;
+        if (graphStackSize > 0)
+        {
+            int32_t *tempGraphStack = (int32_t *)malloc(sizeof(int32_t) * (graphStackSize));
+            memcpy(tempGraphStack, graphStack, sizeof(int32_t) * (graphStackSize));
+            free(graphStack);
+            graphStack = tempGraphStack;
+        }
+
+        int8_t indicesSize = 5;
+        int64_t *indices = (int64_t *)malloc(sizeof(int64_t) * indicesSize);
+        for (int i = 0; i < indicesSize; i++)
+            indices[i] = index - 1;
+
+        int8_t graphsSize = 5;
+        int32_t *graphs = (int32_t *)malloc(sizeof(int32_t) * graphsSize);
+        for (int i = 0; i < graphsSize; i++)
+            graphs[i] = graphsSize - i - 1;
+
+        int64_t pow2indexInner = 1 << ((uint64_t)index);
+        int64_t pow2indexInner_1 = 1 << ((uint64_t)index - 1);
+
+        if (graph == 0)
+        {
+            uint64_t sources = count - pow2indexInner;
+            for (i = 0; i < pow2indexInner_1; i++)
+            {
+                uint8_t *parent0 = GetNode(g, sources + i);
+                uint8_t *parent1 = GetNode(g, sources + i + pow2indexInner_1);
+
+                uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+                WriteVarInt(buf, count);
+
+                uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 4);
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent0, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 3), parent1, NODE_SIZE);
+
+                uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 4, hashOutput, NODE_SIZE);
+
+                NewNode(g, count, hashOutput);
+                count++;
+
+                free(hashOutput);
+                free(hashInput);
+                free(parent0);
+                free(parent1);
+                free(buf);
+            }
+        }
+        else if (graph == 1)
+        {
+            uint64_t firstXi = count;
+            for (i = 0; i < pow2indexInner_1; i++)
+            {
+                uint64_t nodeId = firstXi + i;
+                uint8_t *parent = GetNode(g, firstXi - pow2indexInner_1 + i);
+
+                uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+                WriteVarInt(buf, nodeId);
+
+                uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 3);
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent, NODE_SIZE);
+
+                uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 3, hashOutput, NODE_SIZE);
+
+                NewNode(g, count, hashOutput);
+                count++;
+
+                free(hashOutput);
+                free(hashInput);
+                free(parent);
+                free(buf);
+            }
+        }
+        else if (graph == 2)
+        {
+            uint64_t secondXi = count;
+            for (i = 0; i < pow2indexInner_1; i++)
+            {
+                uint64_t nodeId = secondXi + i;
+                uint8_t *parent = GetNode(g, secondXi - pow2indexInner_1 + i);
+
+                uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+                WriteVarInt(buf, nodeId);
+
+                uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 3);
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent, NODE_SIZE);
+
+                uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 3, hashOutput, NODE_SIZE);
+
+                NewNode(g, count, hashOutput);
+                count++;
+
+                free(hashOutput);
+                free(hashInput);
+                free(parent);
+                free(buf);
+            }
+        }
+        else if (graph == 3)
+        {
+            uint64_t secondButter = count;
+            for (i = 0; i < pow2indexInner_1; i++)
+            {
+                uint64_t nodeId = secondButter + i;
+                uint8_t *parent = GetNode(g, secondButter - pow2indexInner_1 + i);
+
+                uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+                WriteVarInt(buf, nodeId);
+
+                uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 3);
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent, NODE_SIZE);
+
+                uint8_t *hashOutput = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 3, hashOutput, NODE_SIZE);
+
+                NewNode(g, count, hashOutput);
+                count++;
+
+                free(hashOutput);
+                free(hashInput);
+                free(parent);
+                free(buf);
+            }
+        }
+        else
+        {
+            uint64_t sinks = count;
+            uint64_t sources = sinks + pow2indexInner - numXi(index);
+            for (i = 0; i < pow2indexInner_1; i++)
+            {
+                uint64_t nodeId0 = sinks + i;
+                uint64_t nodeId1 = sinks + i + pow2indexInner_1;
+                uint8_t *parent0 = GetNode(g, sinks - pow2indexInner_1 + i);
+                uint8_t *parent1_0 = GetNode(g, sources + i);
+                uint8_t *parent1_1 = GetNode(g, sources + i + pow2indexInner_1);
+
+                uint8_t *buf = (uint8_t *)malloc(NODE_SIZE);
+                WriteVarInt(buf, nodeId0);
+
+                uint8_t *hashInput = (uint8_t *)malloc(NODE_SIZE * 4);
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent0, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 3), parent1_0, NODE_SIZE);
+
+                uint8_t *hashOutput0 = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 4, hashOutput0, NODE_SIZE);
+
+                WriteVarInt(buf, nodeId1);
+
+                memcpy(hashInput, g->pk, NODE_SIZE);
+                memcpy(hashInput + NODE_SIZE, buf, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 2), parent0, NODE_SIZE);
+                memcpy(hashInput + (NODE_SIZE * 3), parent1_1, NODE_SIZE);
+
+                uint8_t *hashOutput1 = (uint8_t *)malloc(NODE_SIZE);
+                sha3(hashInput, NODE_SIZE * 4, hashOutput1, NODE_SIZE);
+
+                NewNode(g, nodeId0, hashOutput0);
+                NewNode(g, nodeId1, hashOutput1);
+                count += 2;
+
+                free(parent0);
+                free(parent1_0);
+                free(parent1_1);
+                free(buf);
+                free(hashInput);
+                free(hashOutput0);
+                free(hashOutput1);
+            }
+        }
+
+        if ((graph == 0 || graph == 3) ||
+            ((graph == 1 || graph == 2) && index == 2))
+        {
+            ButterflyGraph(g, index - 1, &count);
+        }
+        else if (graph == 1 || graph == 2)
+        {
+
+            int64_t *tempStack = (int64_t *)malloc(sizeof(int64_t) * (stackSize + indicesSize));
+            memcpy(tempStack, stack, stackSize * sizeof(int64_t));
+            memcpy(tempStack + stackSize, indices, indicesSize * sizeof(int64_t));
+            stackSize += indicesSize;
+            free(stack);
+            stack = tempStack;
+
+            int32_t *tempGraphStack = (int32_t *)malloc(sizeof(int32_t) * (graphStackSize + graphsSize));
+            memcpy(tempGraphStack, graphStack, graphStackSize * sizeof(int32_t));
+            memcpy(tempGraphStack + graphStackSize, graphs, graphsSize * sizeof(int32_t));
+            graphStackSize += graphsSize;
+            free(graphStack);
+            graphStack = tempGraphStack;
+        }
+
+        free(indices);
+        free(graphs);
+    }
+
+    free(stack);
+    free(graphStack);
+}
+
+struct Graph *NewGraph(int64_t index, const char* targetFile, uint8_t *pk)
+{
+    uint8_t exists = 0;
+    FILE *db;
+    if ((db = fopen_utf8(targetFile, "r")) != NULL)
+    {
+        fclose(db);
+        exists = 1;
+    }
+
+    db = fopen_utf8(targetFile, "wb+");
+    int64_t size = numXi(index);
+    int64_t log2 = Log2(size) + 1;
+    int64_t pow2 = 1 << ((uint64_t)log2);
+
+    struct Graph *g = (struct Graph *)malloc(sizeof(struct Graph));
+
+    if ( !g ) return NULL;
+
+    g->db = db;
+    g->log2 = log2;
+    g->pow2 = pow2;
+    g->pk = pk;
+    g->index = index;
+
+    if (exists == 0)
+    {
+        XiGraphIter(g, index);
+    }
+
+    fclose(db);
+    return g;
+}
+
+//-----------------------------------------------------------------------------
+
+// use info for _mm_malloc, then verify file
+int verthash_generate_data_file(const char* output_file_name)
+{
+    const char *hashInput = "Verthash Proof-of-Space Datafile";
+    uint8_t *pk = (uint8_t*)malloc( NODE_SIZE );
+    
+    if ( !pk )
+    {
+      applog( LOG_ERR, "Verthash data memory allocation failed");
+      return -1;
+    }
+
+    sha3( hashInput, 32, pk, NODE_SIZE );
+
+    int64_t index = 17;
+    if ( !NewGraph( index, output_file_name, pk ) )
+    {
+       applog( LOG_ERR, "Verthash file creation failed");
+       return -1;
+    }
+
+    return 0;
+}
+
--- a/algo/verthash/Verthash.h
+++ b/algo/verthash/Verthash.h
@@ -0,0 +1,57 @@
+/*
+ * Copyright 2018-2021 CryptoGraphics
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version. See LICENSE for more details.
+ */
+
+#ifndef Verthash_INCLUDE_ONCE
+#define Verthash_INCLUDE_ONCE
+
+#include "tiny_sha3/sha3.h"
+#include "fopen_utf8.h"
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdint.h>
+#include <string.h>
+
+// Verthash constants used to compute bitmask, used inside kernel during IO pass
+#define VH_HASH_OUT_SIZE 32
+#define VH_BYTE_ALIGNMENT 16
+#define VH_HEADER_SIZE 80
+
+//-----------------------------------------------------------------------------
+// Verthash data
+//! Verthash C api for data maniputation.
+typedef struct VerthashInfo
+{
+    char* fileName;
+    uint8_t* data;
+    uint64_t dataSize;
+    uint32_t bitmask;
+} verthash_info_t;
+
+//! Must be called before usage. Reset all fields and set a mining data file name.
+//! Error codes
+//! 0 - Success(No error).
+//! 1 - File name is invalid.
+//! 2 - Memory allocation error
+int verthash_info_init(verthash_info_t* info, const char* file_name);
+
+//! Reset all fields and free allocated data.
+void verthash_info_free(verthash_info_t* info);
+
+//! Generate verthash data file and save it to specified location.
+int verthash_generate_data_file(const char* output_file_name);
+
+void verthash_hash( const void *blob_bytes, const size_t blob_size,
+                    const void *input, void *output );
+
+void verthash_sha3_512_prehash_72( const void *input );
+void verthash_sha3_512_final_8( void *hash, const uint64_t nonce );
+
+#endif // !Verthash_INCLUDE_ONCE
+
--- a/algo/verthash/fopen_utf8.c
+++ b/algo/verthash/fopen_utf8.c
@@ -0,0 +1,181 @@
+#ifndef H_FOPEN_UTF8
+#define H_FOPEN_UTF8
+
+#include "fopen_utf8.h"
+#include <stdint.h>
+#include <stddef.h>
+#include <stdlib.h>
+#include <stdio.h>
+
+int utf8_char_size(const uint8_t *c)
+{
+	const uint8_t	m0x	= 0x80, c0x	= 0x00,
+	      		m10x	= 0xC0, c10x	= 0x80,
+	      		m110x	= 0xE0, c110x	= 0xC0,
+	      		m1110x	= 0xF0, c1110x	= 0xE0,
+	      		m11110x	= 0xF8, c11110x	= 0xF0;
+
+	if ((c[0] & m0x) == c0x)
+		return 1;
+
+	if ((c[0] & m110x) == c110x)
+	if ((c[1] & m10x) == c10x)
+		return 2;
+
+	if ((c[0] & m1110x) == c1110x)
+	if ((c[1] & m10x) == c10x)
+	if ((c[2] & m10x) == c10x)
+		return 3;
+
+	if ((c[0] & m11110x) == c11110x)
+	if ((c[1] & m10x) == c10x)
+	if ((c[2] & m10x) == c10x)
+	if ((c[3] & m10x) == c10x)
+		return 4;
+
+	if ((c[0] & m10x) == c10x)	// not a first UTF-8 byte
+		return 0;
+
+	return -1;			// if c[0] is a first byte but the other bytes don't match
+}
+
+uint32_t utf8_to_unicode32(const uint8_t *c, size_t *index)
+{
+	uint32_t v;
+	int size;
+	const uint8_t m6 = 63, m5 = 31, m4 = 15, m3 = 7;
+
+	if (c==NULL)
+		return 0;
+
+	size = utf8_char_size(c);
+
+	if (size > 0 && index)
+		*index += size-1;
+
+	switch (size)
+	{
+		case 1:
+			v = c[0];
+			break;
+		case 2:
+			v = c[0] & m5;
+			v = v << 6 | (c[1] & m6);
+			break;
+		case 3:
+			v = c[0] & m4;
+			v = v << 6 | (c[1] & m6);
+			v = v << 6 | (c[2] & m6);
+			break;
+		case 4:
+			v = c[0] & m3;
+			v = v << 6 | (c[1] & m6);
+			v = v << 6 | (c[2] & m6);
+			v = v << 6 | (c[3] & m6);
+			break;
+		case 0:				// not a first UTF-8 byte
+		case -1:			// corrupt UTF-8 letter
+		default:
+			v = -1;
+			break;
+	}
+
+	return v;
+}
+
+int codepoint_utf16_size(uint32_t c)
+{
+	if (c < 0x10000) return 1;
+	if (c < 0x110000) return 2;
+
+	return 0;
+}
+
+uint16_t *sprint_utf16(uint16_t *str, uint32_t c)	// str must be able to hold 1 to 3 entries and will be null-terminated by this function
+{
+	int c_size;
+
+	if (str==NULL)
+		return NULL;
+
+	c_size = codepoint_utf16_size(c);
+
+	switch (c_size)
+	{
+		case 1:
+			str[0] = c;
+			if (c > 0)
+				str[1] = '\0';
+			break;
+
+		case 2:
+			c -= 0x10000;
+			str[0] = 0xD800 + (c >> 10);
+			str[1] = 0xDC00 + (c & 0x3FF);
+			str[2] = '\0';
+			break;
+
+		default:
+			str[0] = '\0';
+	}
+
+	return str;
+}
+
+size_t strlen_utf8_to_utf16(const uint8_t *str)
+{
+	size_t i, count;
+	uint32_t c;
+
+	for (i=0, count=0; ; i++)
+	{
+		if (str[i]==0)
+			return count;
+
+		c = utf8_to_unicode32(&str[i], &i);
+		count += codepoint_utf16_size(c);
+	}
+}
+
+uint16_t *utf8_to_utf16(const uint8_t *utf8, uint16_t *utf16)
+{
+	size_t i, j;
+	uint32_t c;
+
+	if (utf8==NULL)
+		return NULL;
+
+	if (utf16==NULL)
+		utf16 = (uint16_t *) calloc(strlen_utf8_to_utf16(utf8) + 1, sizeof(uint16_t));
+
+	for (i=0, j=0, c=1; c; i++)
+	{
+		c = utf8_to_unicode32(&utf8[i], &i);
+		sprint_utf16(&utf16[j], c);
+		j += codepoint_utf16_size(c);
+	}
+
+	return utf16;
+}
+
+FILE *fopen_utf8(const char *path, const char *mode)
+{
+	#ifdef _WIN32
+	wchar_t *wpath, wmode[8];
+	FILE *file;
+
+	if (utf8_to_utf16((const uint8_t *) mode, (uint16_t *) wmode)==NULL)
+		return NULL;
+
+	wpath = (wchar_t *) utf8_to_utf16((const uint8_t *) path, NULL);
+	if (wpath==NULL)
+		return NULL;
+
+	file = _wfopen(wpath, wmode);
+	free(wpath);
+	return file;
+	#else
+	return fopen(path, mode);
+	#endif
+}
+#endif
--- a/algo/verthash/fopen_utf8.h
+++ b/algo/verthash/fopen_utf8.h
@@ -0,0 +1,25 @@
+#ifndef H_FOPEN_UTF8
+#define H_FOPEN_UTF8
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <stdint.h>
+#include <stddef.h>
+
+int utf8_char_size(const uint8_t *c);
+uint32_t utf8_to_unicode32(const uint8_t *c, size_t *index);
+int codepoint_utf16_size(uint32_t c);
+uint16_t *sprint_utf16(uint16_t *str, uint32_t c);
+size_t strlen_utf8_to_utf16(const uint8_t *str);
+uint16_t *utf8_to_utf16(const uint8_t *utf8, uint16_t *utf16);
+
+FILE *fopen_utf8(const char *path, const char *mode);
+
+#ifdef __cplusplus
+}
+#endif
+#endif
+
--- a/algo/verthash/tiny_sha3/sha3-4way.c
+++ b/algo/verthash/tiny_sha3/sha3-4way.c
@@ -0,0 +1,301 @@
+#if defined(__AVX2__)
+
+// sha3-4way.c
+// 19-Nov-11  Markku-Juhani O. Saarinen <mjos@iki.fi>
+// vectorization by JayDDee 2021-03-27
+//
+// Revised 07-Aug-15 to match with official release of FIPS PUB 202 "SHA3"
+// Revised 03-Sep-15 for portability + OpenSSL - style API
+
+#include "sha3-4way.h"
+
+// constants
+static const uint64_t keccakf_rndc[24] = {
+        0x0000000000000001, 0x0000000000008082, 0x800000000000808a,
+        0x8000000080008000, 0x000000000000808b, 0x0000000080000001,
+        0x8000000080008081, 0x8000000000008009, 0x000000000000008a,
+        0x0000000000000088, 0x0000000080008009, 0x000000008000000a,
+        0x000000008000808b, 0x800000000000008b, 0x8000000000008089,
+        0x8000000000008003, 0x8000000000008002, 0x8000000000000080,
+        0x000000000000800a, 0x800000008000000a, 0x8000000080008081,
+        0x8000000000008080, 0x0000000080000001, 0x8000000080008008
+    };
+
+void sha3_4way_keccakf( __m256i st[25] )
+{
+   int i, j, r;
+   __m256i t, bc[5];
+
+   for ( r = 0; r < KECCAKF_ROUNDS; r++ )
+   {
+      // Theta
+      bc[0] = _mm256_xor_si256( st[0],
+                           mm256_xor4( st[5], st[10], st[15], st[20] ) );
+      bc[1] = _mm256_xor_si256( st[1],
+                           mm256_xor4( st[6], st[11], st[16], st[21] ) );
+      bc[2] = _mm256_xor_si256( st[2],
+                           mm256_xor4( st[7], st[12], st[17], st[22] ) );
+      bc[3] = _mm256_xor_si256( st[3],
+                           mm256_xor4( st[8], st[13], st[18], st[23] ) );
+      bc[4] = _mm256_xor_si256( st[4],
+                           mm256_xor4( st[9], st[14], st[19], st[24] ) );
+
+      for ( i = 0; i < 5; i++ )
+      {
+         t = _mm256_xor_si256( bc[ (i+4) % 5 ],
+                               mm256_rol_64( bc[ (i+1) % 5 ], 1 ) );
+         st[ i    ]  = _mm256_xor_si256( st[ i    ],  t );
+         st[ i+5  ]  = _mm256_xor_si256( st[ i+ 5 ],  t );
+         st[ i+10 ]  = _mm256_xor_si256( st[ i+10 ],  t );
+         st[ i+15 ]  = _mm256_xor_si256( st[ i+15 ],  t );
+         st[ i+20 ]  = _mm256_xor_si256( st[ i+20 ],  t );
+      }
+
+      // Rho Pi
+#define RHO_PI( i, c ) \
+   bc[0] = st[ i ]; \
+   st[ i ] = mm256_rol_64( t, c ); \
+   t = bc[0]
+
+      t = st[1];
+
+      RHO_PI( 10,  1 );
+      RHO_PI(  7,  3 );
+      RHO_PI( 11,  6 );
+      RHO_PI( 17, 10 );
+      RHO_PI( 18, 15 );
+      RHO_PI(  3, 21 );
+      RHO_PI(  5, 28 );
+      RHO_PI( 16, 36 );
+      RHO_PI(  8, 45 );
+      RHO_PI( 21, 55 );
+      RHO_PI( 24,  2 );
+      RHO_PI(  4, 14 );
+      RHO_PI( 15, 27 );
+      RHO_PI( 23, 41 );
+      RHO_PI( 19, 56 );
+      RHO_PI( 13,  8 );
+      RHO_PI( 12, 25 );
+      RHO_PI(  2, 43 );
+      RHO_PI( 20, 62 );
+      RHO_PI( 14, 18 );
+      RHO_PI( 22, 39 );
+      RHO_PI(  9, 61 );
+      RHO_PI(  6, 20 );
+      RHO_PI(  1, 44 );
+
+#undef RHO_PI        
+
+      //  Chi
+      for ( j = 0; j < 25; j += 5 )
+      {
+         memcpy( bc, &st[ j ], 5*32 );
+         st[ j   ] = _mm256_xor_si256( st[ j   ],
+                                       _mm256_andnot_si256( bc[1], bc[2] ) );
+         st[ j+1 ] = _mm256_xor_si256( st[ j+1 ],
+                                       _mm256_andnot_si256( bc[2], bc[3] ) );
+         st[ j+2 ] = _mm256_xor_si256( st[ j+2 ],
+                                       _mm256_andnot_si256( bc[3], bc[4] ) );
+         st[ j+3 ] = _mm256_xor_si256( st[ j+3 ],
+                                       _mm256_andnot_si256( bc[4], bc[0] ) );
+         st[ j+4 ] = _mm256_xor_si256( st[ j+4 ],
+                                       _mm256_andnot_si256( bc[0], bc[1] ) );
+      }
+
+      //  Iota
+      st[0] = _mm256_xor_si256( st[0],
+                                _mm256_set1_epi64x( keccakf_rndc[ r ] ) );
+   }
+}
+
+int sha3_4way_init( sha3_4way_ctx_t *c, int mdlen )
+{
+    for ( int i = 0; i < 25; i++ )  c->st[ i ] = m256_zero;
+    c->mdlen = mdlen;
+    c->rsiz = 200 - 2 * mdlen;
+    c->pt = 0;
+    return 1;
+}
+
+int sha3_4way_update( sha3_4way_ctx_t *c, const void *data, size_t len )
+{
+    size_t i;
+    int j =  c->pt;
+    const int rsiz = c->rsiz / 8;
+    const int l = len / 8;
+
+    for ( i = 0; i < l; i++ )
+    {
+        c->st[ j ] = _mm256_xor_si256( c->st[ j ],
+                                       ( (const __m256i*)data )[i] );
+        j++;
+        if ( j >= rsiz )
+        {
+            sha3_4way_keccakf( c->st );
+            j = 0;
+        }
+    }
+    c->pt = j;
+
+    return 1;
+}
+
+int sha3_4way_final( void *md, sha3_4way_ctx_t *c )
+{
+    c->st[ c->pt ] = _mm256_xor_si256( c->st[ c->pt ],
+                                       m256_const1_64( 6 ) );
+    c->st[ c->rsiz / 8 - 1 ] =
+                       _mm256_xor_si256( c->st[ c->rsiz / 8 - 1 ],
+                                         m256_const1_64( 0x8000000000000000 ) );
+    sha3_4way_keccakf( c->st );
+    memcpy( md, c->st, c->mdlen * 4 );
+    return 1;
+}
+
+void *sha3_4way( const void *in, size_t inlen, void *md, int mdlen )
+{
+    sha3_4way_ctx_t ctx;
+    sha3_4way_init( &ctx, mdlen);
+    sha3_4way_update( &ctx, in, inlen );
+    sha3_4way_final( md, &ctx );
+    return md;
+}
+
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
+void sha3_8way_keccakf( __m512i st[25] )
+{
+    int i, j, r;
+    __m512i t, bc[5];
+
+    // actual iteration
+    for ( r = 0; r < KECCAKF_ROUNDS; r++ )
+    {
+
+        // Theta
+        for ( i = 0; i < 5; i++ )
+           bc[i] = _mm512_xor_si512( st[i], 
+              mm512_xor4( st[ i+5 ], st[ i+10 ], st[ i+15 ], st[i+20 ] ) );
+
+        for ( i = 0; i < 5; i++ )
+        {
+            t = _mm512_xor_si512( bc[(i + 4) % 5],
+                                  _mm512_rol_epi64( bc[(i + 1) % 5], 1 ) );
+            for ( j = 0; j < 25; j += 5 )
+                st[j + i]  = _mm512_xor_si512( st[j + i],  t );
+        }
+
+        // Rho Pi
+#define RHO_PI( i, c ) \
+   bc[0] = st[ i ]; \
+   st[ i ] = _mm512_rol_epi64( t, c ); \
+   t = bc[0]
+
+        t = st[1];
+
+        RHO_PI( 10,  1 );        
+        RHO_PI(  7,  3 );
+        RHO_PI( 11,  6 );
+        RHO_PI( 17, 10 );
+        RHO_PI( 18, 15 );
+        RHO_PI(  3, 21 );
+        RHO_PI(  5, 28 );
+        RHO_PI( 16, 36 );
+        RHO_PI(  8, 45 );
+        RHO_PI( 21, 55 );
+        RHO_PI( 24,  2 );
+        RHO_PI(  4, 14 );
+        RHO_PI( 15, 27 );
+        RHO_PI( 23, 41 );
+        RHO_PI( 19, 56 );
+        RHO_PI( 13,  8 );
+        RHO_PI( 12, 25 );
+        RHO_PI(  2, 43 );
+        RHO_PI( 20, 62 );
+        RHO_PI( 14, 18 );
+        RHO_PI( 22, 39 );
+        RHO_PI(  9, 61 );
+        RHO_PI(  6, 20 );
+        RHO_PI(  1, 44 );
+
+#undef RHO_PI        
+
+        //  Chi
+        for ( j = 0; j < 25; j += 5 )
+        {
+            for ( i = 0; i < 5; i++ )
+                bc[i] = st[j + i];
+            for ( i = 0; i < 5; i++ )
+                st[ j+i ] = _mm512_xor_si512(  st[ j+i ],  _mm512_andnot_si512(
+                                         bc[ (i+1) % 5 ], bc[ (i+2) % 5 ] ) );
+        }
+
+        //  Iota
+        st[0] = _mm512_xor_si512( st[0], _mm512_set1_epi64( keccakf_rndc[r] ) );
+    }
+}
+
+// Initialize the context for SHA3
+
+int sha3_8way_init( sha3_8way_ctx_t *c, int mdlen )
+{
+    for ( int i = 0; i < 25; i++ )  c->st[ i ] = m512_zero;
+    c->mdlen = mdlen;
+    c->rsiz = 200 - 2 * mdlen;
+    c->pt = 0;
+    return 1;
+}
+
+// update state with more data
+
+int sha3_8way_update( sha3_8way_ctx_t *c, const void *data, size_t len )
+{
+    size_t i;
+    int j =  c->pt;
+    const int rsiz = c->rsiz / 8;
+    const int l = len / 8;
+
+    for ( i = 0; i < l; i++ )
+    {
+        c->st[ j ] = _mm512_xor_si512( c->st[ j ],
+                                        ( (const __m512i*)data )[i] );
+        j++;
+        if ( j >= rsiz )
+        {
+            sha3_8way_keccakf( c->st );
+            j = 0;
+        }
+    }
+    c->pt = j;
+
+    return 1;
+}
+
+// finalize and output a hash
+
+int sha3_8way_final( void *md, sha3_8way_ctx_t *c )
+{
+    c->st[ c->pt ] =
+                       _mm512_xor_si512( c->st[ c->pt ],
+                                         m512_const1_64( 6 ) );
+    c->st[ c->rsiz / 8 - 1 ] =
+                       _mm512_xor_si512( c->st[ c->rsiz / 8 - 1 ],
+                                         m512_const1_64( 0x8000000000000000 ) );
+    sha3_8way_keccakf( c->st );
+    memcpy( md, c->st, c->mdlen * 8 );
+    return 1;
+}
+
+// compute a SHA-3 hash (md) of given byte length from "in"
+
+void *sha3_8way( const void *in, size_t inlen, void *md, int mdlen )
+{
+    sha3_8way_ctx_t sha3;
+    sha3_8way_init( &sha3, mdlen);
+    sha3_8way_update( &sha3, in, inlen );
+    sha3_8way_final( md, &sha3 );
+    return md;
+}
+
+#endif  // AVX512
+#endif  // AVX2
--- a/algo/verthash/tiny_sha3/sha3-4way.h
+++ b/algo/verthash/tiny_sha3/sha3-4way.h
@@ -0,0 +1,67 @@
+// sha3.h
+// 19-Nov-11  Markku-Juhani O. Saarinen <mjos@iki.fi>
+// 2021-03-27 JayDDee
+//
+#ifndef SHA3_4WAY_H
+#define SHA3_4WAY_H
+
+#include <stddef.h>
+#include <stdint.h>
+#include "simd-utils.h"
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+#ifndef KECCAKF_ROUNDS
+#define KECCAKF_ROUNDS 24
+#endif
+
+#if defined(__AVX2__)
+
+typedef struct
+{
+   __m256i st[25];                     // 64-bit words * 4 lanes
+    int pt, rsiz, mdlen;                    // these don't overflow
+} sha3_4way_ctx_t __attribute__ ((aligned (64)));;
+
+// Compression function.
+void sha3_4way_keccakf( __m256i st[25] );
+
+// OpenSSL - like interfece
+int sha3_4way_init( sha3_4way_ctx_t *c, int mdlen );    // mdlen = hash output in bytes
+int sha3_4way_update( sha3_4way_ctx_t *c, const void *data, size_t len );
+int sha3_4way_final( void *md, sha3_4way_ctx_t *c );    // digest goes to md
+
+// compute a sha3 hash (md) of given byte length from "in"
+void *sha3_4way( const void *in, size_t inlen, void *md, int mdlen );
+
+
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+
+// state context
+typedef struct
+{
+   __m512i st[25];                     // 64-bit words * 8 lanes
+    int pt, rsiz, mdlen;                    // these don't overflow
+} sha3_8way_ctx_t __attribute__ ((aligned (64)));;
+
+// Compression function.
+void sha3_8way_keccakf( __m512i st[25] );
+
+// OpenSSL - like interfece
+int sha3_8way_init( sha3_8way_ctx_t *c, int mdlen );    // mdlen = hash output in bytes
+int sha3_8way_update( sha3_8way_ctx_t *c, const void *data, size_t len );
+int sha3_8way_final( void *md, sha3_8way_ctx_t *c );    // digest goes to md
+
+// compute a sha3 hash (md) of given byte length from "in"
+void *sha3_8way( const void *in, size_t inlen, void *md, int mdlen );
+
+#endif // AVX512
+#endif // AVX2
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif
--- a/algo/verthash/tiny_sha3/sha3.c
+++ b/algo/verthash/tiny_sha3/sha3.c
@@ -0,0 +1,226 @@
+// sha3.c
+// 19-Nov-11  Markku-Juhani O. Saarinen <mjos@iki.fi>
+
+// Revised 07-Aug-15 to match with official release of FIPS PUB 202 "SHA3"
+// Revised 03-Sep-15 for portability + OpenSSL - style API
+
+#include "sha3.h"
+#include <string.h>
+
+// update the state with given number of rounds
+
+void sha3_keccakf(uint64_t st[25])
+{
+    // constants
+    const uint64_t keccakf_rndc[24] = {
+        0x0000000000000001, 0x0000000000008082, 0x800000000000808a,
+        0x8000000080008000, 0x000000000000808b, 0x0000000080000001,
+        0x8000000080008081, 0x8000000000008009, 0x000000000000008a,
+        0x0000000000000088, 0x0000000080008009, 0x000000008000000a,
+        0x000000008000808b, 0x800000000000008b, 0x8000000000008089,
+        0x8000000000008003, 0x8000000000008002, 0x8000000000000080,
+        0x000000000000800a, 0x800000008000000a, 0x8000000080008081,
+        0x8000000000008080, 0x0000000080000001, 0x8000000080008008
+    };
+/*
+    const int keccakf_rotc[24] = {
+        1,  3,  6,  10, 15, 21, 28, 36, 45, 55, 2,  14,
+        27, 41, 56, 8,  25, 43, 62, 18, 39, 61, 20, 44
+    };
+    const int keccakf_piln[24] = {
+        10, 7,  11, 17, 18, 3, 5,  16, 8,  21, 24, 4,
+        15, 23, 19, 13, 12, 2, 20, 14, 22, 9,  6,  1
+    };
+*/
+
+    // variables
+    int i, j, r;
+    uint64_t t, bc[5];
+
+#if __BYTE_ORDER__ != __ORDER_LITTLE_ENDIAN__
+    uint8_t *v;
+
+    // endianess conversion. this is redundant on little-endian targets
+    for (i = 0; i < 25; i++) {
+        v = (uint8_t *) &st[i];
+        st[i] = ((uint64_t) v[0])     | (((uint64_t) v[1]) << 8) |
+            (((uint64_t) v[2]) << 16) | (((uint64_t) v[3]) << 24) |
+            (((uint64_t) v[4]) << 32) | (((uint64_t) v[5]) << 40) |
+            (((uint64_t) v[6]) << 48) | (((uint64_t) v[7]) << 56);
+    }
+#endif
+
+    // actual iteration
+    for (r = 0; r < KECCAKF_ROUNDS; r++) {
+
+        // Theta
+        for (i = 0; i < 5; i++)
+            bc[i] = st[i] ^ st[i + 5] ^ st[i + 10] ^ st[i + 15] ^ st[i + 20];
+
+        for (i = 0; i < 5; i++) {
+            t = bc[(i + 4) % 5] ^ ROTL64(bc[(i + 1) % 5], 1);
+            for (j = 0; j < 25; j += 5)
+                st[j + i] ^= t;
+        }
+
+        
+        // Rho Pi
+#define RHO_PI( i, c ) \
+   bc[0] = st[ i ]; \
+   st[ i ] = ROTL64( t, c ); \
+   t = bc[0]
+
+        t = st[1];
+
+        RHO_PI( 10,  1 );
+        RHO_PI(  7,  3 );
+        RHO_PI( 11,  6 );
+        RHO_PI( 17, 10 );
+        RHO_PI( 18, 15 );
+        RHO_PI(  3, 21 );
+        RHO_PI(  5, 28 );
+        RHO_PI( 16, 36 );
+        RHO_PI(  8, 45 );
+        RHO_PI( 21, 55 );
+        RHO_PI( 24,  2 );
+        RHO_PI(  4, 14 );
+        RHO_PI( 15, 27 );
+        RHO_PI( 23, 41 );
+        RHO_PI( 19, 56 );
+        RHO_PI( 13,  8 );
+        RHO_PI( 12, 25 );
+        RHO_PI(  2, 43 );
+        RHO_PI( 20, 62 );
+        RHO_PI( 14, 18 );
+        RHO_PI( 22, 39 );
+        RHO_PI(  9, 61 );
+        RHO_PI(  6, 20 );
+        RHO_PI(  1, 44 );
+
+#undef RHO_PI        
+
+/*        
+        for (i = 0; i < 24; i++) {
+            j = keccakf_piln[i];
+            bc[0] = st[j];
+            st[j] = ROTL64(t, keccakf_rotc[i]);
+            t = bc[0];
+        }
+*/
+
+        //  Chi
+        for (j = 0; j < 25; j += 5) {
+            for (i = 0; i < 5; i++)
+                bc[i] = st[j + i];
+            for (i = 0; i < 5; i++)
+                st[j + i] ^= (~bc[(i + 1) % 5]) & bc[(i + 2) % 5];
+        }
+
+        //  Iota
+        st[0] ^= keccakf_rndc[r];
+    }
+
+#if __BYTE_ORDER__ != __ORDER_LITTLE_ENDIAN__
+    // endianess conversion. this is redundant on little-endian targets
+    for (i = 0; i < 25; i++) {
+        v = (uint8_t *) &st[i];
+        t = st[i];
+        v[0] = t & 0xFF;
+        v[1] = (t >> 8) & 0xFF;
+        v[2] = (t >> 16) & 0xFF;
+        v[3] = (t >> 24) & 0xFF;
+        v[4] = (t >> 32) & 0xFF;
+        v[5] = (t >> 40) & 0xFF;
+        v[6] = (t >> 48) & 0xFF;
+        v[7] = (t >> 56) & 0xFF;
+    }
+#endif
+}
+
+// Initialize the context for SHA3
+
+int sha3_init(sha3_ctx_t *c, int mdlen)
+{
+    int i;
+
+    for (i = 0; i < 25; i++)
+        c->st.q[i] = 0;
+    c->mdlen = mdlen;
+    c->rsiz = 200 - 2 * mdlen;
+    c->pt = 0;
+
+    return 1;
+}
+
+// update state with more data
+
+int sha3_update(sha3_ctx_t *c, const void *data, size_t len)
+{
+    size_t i;
+    int j = c->pt / 8;
+    const int rsiz = c->rsiz / 8;
+    const int l = len / 8;
+
+    for ( i = 0; i < l; i++ )
+    {
+        c->st.q[ j++ ] ^= ( ((const uint64_t *) data) [i] );
+        if ( j >= rsiz )
+        {
+            sha3_keccakf( c->st.q );
+            j = 0;
+        }
+    }
+    c->pt = j*8;
+
+    return 1;
+}
+
+// finalize and output a hash
+
+int sha3_final(void *md, sha3_ctx_t *c)
+{
+    c->st.q[ c->pt / 8 ] ^= 6;
+    c->st.q[ c->rsiz / 8 - 1 ] ^= 0x8000000000000000;
+    sha3_keccakf(c->st.q);
+    memcpy( md, c->st.q, c->mdlen );
+    return 1;
+}
+
+// compute a SHA-3 hash (md) of given byte length from "in"
+
+void *sha3(const void *in, size_t inlen, void *md, int mdlen)
+{
+    sha3_ctx_t sha3;
+    sha3_init(&sha3, mdlen);
+    sha3_update(&sha3, in, inlen);
+    sha3_final(md, &sha3);
+
+    return md;
+}
+
+// SHAKE128 and SHAKE256 extensible-output functionality
+
+void shake_xof(sha3_ctx_t *c)
+{
+    c->st.b[c->pt] ^= 0x1F;
+    c->st.b[c->rsiz - 1] ^= 0x80;
+    sha3_keccakf(c->st.q);
+    c->pt = 0;
+}
+
+void shake_out(sha3_ctx_t *c, void *out, size_t len)
+{
+    size_t i;
+    int j;
+
+    j = c->pt;
+    for (i = 0; i < len; i++) {
+        if (j >= c->rsiz) {
+            sha3_keccakf(c->st.q);
+            j = 0;
+        }
+        ((uint8_t *) out)[i] = c->st.b[j++];
+    }
+    c->pt = j;
+}
+
--- a/algo/verthash/tiny_sha3/sha3.h
+++ b/algo/verthash/tiny_sha3/sha3.h
@@ -0,0 +1,55 @@
+// sha3.h
+// 19-Nov-11  Markku-Juhani O. Saarinen <mjos@iki.fi>
+
+#ifndef SHA3_H
+#define SHA3_H
+
+#include <stddef.h>
+#include <stdint.h>
+
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+#ifndef KECCAKF_ROUNDS
+#define KECCAKF_ROUNDS 24
+#endif
+
+#ifndef ROTL64
+#define ROTL64(x, y) (((x) << (y)) | ((x) >> (64 - (y))))
+#endif
+
+// state context
+typedef struct {
+    union {                                 // state:
+        uint8_t b[200];                     // 8-bit bytes
+        uint64_t q[25];                     // 64-bit words
+    } st;
+    int pt, rsiz, mdlen;                    // these don't overflow
+} sha3_ctx_t;
+
+// Compression function.
+void sha3_keccakf(uint64_t st[25]);
+
+// OpenSSL - like interfece
+int sha3_init(sha3_ctx_t *c, int mdlen);    // mdlen = hash output in bytes
+int sha3_update(sha3_ctx_t *c, const void *data, size_t len);
+int sha3_final(void *md, sha3_ctx_t *c);    // digest goes to md
+
+// compute a sha3 hash (md) of given byte length from "in"
+void *sha3(const void *in, size_t inlen, void *md, int mdlen);
+
+// SHAKE128 and SHAKE256 extensible-output functions
+#define shake128_init(c) sha3_init(c, 16)
+#define shake256_init(c) sha3_init(c, 32)
+#define shake_update sha3_update
+
+void shake_xof(sha3_ctx_t *c);
+void shake_out(sha3_ctx_t *c, void *out, size_t len);
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif
--- a/algo/verthash/verthash-gate.c
+++ b/algo/verthash/verthash-gate.c
@@ -0,0 +1,176 @@
+#include "algo-gate-api.h"
+#include "algo/sha/sph_sha2.h"
+#include "Verthash.h"
+#include "tiny_sha3/sha3-4way.h"
+
+static verthash_info_t verthashInfo;
+
+// Verthash data file hash in bytes for verification
+// 0x48aa21d7afededb63976d48a8ff8ec29d5b02563af4a1110b056cd43e83155a5
+static const uint8_t verthashDatFileHash_bytes[32] =
+{ 0xa5, 0x55, 0x31, 0xe8, 0x43, 0xcd, 0x56, 0xb0,
+  0x10, 0x11, 0x4a, 0xaf, 0x63, 0x25, 0xb0, 0xd5,
+  0x29, 0xec, 0xf8, 0x8f, 0x8a, 0xd4, 0x76, 0x39,
+  0xb6, 0xed, 0xed, 0xaf, 0xd7, 0x21, 0xaa, 0x48 };
+
+#if defined(__AVX2__)
+
+static __thread sha3_4way_ctx_t sha3_mid_ctxA;
+static __thread sha3_4way_ctx_t sha3_mid_ctxB;
+
+#else
+
+static __thread sha3_ctx_t sha3_mid_ctx[8];
+
+#endif
+
+void verthash_sha3_512_prehash_72( const void *input )
+{
+#if defined(__AVX2__)
+   
+   __m256i vin[10];
+   mm256_intrlv80_4x64( vin, input );
+
+   sha3_4way_init( &sha3_mid_ctxA, 64 );
+   sha3_4way_init( &sha3_mid_ctxB, 64 );
+
+   vin[0] = _mm256_add_epi8( vin[0], _mm256_set_epi64x( 4,3,2,1 ) );
+   sha3_4way_update( &sha3_mid_ctxA, vin, 72 );
+
+   vin[0] = _mm256_add_epi8( vin[0], _mm256_set1_epi64x( 4 ) );
+   sha3_4way_update( &sha3_mid_ctxB, vin, 72 );
+
+#else
+
+   char in[80] __attribute__ ((aligned (64)));
+   memcpy( in, input, 80 );   
+   for ( int i = 0; i < 8; i++ )
+   {
+      in[0] += 1;
+      sha3_init( &sha3_mid_ctx[i], 64 );
+      sha3_update( &sha3_mid_ctx[i], in, 72 );
+   }
+
+#endif
+}
+
+void verthash_sha3_512_final_8( void *hash, const uint64_t nonce )
+{
+#if defined(__AVX2__)
+
+    __m256i vhashA[ 10 ] __attribute__ ((aligned (64)));
+    __m256i vhashB[ 10 ] __attribute__ ((aligned (64)));
+
+   sha3_4way_ctx_t ctx;
+   const __m256i vnonce = _mm256_set1_epi64x( nonce );
+
+   memcpy( &ctx, &sha3_mid_ctxA, sizeof ctx );
+   sha3_4way_update( &ctx, &vnonce, 8 );
+   sha3_4way_final( vhashA, &ctx );
+
+   memcpy( &ctx, &sha3_mid_ctxB, sizeof ctx );
+   sha3_4way_update( &ctx, &vnonce, 8 );
+   sha3_4way_final( vhashB, &ctx );
+
+   dintrlv_4x64( hash,     hash+64,  hash+128, hash+192, vhashA, 512 );
+   dintrlv_4x64( hash+256, hash+320, hash+384, hash+448, vhashB, 512 );
+   
+#else
+
+   for ( int i = 0; i < 8; i++ )
+   {
+      sha3_ctx_t ctx;
+      memcpy( &ctx, &sha3_mid_ctx[i], sizeof ctx );
+      sha3_update( &ctx, &nonce, 8 );
+      sha3_final( hash + i*64, &ctx );
+   }
+   
+#endif
+}
+
+int scanhash_verthash( struct work *work, uint32_t max_nonce,
+                      uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t edata[20] __attribute__((aligned(64)));
+   uint32_t hash[8] __attribute__((aligned(64)));
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 1;
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+
+   mm128_bswap32_80( edata, pdata );
+   verthash_sha3_512_prehash_72( edata );
+
+   do
+   {
+      edata[19] = n;
+      verthash_hash( verthashInfo.data, verthashInfo.dataSize, 
+                     edata,  hash );
+      if ( valid_hash( hash, ptarget ) && !bench )
+      {
+         pdata[19] = bswap_32( n );
+         submit_solution( work, hash, mythr );
+      }
+      n++;
+   } while ( n < last_nonce && !work_restart[thr_id].restart );
+   *hashes_done = n - first_nonce;
+   pdata[19] = n;
+   return 0;
+}
+
+static const char *default_verthash_data_file = "verthash.dat";
+
+bool register_verthash_algo( algo_gate_t* gate )
+{
+  opt_target_factor = 256.0;
+  gate->scanhash  = (void*)&scanhash_verthash;
+  gate->optimizations = AVX2_OPT;
+   
+  const char *verthash_data_file = opt_data_file ? opt_data_file
+                                                 : default_verthash_data_file;
+  
+   int vhLoadResult = verthash_info_init( &verthashInfo, verthash_data_file );
+   if (vhLoadResult == 0) // No Error
+   {
+      if ( opt_verify )
+      {
+         uint8_t vhDataFileHash[32] = { 0 };
+
+         applog( LOG_NOTICE, "Verifying Verthash data" );
+         sph_sha256_full( vhDataFileHash, verthashInfo.data,
+                          verthashInfo.dataSize );
+         if ( memcmp( vhDataFileHash, verthashDatFileHash_bytes,
+                      sizeof(verthashDatFileHash_bytes) ) == 0 )
+            applog( LOG_NOTICE, "Verthash data has been verified" );
+         else
+         {
+            applog( LOG_ERR, "Verthash data verification has failed" );
+            return false;
+         }
+      }
+   }
+   else
+   {
+      // Handle Verthash error codes
+      if ( vhLoadResult == 1 )
+      {
+         applog( LOG_ERR, "Verthash data file not found: %s",
+                 verthash_data_file );
+         if ( !opt_data_file )
+            applog( LOG_NOTICE, "Add '--verify' to create verthash.dat");
+      }
+      else if ( vhLoadResult == 2 )
+         applog( LOG_ERR, "Failed to allocate memory for Verthash data" );
+//       else // for debugging purposes
+//          applog( LOG_ERR, "Verthash data initialization unknown error code: %d",
+//                 vhLoadResult );
+      return false;
+   }
+
+   printf("\n");
+   return true;
+}
+
--- a/algo/x13/phi1612.c
+++ b/algo/x13/phi1612.c
@@ -42,7 +42,6 @@ void init_phi1612_ctx()
     sph_skein512_init( &phi_ctx.skein );
     sph_jh512_init( &phi_ctx.jh );
     cubehashInit( &phi_ctx.cube, 512, 16, 32 );
-     sph_fugue512_init( &phi_ctx.fugue );
     sph_gost512_init( &phi_ctx.gost );
 #ifdef __AES__
     init_echo( &phi_ctx.echo, 512 );
--- a/algo/x16/hex.c
+++ b/algo/x16/hex.c
@@ -161,9 +161,9 @@ int hex_hash( void* output, const void* input, int thrid )
                sph_whirlpool512_full( &ctx.whirlpool, hash, in,  size );
         break;
         case SHA_512:
-             SHA512_Init( &ctx.sha512 );
-             SHA512_Update( &ctx.sha512, in, size );
-             SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+             sph_sha512_init( &ctx.sha512 );
+             sph_sha512( &ctx.sha512, in, size );
+             sph_sha512_close( &ctx.sha512, hash );
         break;
      }

--- a/algo/x16/minotaur.c
+++ b/algo/x16/minotaur.c
@@ -17,7 +17,7 @@
 #include "algo/hamsi/sph_hamsi.h"
 #include "algo/shabal/sph_shabal.h"
 #include "algo/whirlpool/sph_whirlpool.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/echo/aes_ni/hash_api.h"
  #include "algo/groestl/aes_ni/hash-groestl.h"
@@ -58,7 +58,7 @@ struct TortureGarden
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;

    struct TortureNode {
        unsigned int algo;
@@ -120,9 +120,9 @@ static void get_hash( void *output, const void *input, TortureGarden *garden,
            sph_hamsi512_close(&garden->hamsi, hash);          
            break;
        case 7:
-            SHA512_Init( &garden->sha512 );
-            SHA512_Update( &garden->sha512, input, 64 );
-            SHA512_Final( (unsigned char*)hash, &garden->sha512 );
+            sph_sha512_init( &garden->sha512 );
+            sph_sha512( &garden->sha512, input, 64 );
+            sph_sha512_close( &garden->sha512, hash );
            break;
        case 8:
            sph_jh512_init(&garden->jh);
@@ -229,9 +229,9 @@ int minotaur_hash( void *output, const void *input, int thr_id )
    unsigned char hash[64] __attribute__ ((aligned (64)));

    // Find initial sha512 hash
-    SHA512_Init( &garden.sha512 );
-    SHA512_Update( &garden.sha512, input, 80 );
-    SHA512_Final( (unsigned char*) hash, &garden.sha512 );
+    sph_sha512_init( &garden.sha512 );
+    sph_sha512( &garden.sha512, input, 80 );
+    sph_sha512_close( &garden.sha512, hash );

    // algo 6 (Hamsi) is very slow. It's faster to skip hashing this nonce
    // if Hamsi is needed but only the first and last functions are
--- a/algo/x16/x16r-4way.c
+++ b/algo/x16/x16r-4way.c
@@ -16,8 +16,7 @@

 #if defined (X16R_8WAY)

-// Perform midstate prehash of hash functions with block size <= 64 bytes
-// and interleave 4x64 before nonce insertion for final hash.
+// Perform midstate prehash of hash functions with block size <= 72 bytes.

 void x16r_8way_prehash( void *vdata, void *pdata )
 {
@@ -34,6 +33,11 @@ void x16r_8way_prehash( void *vdata, void *pdata )
         jh512_8way_init( &x16r_ctx.jh );
         jh512_8way_update( &x16r_ctx.jh, vdata, 64 );
      break;
+      case KECCAK:
+         mm512_bswap32_intrlv80_8x64( vdata, pdata );
+         keccak512_8way_init( &x16r_ctx.keccak );
+         keccak512_8way_update( &x16r_ctx.keccak, vdata, 72 );
+      break;
      case SKEIN:
         mm512_bswap32_intrlv80_8x64( vdata, pdata );
         skein512_8way_init( &x16r_ctx.skein );
@@ -173,13 +177,13 @@ int x16r_8way_hash_generic( void* output, const void* input, int thrid )
                          hash7, vhash );
         break;
         case KECCAK:
-            keccak512_8way_init( &ctx.keccak );
-            if ( i == 0 )
-               keccak512_8way_update( &ctx.keccak, input, size );
+           if ( i == 0 )
+               keccak512_8way_update( &ctx.keccak, input + (72<<3), 8 );
            else
            {
               intrlv_8x64( vhash, in0, in1, in2, in3, in4, in5, in6, in7, 
                            size<<3 );
+               keccak512_8way_init( &ctx.keccak );
               keccak512_8way_update( &ctx.keccak, vhash, size );
            }
            keccak512_8way_close( &ctx.keccak, vhash );
@@ -490,6 +494,7 @@ int scanhash_x16r_8way( struct work *work, uint32_t max_nonce,
   {
      x16_r_s_getAlgoString( (const uint8_t*)bedata1, x16r_hash_order );
      s_ntime = ntime;
+
      if ( opt_debug && !thr_id )
          applog( LOG_INFO, "hash order %s (%08x)", x16r_hash_order, ntime );
   }
@@ -533,6 +538,11 @@ void x16r_4way_prehash( void *vdata, void *pdata )
         jh512_4way_init( &x16r_ctx.jh );
         jh512_4way_update( &x16r_ctx.jh, vdata, 64 );
      break;
+      case KECCAK:
+         mm256_bswap32_intrlv80_4x64( vdata, pdata );
+         keccak512_4way_init( &x16r_ctx.keccak );
+         keccak512_4way_update( &x16r_ctx.keccak, vdata, 72 );
+      break;
      case SKEIN:
         mm256_bswap32_intrlv80_4x64( vdata, pdata );
         skein512_4way_prehash64( &x16r_ctx.skein, vdata );
@@ -619,11 +629,20 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
            dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
         break;
         case GROESTL:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            groestl512_full( &ctx.groestl, (char*)hash0, (char*)in0, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash1, (char*)in1, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash2, (char*)in2, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash3, (char*)in3, size<<3 );
-         break;
+#endif
+   	    break;
         case JH:
            if ( i == 0 )
               jh512_4way_update( &ctx.jh, input + (64<<2), 16 );
@@ -637,12 +656,12 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
            dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
         break;
         case KECCAK:
-            keccak512_4way_init( &ctx.keccak );
-            if ( i == 0 )
-               keccak512_4way_update( &ctx.keccak, input, size );
+           if ( i == 0 )
+               keccak512_4way_update( &ctx.keccak, input + (72<<2), 8 );
            else
            {
               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               keccak512_4way_init( &ctx.keccak );
               keccak512_4way_update( &ctx.keccak, vhash, size );
            }
            keccak512_4way_close( &ctx.keccak, vhash );
@@ -711,11 +730,20 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
            }
         break;
         case SHAVITE:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            shavite512_full( &ctx.shavite, hash0, in0, size );
            shavite512_full( &ctx.shavite, hash1, in1, size );
            shavite512_full( &ctx.shavite, hash2, in2, size );
            shavite512_full( &ctx.shavite, hash3, in3, size );
-         break;
+#endif
+   	    break;
         case SIMD:
            intrlv_2x128( vhash, in0, in1, size<<3 );
            simd512_2way_full( &ctx.simd, vhash, vhash, size );
@@ -725,6 +753,14 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
            dintrlv_2x128_512( hash2, hash3, vhash );
         break;
         case ECHO:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            echo_full( &ctx.echo, (BitSequence *)hash0, 512,
                              (const BitSequence *)in0, size );
            echo_full( &ctx.echo, (BitSequence *)hash1, 512,
@@ -733,7 +769,8 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
                              (const BitSequence *)in2, size );
            echo_full( &ctx.echo, (BitSequence *)hash3, 512,
                              (const BitSequence *)in3, size );
-         break;
+#endif
+   	    break;
         case HAMSI:
            if ( i == 0 )
               hamsi512_4way_update( &ctx.hamsi, input + (64<<2), 16 );
@@ -856,7 +893,7 @@ int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
      x16_r_s_getAlgoString( (const uint8_t*)bedata1, x16r_hash_order );
      s_ntime = ntime;
      if ( opt_debug && !thr_id )
-              applog( LOG_INFO, "hash order %s (%08x)", x16r_hash_order, ntime );
+         applog( LOG_INFO, "hash order %s (%08x)", x16r_hash_order, ntime );
   }

   x16r_4way_prehash( vdata, pdata );
--- a/algo/x16/x16r-gate.c
+++ b/algo/x16/x16r-gate.c
@@ -61,7 +61,8 @@ bool register_x16r_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x16r;
  gate->hash      = (void*)&x16r_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                VAES_OPT | VAES256_OPT;
  x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
  opt_target_factor = 256.0;
  return true;
@@ -79,7 +80,8 @@ bool register_x16rv2_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x16rv2;
  gate->hash      = (void*)&x16rv2_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                VAES_OPT | VAES256_OPT;
  x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
  opt_target_factor = 256.0;
  return true;
@@ -97,7 +99,8 @@ bool register_x16s_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x16r;
  gate->hash      = (void*)&x16r_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                VAES_OPT | VAES256_OPT;
  x16_r_s_getAlgoString = (void*)&x16s_getAlgoString;
  opt_target_factor = 256.0;
  return true;
@@ -230,7 +233,8 @@ bool register_x16rt_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x16rt;
  gate->hash      = (void*)&x16r_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                VAES_OPT | VAES256_OPT;
  opt_target_factor = 256.0;
  return true;
 };
@@ -247,7 +251,8 @@ bool register_x16rt_veil_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x16rt;
  gate->hash      = (void*)&x16r_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                VAES_OPT | VAES256_OPT;
  gate->build_extraheader = (void*)&veil_build_extraheader;
  opt_target_factor = 256.0;
  return true;
@@ -277,22 +282,17 @@ bool register_x21s_algo( algo_gate_t* gate )
  gate->scanhash          = (void*)&scanhash_x21s_8way;
  gate->hash              = (void*)&x21s_8way_hash;
  gate->miner_thread_init = (void*)&x21s_8way_thread_init;
-  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT
-                            | VAES_OPT;
 #elif defined (X16R_4WAY)
  gate->scanhash          = (void*)&scanhash_x21s_4way;
  gate->hash              = (void*)&x21s_4way_hash;
  gate->miner_thread_init = (void*)&x21s_4way_thread_init;
-  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                            | AVX512_OPT | VAES_OPT;
 #else
  gate->scanhash          = (void*)&scanhash_x21s;
  gate->hash              = (void*)&x21s_hash;
  gate->miner_thread_init = (void*)&x21s_thread_init;
-  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                          | AVX512_OPT | VAES_OPT;
 #endif
-//  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT | AVX512_OPT;
+  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
+	                    VAES_OPT | VAES256_OPT;
  x16_r_s_getAlgoString   = (void*)&x16s_getAlgoString;
  opt_target_factor = 256.0;
  return true;
--- a/algo/x16/x16r-gate.h
+++ b/algo/x16/x16r-gate.h
@@ -20,13 +20,16 @@
 #include "algo/fugue/sph_fugue.h"
 #include "algo/shabal/sph_shabal.h"
 #include "algo/whirlpool/sph_whirlpool.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
+
 #if defined(__AES__)
-  #include "algo/echo/aes_ni/hash_api.h"
-  #include "algo/groestl/aes_ni/hash-groestl.h"
-  #include "algo/fugue/fugue-aesni.h"
+#include "algo/echo/aes_ni/hash_api.h"
+#include "algo/groestl/aes_ni/hash-groestl.h"
+#include "algo/fugue/fugue-aesni.h"
 #endif
+
 #if defined (__AVX2__)
+
 #include "algo/blake/blake-hash-4way.h"
 #include "algo/bmw/bmw-hash-4way.h"
 #include "algo/groestl/aes_ni/hash-groestl.h"
@@ -39,11 +42,14 @@
 #include "algo/hamsi/hamsi-hash-4way.h"
 #include "algo/shabal/shabal-hash-4way.h"
 #include "algo/sha/sha-hash-4way.h"
+
 #if defined(__VAES__)
-  #include "algo/groestl/groestl512-hash-4way.h"
-  #include "algo/shavite/shavite-hash-4way.h"
-  #include "algo/echo/echo-hash-4way.h"
+#include "algo/groestl/groestl512-hash-4way.h"
+#include "algo/shavite/shavite-hash-2way.h"
+#include "algo/shavite/shavite-hash-4way.h"
+#include "algo/echo/echo-hash-4way.h"
 #endif
+
 #endif // AVX2

 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
@@ -145,15 +151,21 @@ union _x16r_4way_context_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
-    hashState_echo          echo;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    shavite512_2way_context shavite;
+    echo_2way_context       echo;
+#else
    hashState_groestl       groestl;
+    shavite512_context      shavite;
+    hashState_echo          echo;
+#endif
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
    luffa_2way_context      luffa;
    hashState_luffa         luffa1;
    cubehashParam           cube;
-    shavite512_context      shavite;
    simd_2way_context       simd;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
@@ -199,7 +211,7 @@ union _x16r_context_overlay
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
 } __attribute__ ((aligned (64)));

 typedef union _x16r_context_overlay x16r_context_overlay;
--- a/algo/x16/x16r.c
+++ b/algo/x16/x16r.c
@@ -177,9 +177,9 @@ int x16r_hash_generic( void* output, const void* input, int thrid )
               sph_whirlpool512_full( &ctx.whirlpool, hash, in, size );
         break;
         case SHA_512:
-            SHA512_Init( &ctx.sha512 );
-            SHA512_Update( &ctx.sha512, in, size );
-            SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+            sph_sha512_init( &ctx.sha512 );
+            sph_sha512( &ctx.sha512, in, size );
+            sph_sha512_close( &ctx.sha512, hash );
         break;
      }

--- a/algo/x16/x16rv2-4way.c
+++ b/algo/x16/x16rv2-4way.c
@@ -672,14 +672,20 @@ union _x16rv2_4way_context_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
-    hashState_echo          echo;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    shavite512_2way_context shavite;
+    echo_2way_context       echo;
+#else
    hashState_groestl       groestl;
+    shavite512_context      shavite;
+    hashState_echo          echo;
+#endif
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
    luffa_2way_context      luffa;
    cubehashParam           cube;
-    shavite512_context      shavite;
    simd_2way_context       simd;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
@@ -745,10 +751,19 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
            dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
         break;
         case GROESTL:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            groestl512_full( &ctx.groestl, (char*)hash0, (char*)in0, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash1, (char*)in1, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash2, (char*)in2, size<<3 );
            groestl512_full( &ctx.groestl, (char*)hash3, (char*)in3, size<<3 );
+#endif
         break;
         case JH:
            if ( i == 0 )
@@ -887,10 +902,19 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
            }
         break;
         case SHAVITE:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            shavite512_full( &ctx.shavite, hash0, in0, size );
            shavite512_full( &ctx.shavite, hash1, in1, size );
            shavite512_full( &ctx.shavite, hash2, in2, size );
            shavite512_full( &ctx.shavite, hash3, in3, size );
+#endif
         break;
         case SIMD:
            intrlv_2x128( vhash, in0, in1, size<<3 );
@@ -901,6 +925,14 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
            dintrlv_2x128_512( hash2, hash3, vhash );
         break;
         case ECHO:
+#if defined(__VAES__)
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
+            dintrlv_2x128_512( hash0, hash1, vhash );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
+            dintrlv_2x128_512( hash2, hash3, vhash );
+#else
            echo_full( &ctx.echo, (BitSequence *)hash0, 512,
                              (const BitSequence *)in0, size );
            echo_full( &ctx.echo, (BitSequence *)hash1, 512,
@@ -909,6 +941,7 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
                              (const BitSequence *)in2, size );
            echo_full( &ctx.echo, (BitSequence *)hash3, 512,
                              (const BitSequence *)in3, size );
+#endif
         break;
         case HAMSI:
            if ( i == 0 )
--- a/algo/x16/x16rv2.c
+++ b/algo/x16/x16rv2.c
@@ -33,7 +33,7 @@ union _x16rv2_context_overlay
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_tiger_context       tiger;
 };
 typedef union _x16rv2_context_overlay x16rv2_context_overlay;
@@ -155,9 +155,9 @@ int x16rv2_hash( void* output, const void* input, int thrid )
             sph_tiger( &ctx.tiger, in, size );
             sph_tiger_close( &ctx.tiger, hash );
             padtiger512( hash );
-             SHA512_Init( &ctx.sha512 );
-             SHA512_Update( &ctx.sha512, hash, 64 );
-             SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+             sph_sha512_init( &ctx.sha512 );
+             sph_sha512( &ctx.sha512, hash, 64 );
+             sph_sha512_close( &ctx.sha512, hash );
         break;
      }

--- a/algo/x16/x21s-4way.c
+++ b/algo/x16/x21s-4way.c
@@ -13,7 +13,7 @@
 #include "algo/gost/sph_gost.h"
 #include "algo/lyra2/lyra2.h"
 #if defined(__SHA__)
- #include <openssl/sha.h>
+  #include "algo/sha/sph_sha2.h"
 #endif

 #if defined (X21S_8WAY)
@@ -209,7 +209,7 @@ union _x21s_4way_context_overlay
    sph_tiger_context       tiger;
    sph_gost512_context     gost;
 #if defined(__SHA__)
-    SHA256_CTX              sha256;
+    sph_sha256_context      sha256;
 #else
    sha256_4way_context     sha256;
 #endif
@@ -275,23 +275,18 @@ int x21s_4way_hash( void* output, const void* input, int thrid )

 #if defined(__SHA__)

-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update( &ctx.sha256, hash0, 64 );
-   SHA256_Final( (unsigned char*)hash0, &ctx.sha256 );
-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update( &ctx.sha256, hash1, 64 );
-   SHA256_Final( (unsigned char*)hash1, &ctx.sha256 );
-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update( &ctx.sha256, hash2, 64 );
-   SHA256_Final( (unsigned char*)hash2, &ctx.sha256 );
-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update( &ctx.sha256, hash3, 64 );
-   SHA256_Final( (unsigned char*)hash3, &ctx.sha256 );
-
-   memcpy( output,    hash0, 32 );
-   memcpy( output+32, hash1, 32 );
-   memcpy( output+64, hash2, 32 );
-   memcpy( output+96, hash3, 32 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash0, 64 );
+   sph_sha256_close( &ctx.sha256, output );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash1, 64 );
+   sph_sha256_close( &ctx.sha256, output+32 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash2, 64 );
+   sph_sha256_close( &ctx.sha256, output+64 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash3, 64 );
+   sph_sha256_close( &ctx.sha256, output+96 );

 #else

--- a/algo/x16/x21s.c
+++ b/algo/x16/x21s.c
@@ -8,7 +8,7 @@
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #include "algo/haval/sph-haval.h"
 #include "algo/tiger/sph_tiger.h"
 #include "algo/gost/sph_gost.h"
@@ -23,7 +23,7 @@ union _x21s_context_overlay
        sph_haval256_5_context  haval;
        sph_tiger_context       tiger;
        sph_gost512_context     gost;
-        SHA256_CTX              sha256;
+        sph_sha256_context      sha256;
 };
 typedef union _x21s_context_overlay x21s_context_overlay;

@@ -50,9 +50,9 @@ int x21s_hash( void* output, const void* input, int thrid )
   sph_gost512 ( &ctx.gost, (const void*) hash, 64 );
   sph_gost512_close( &ctx.gost, (void*) hash );

-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update( &ctx.sha256, hash, 64 );
-   SHA256_Final( (unsigned char*)hash, &ctx.sha256 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash, 64 );
+   sph_sha256_close( &ctx.sha256, hash );

   memcpy( output, hash, 32 );

--- a/algo/x17/sonoa-4way.c
+++ b/algo/x17/sonoa-4way.c
@@ -1124,7 +1124,13 @@ union _sonoa_4way_context_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    echo512_2way_context    echo;
+#else
    hashState_groestl       groestl;
+    hashState_echo          echo;
+#endif
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
@@ -1132,7 +1138,6 @@ union _sonoa_4way_context_overlay
    cube_2way_context       cube;
    shavite512_2way_context shavite;
    simd_2way_context       simd;
-    hashState_echo          echo;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
    shabal512_4way_context  shabal;
@@ -1162,6 +1167,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1171,6 +1187,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     
     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1195,6 +1213,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1206,16 +1233,29 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
                     (const BitSequence *)hash2, 64 );
     echo_full( &ctx.echo, (BitSequence *)hash3, 512,
                     (const BitSequence *)hash3, 64 );
+
+     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
+
+#endif
     
     if ( work_restart[thr_id].restart ) return 0;
 // 2

-     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
-
     bmw512_4way_init( &ctx.bmw );
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 ); 
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1225,6 +1265,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1249,6 +1291,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1263,6 +1314,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1274,6 +1327,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 ); 
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1283,6 +1347,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1307,6 +1373,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1321,6 +1396,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1340,6 +1417,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 ); 
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1349,6 +1437,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1373,6 +1463,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1387,6 +1486,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1410,6 +1511,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     hamsi512_4way_update( &ctx.hamsi, vhashB, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     echo_full( &ctx.echo, (BitSequence *)hash0, 512,
@@ -1424,6 +1534,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     intrlv_2x128_512( vhashA, hash0, hash1 );
     intrlv_2x128_512( vhashB, hash2, hash3 );

+#endif
+
     shavite512_2way_init( &ctx.shavite );
     shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
     shavite512_2way_init( &ctx.shavite );
@@ -1443,6 +1555,20 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     shabal512_4way_update( &ctx.shabal, vhashB, 64 );
     shabal512_4way_close( &ctx.shabal, vhash );

+#if defined(__VAES__)
+
+//     rintrlv_4x32_2x128( vhashA, vhashB, vhash, 512 ); 
+     dintrlv_4x32_512( hash0, hash1, hash2, hash3, vhash );
+     intrlv_2x128_512( vhashA, hash0, hash1 );
+     intrlv_2x128_512( vhashB, hash2, hash3 );
+     
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x32_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1452,6 +1578,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1476,6 +1604,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1490,6 +1627,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1523,6 +1662,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 ); 
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1532,6 +1682,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1556,6 +1708,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1570,6 +1731,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1616,6 +1779,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 ); 
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+ 
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1625,6 +1799,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif     
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -1649,6 +1825,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -1663,6 +1848,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
--- a/algo/x17/sonoa-gate.c
+++ b/algo/x17/sonoa-gate.c
@@ -12,7 +12,7 @@ bool register_sonoa_algo( algo_gate_t* gate )
  init_sonoa_ctx();
  gate->hash      = (void*)&sonoa_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
  return true;
 };

--- a/algo/x17/sonoa.c
+++ b/algo/x17/sonoa.c
@@ -20,7 +20,7 @@
 #include "algo/luffa/luffa_for_sse2.h"
 #include "algo/cubehash/cubehash_sse2.h"
 #include "algo/simd/nist.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/echo/aes_ni/hash_api.h"
  #include "algo/groestl/aes_ni/hash-groestl.h"
@@ -53,7 +53,7 @@ typedef struct {
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_haval256_5_context  haval;
 } sonoa_ctx_holder;

@@ -82,7 +82,7 @@ void init_sonoa_ctx()
        sph_hamsi512_init( &sonoa_ctx.hamsi );
        sph_shabal512_init( &sonoa_ctx.shabal );
        sph_whirlpool_init( &sonoa_ctx.whirlpool );
-        SHA512_Init( &sonoa_ctx.sha512 );
+        sph_sha512_init( &sonoa_ctx.sha512 );
        sph_haval256_5_init(&sonoa_ctx.haval);
 };

@@ -494,8 +494,8 @@ int sonoa_hash( void *state, const void *input, int thr_id )
   sph_whirlpool(&ctx.whirlpool, hash, 64);
   sph_whirlpool_close(&ctx.whirlpool, hash);

-   SHA512_Update( &ctx.sha512, hash, 64 );
-   SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+   sph_sha512( &ctx.sha512, hash, 64 );
+   sph_sha512_close( &ctx.sha512, hash );

   sph_whirlpool_init( &ctx.whirlpool );
   sph_whirlpool(&ctx.whirlpool, hash, 64);
@@ -574,9 +574,9 @@ int sonoa_hash( void *state, const void *input, int thr_id )
   sph_whirlpool(&ctx.whirlpool, hash, 64);
   sph_whirlpool_close(&ctx.whirlpool, hash);

-   SHA512_Init( &ctx.sha512 );
-   SHA512_Update( &ctx.sha512, hash, 64 );
-   SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+   sph_sha512_init( &ctx.sha512 );
+   sph_sha512( &ctx.sha512, hash, 64 );
+   sph_sha512_close( &ctx.sha512, hash );

   sph_haval256_5(&ctx.haval,(const void*) hash, 64);
   sph_haval256_5_close(&ctx.haval, hash);
--- a/algo/x17/x17-4way.c
+++ b/algo/x17/x17-4way.c
@@ -240,7 +240,13 @@ union _x17_4way_context_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    echo512_2way_context    echo;
+#else
    hashState_groestl       groestl;
+    hashState_echo          echo;
+#endif
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
@@ -248,7 +254,6 @@ union _x17_4way_context_overlay
    cube_2way_context       cube;
    shavite512_2way_context shavite;
    simd_2way_context       simd;
-    hashState_echo          echo;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
    shabal512_4way_context  shabal;
@@ -275,6 +280,17 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, 64 );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+     
     dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -284,6 +300,8 @@ int x17_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

     jh512_4way_init( &ctx.jh );
@@ -308,6 +326,15 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
     dintrlv_2x128_512( hash0, hash1, vhashA );
     dintrlv_2x128_512( hash2, hash3, vhashB );

@@ -322,6 +349,8 @@ int x17_4way_hash( void *state, const void *input, int thr_id )

     intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
     hamsi512_4way_close( &ctx.hamsi, vhash );
--- a/algo/x17/x17-gate.c
+++ b/algo/x17/x17-gate.c
@@ -11,7 +11,7 @@ bool register_x17_algo( algo_gate_t* gate )
 #else
  gate->hash      = (void*)&x17_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
  return true;
 };

--- a/algo/x17/x17.c
+++ b/algo/x17/x17.c
@@ -19,7 +19,7 @@
 #include "algo/luffa/luffa_for_sse2.h" 
 #include "algo/cubehash/cubehash_sse2.h"
 #include "algo/simd/nist.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/fugue/fugue-aesni.h"
  #include "algo/echo/aes_ni/hash_api.h"
@@ -53,7 +53,7 @@ union _x17_context_overlay
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_haval256_5_context  haval;
 };
 typedef union _x17_context_overlay x17_context_overlay;
@@ -140,9 +140,9 @@ int x17_hash(void *output, const void *input, int thr_id )
    sph_whirlpool( &ctx.whirlpool, hash, 64 );
    sph_whirlpool_close( &ctx.whirlpool, hash );

-    SHA512_Init( &ctx.sha512 );
-    SHA512_Update( &ctx.sha512, hash, 64 );
-    SHA512_Final( (unsigned char*)hash, &ctx.sha512 );
+    sph_sha512_init( &ctx.sha512 );
+    sph_sha512( &ctx.sha512, hash, 64 );
+    sph_sha512_close( &ctx.sha512, hash );

    sph_haval256_5_init(&ctx.haval);
    sph_haval256_5( &ctx.haval, (const void*)hash, 64 );
--- a/algo/x17/xevan-4way.c
+++ b/algo/x17/xevan-4way.c
@@ -405,15 +405,20 @@ union _xevan_4way_context_overlay
 {
 	blake512_4way_context   blake;
        bmw512_4way_context     bmw;
-        hashState_groestl       groestl;
-        skein512_4way_context   skein;
+#if defined(__VAES__)
+        groestl512_2way_context groestl;
+        echo_2way_context       echo;
+#else
+	hashState_groestl       groestl;
+        hashState_echo          echo;
+#endif
+	skein512_4way_context   skein;
        jh512_4way_context      jh;
        keccak512_4way_context  keccak;
        luffa_2way_context      luffa;
        cube_2way_context       cube;
        shavite512_2way_context shavite;
        simd_2way_context       simd;
-        hashState_echo          echo;
        hamsi512_4way_context   hamsi;
        hashState_fugue         fugue;
        shabal512_4way_context  shabal;
@@ -442,7 +447,17 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, dataLen );
     bmw512_4way_close( &ctx.bmw, vhash );

-     // Serial
+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, dataLen );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, dataLen );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
+
+#else
+     
     dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, dataLen<<3 );
@@ -450,9 +465,10 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, dataLen<<3 );
     groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, dataLen<<3 );

-     // Parallel 4way
     intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );

+#endif
+
     skein512_4way_full( &ctx.skein, vhash, vhash, dataLen );

     jh512_4way_init( &ctx.jh );
@@ -477,6 +493,15 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, dataLen );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, dataLen );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, dataLen );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, dataLen );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
+
+#else
+     
     dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
     dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );

@@ -489,9 +514,10 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     echo_full( &ctx.echo, (BitSequence *)hash3, 512,
                     (const BitSequence *)hash3, dataLen );

-     // Parallel
     intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, dataLen );
     hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -542,6 +568,17 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     bmw512_4way_update( &ctx.bmw, vhash, dataLen );
     bmw512_4way_close( &ctx.bmw, vhash );

+#if defined(__VAES__)
+
+     rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
+
+     groestl512_2way_full( &ctx.groestl, vhashA, vhashA, dataLen );
+     groestl512_2way_full( &ctx.groestl, vhashB, vhashB, dataLen );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
+
+#else
+
     dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );

     groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, dataLen<<3 );
@@ -551,6 +588,8 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )

     intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );

+#endif
+
     skein512_4way_full( &ctx.skein, vhash, vhash, dataLen );

     jh512_4way_init( &ctx.jh );
@@ -575,6 +614,15 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
     simd512_2way_full( &ctx.simd, vhashA, vhashA, dataLen );
     simd512_2way_full( &ctx.simd, vhashB, vhashB, dataLen );

+#if defined(__VAES__)
+
+     echo_2way_full( &ctx.echo, vhashA, 512, vhashA, dataLen );
+     echo_2way_full( &ctx.echo, vhashB, 512, vhashB, dataLen );
+
+     rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
+
+#else
+
     dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
     dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );

@@ -589,6 +637,8 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )

     intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );

+#endif
+
     hamsi512_4way_init( &ctx.hamsi );
     hamsi512_4way_update( &ctx.hamsi, vhash, dataLen );
     hamsi512_4way_close( &ctx.hamsi, vhash );
--- a/algo/x17/xevan-gate.c
+++ b/algo/x17/xevan-gate.c
@@ -12,7 +12,7 @@ bool register_xevan_algo( algo_gate_t* gate )
  init_xevan_ctx();
  gate->hash      = (void*)&xevan_hash;
 #endif
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
  opt_target_factor = 256.0;
  return true;
 };
--- a/algo/x17/xevan.c
+++ b/algo/x17/xevan.c
@@ -20,7 +20,7 @@
 #include "algo/haval/sph-haval.h"
 #include "algo/simd/nist.h"
 #include "algo/cubehash/cubehash_sse2.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/groestl/aes_ni/hash-groestl.h"
  #include "algo/echo/aes_ni/hash_api.h"
@@ -44,7 +44,7 @@ typedef struct {
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_haval256_5_context  haval;
 #if defined(__AES__)
        hashState_echo          echo;
@@ -73,7 +73,7 @@ void init_xevan_ctx()
        sph_hamsi512_init( &xevan_ctx.hamsi );
        sph_shabal512_init( &xevan_ctx.shabal );
        sph_whirlpool_init( &xevan_ctx.whirlpool );
-        SHA512_Init( &xevan_ctx.sha512 );
+        sph_sha512_init( &xevan_ctx.sha512 );
        sph_haval256_5_init(&xevan_ctx.haval);
 #if defined(__AES__)
        init_groestl( &xevan_ctx.groestl, 64 );
@@ -95,97 +95,27 @@ int xevan_hash(void *output, const void *input, int thr_id )

   sph_blake512( &ctx.blake, input, 80 );
   sph_blake512_close( &ctx.blake, hash );
-	memset(&hash[16], 0, 64);
+   memset(&hash[16], 0, 64);

-	sph_bmw512(&ctx.bmw, hash, dataLen);
-	sph_bmw512_close(&ctx.bmw, hash);
+   sph_bmw512(&ctx.bmw, hash, dataLen);
+   sph_bmw512_close(&ctx.bmw, hash);

 #if defined(__AES__)
   update_and_final_groestl( &ctx.groestl, (char*)hash,
                                     (const char*)hash, dataLen*8 );
 #else
-	sph_groestl512(&ctx.groestl, hash, dataLen);
-	sph_groestl512_close(&ctx.groestl, hash);
-#endif
-
-	sph_skein512(&ctx.skein, hash, dataLen);
-	sph_skein512_close(&ctx.skein, hash);
-
-	sph_jh512(&ctx.jh, hash, dataLen);
-	sph_jh512_close(&ctx.jh, hash);
-
-	sph_keccak512(&ctx.keccak, hash, dataLen);
-	sph_keccak512_close(&ctx.keccak, hash);
-
-   update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
-                                 (const BitSequence*)hash, dataLen );
-
-   cubehashUpdateDigest( &ctx.cubehash, (byte*)hash,
-                                 (const byte*) hash, dataLen );
-
-	sph_shavite512(&ctx.shavite, hash, dataLen);
-	sph_shavite512_close(&ctx.shavite, hash);
-
-   update_final_sd( &ctx.simd, (BitSequence *)hash,
-                         (const BitSequence *)hash, dataLen*8 );
-
-#if defined(__AES__)
-   update_final_echo( &ctx.echo, (BitSequence *) hash,
-                           (const BitSequence *) hash, dataLen*8 );
-#else
-	sph_echo512(&ctx.echo, hash, dataLen);
-	sph_echo512_close(&ctx.echo, hash);
-#endif
-
-	sph_hamsi512(&ctx.hamsi, hash, dataLen);
-	sph_hamsi512_close(&ctx.hamsi, hash);
-
-#if defined(__AES__)
-    fugue512_Update( &ctx.fugue, hash, dataLen*8 );
-    fugue512_Final( &ctx.fugue, hash ); 
-#else
-	sph_fugue512(&ctx.fugue, hash, dataLen);
-	sph_fugue512_close(&ctx.fugue, hash);
-#endif
-
-	sph_shabal512(&ctx.shabal, hash, dataLen);
-	sph_shabal512_close(&ctx.shabal, hash);
-
-	sph_whirlpool(&ctx.whirlpool, hash, dataLen);
-	sph_whirlpool_close(&ctx.whirlpool, hash);
-
-   SHA512_Update( &ctx.sha512, hash, dataLen );
-   SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
-
-	sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
-	sph_haval256_5_close(&ctx.haval, hash);
-
-	memset(&hash[8], 0, dataLen - 32);
-
-   memcpy( &ctx, &xevan_ctx, sizeof(xevan_ctx) );
-
-	sph_blake512(&ctx.blake, hash, dataLen);
-	sph_blake512_close(&ctx.blake, hash);
-
-	sph_bmw512(&ctx.bmw, hash, dataLen);
-	sph_bmw512_close(&ctx.bmw, hash);
-
-#if defined(__AES__)
-   update_and_final_groestl( &ctx.groestl, (char*)hash,
-                              (const BitSequence*)hash, dataLen*8 );
-#else
-	sph_groestl512(&ctx.groestl, hash, dataLen);
+   sph_groestl512(&ctx.groestl, hash, dataLen);
   sph_groestl512_close(&ctx.groestl, hash);
 #endif

-	sph_skein512(&ctx.skein, hash, dataLen);
-	sph_skein512_close(&ctx.skein, hash);
+   sph_skein512(&ctx.skein, hash, dataLen);
+   sph_skein512_close(&ctx.skein, hash);

-	sph_jh512(&ctx.jh, hash, dataLen);
-	sph_jh512_close(&ctx.jh, hash);
+   sph_jh512(&ctx.jh, hash, dataLen);
+   sph_jh512_close(&ctx.jh, hash);

-	sph_keccak512(&ctx.keccak, hash, dataLen);
-	sph_keccak512_close(&ctx.keccak, hash);
+   sph_keccak512(&ctx.keccak, hash, dataLen);
+   sph_keccak512_close(&ctx.keccak, hash);

   update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
                                 (const BitSequence*)hash, dataLen );
@@ -193,8 +123,8 @@ int xevan_hash(void *output, const void *input, int thr_id )
   cubehashUpdateDigest( &ctx.cubehash, (byte*)hash,
                                 (const byte*) hash, dataLen );

-	sph_shavite512(&ctx.shavite, hash, dataLen);
-	sph_shavite512_close(&ctx.shavite, hash);
+   sph_shavite512(&ctx.shavite, hash, dataLen);
+   sph_shavite512_close(&ctx.shavite, hash);

   update_final_sd( &ctx.simd, (BitSequence *)hash,
                         (const BitSequence *)hash, dataLen*8 );
@@ -207,30 +137,100 @@ int xevan_hash(void *output, const void *input, int thr_id )
   sph_echo512_close(&ctx.echo, hash);
 #endif

-	sph_hamsi512(&ctx.hamsi, hash, dataLen);
-	sph_hamsi512_close(&ctx.hamsi, hash);
+   sph_hamsi512(&ctx.hamsi, hash, dataLen);
+   sph_hamsi512_close(&ctx.hamsi, hash);

 #if defined(__AES__)
-    fugue512_Update( &ctx.fugue, hash, dataLen*8 );
-    fugue512_Final( &ctx.fugue, hash );   
+   fugue512_Update( &ctx.fugue, hash, dataLen*8 );
+   fugue512_Final( &ctx.fugue, hash ); 
 #else
-	sph_fugue512(&ctx.fugue, hash, dataLen);
-	sph_fugue512_close(&ctx.fugue, hash);
+   sph_fugue512(&ctx.fugue, hash, dataLen);
+   sph_fugue512_close(&ctx.fugue, hash);
 #endif

-	sph_shabal512(&ctx.shabal, hash, dataLen);
-	sph_shabal512_close(&ctx.shabal, hash);
+   sph_shabal512(&ctx.shabal, hash, dataLen);
+   sph_shabal512_close(&ctx.shabal, hash);

-	sph_whirlpool(&ctx.whirlpool, hash, dataLen);
-	sph_whirlpool_close(&ctx.whirlpool, hash);
+   sph_whirlpool(&ctx.whirlpool, hash, dataLen);
+   sph_whirlpool_close(&ctx.whirlpool, hash);

-   SHA512_Update( &ctx.sha512, hash, dataLen );
-   SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+   sph_sha512( &ctx.sha512, hash, dataLen );
+   sph_sha512_close( &ctx.sha512, hash );

-	sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
-	sph_haval256_5_close(&ctx.haval, hash);
+   sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
+   sph_haval256_5_close(&ctx.haval, hash);

-	memcpy(output, hash, 32);
+   memset(&hash[8], 0, dataLen - 32);
+
+   memcpy( &ctx, &xevan_ctx, sizeof(xevan_ctx) );
+
+   sph_blake512(&ctx.blake, hash, dataLen);
+   sph_blake512_close(&ctx.blake, hash);
+
+   sph_bmw512(&ctx.bmw, hash, dataLen);
+   sph_bmw512_close(&ctx.bmw, hash);
+
+#if defined(__AES__)
+   update_and_final_groestl( &ctx.groestl, (char*)hash,
+                              (const BitSequence*)hash, dataLen*8 );
+#else
+   sph_groestl512(&ctx.groestl, hash, dataLen);
+   sph_groestl512_close(&ctx.groestl, hash);
+#endif
+
+   sph_skein512(&ctx.skein, hash, dataLen);
+   sph_skein512_close(&ctx.skein, hash);
+
+   sph_jh512(&ctx.jh, hash, dataLen);
+   sph_jh512_close(&ctx.jh, hash);
+
+   sph_keccak512(&ctx.keccak, hash, dataLen);
+   sph_keccak512_close(&ctx.keccak, hash);
+
+   update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
+                                 (const BitSequence*)hash, dataLen );
+
+   cubehashUpdateDigest( &ctx.cubehash, (byte*)hash,
+                                 (const byte*) hash, dataLen );
+
+   sph_shavite512(&ctx.shavite, hash, dataLen);
+   sph_shavite512_close(&ctx.shavite, hash);
+
+   update_final_sd( &ctx.simd, (BitSequence *)hash,
+                         (const BitSequence *)hash, dataLen*8 );
+
+#if defined(__AES__)
+   update_final_echo( &ctx.echo, (BitSequence *) hash,
+                           (const BitSequence *) hash, dataLen*8 );
+#else
+   sph_echo512(&ctx.echo, hash, dataLen);
+   sph_echo512_close(&ctx.echo, hash);
+#endif
+
+   sph_hamsi512(&ctx.hamsi, hash, dataLen);
+   sph_hamsi512_close(&ctx.hamsi, hash);
+
+#if defined(__AES__)
+   fugue512_Update( &ctx.fugue, hash, dataLen*8 );
+   fugue512_Final( &ctx.fugue, hash );   
+#else
+   sph_fugue512(&ctx.fugue, hash, dataLen);
+   sph_fugue512_close(&ctx.fugue, hash);
+#endif
+
+   sph_shabal512(&ctx.shabal, hash, dataLen);
+   sph_shabal512_close(&ctx.shabal, hash);
+
+   sph_whirlpool(&ctx.whirlpool, hash, dataLen);
+   sph_whirlpool_close(&ctx.whirlpool, hash);
+
+   sph_sha512( &ctx.sha512, hash, dataLen );
+   sph_sha512_close( &ctx.sha512, hash );
+
+   sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
+   sph_haval256_5_close(&ctx.haval, hash);
+
+   memcpy(output, hash, 32);

   return 1;
 }
--- a/algo/x20/x20r.c
+++ b/algo/x20/x20r.c
@@ -18,7 +18,7 @@
 #include "algo/radiogatun/sph_radiogatun.h"
 #include "algo/panama/sph_panama.h"
 #include "algo/gost/sph_gost.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #if defined(__AES__)
  #include "algo/echo/aes_ni/hash_api.h"
  #include "algo/groestl/aes_ni/hash-groestl.h"
@@ -56,7 +56,7 @@ union _x20r_context_overlay
    sph_fugue512_context     fugue;
    sph_shabal512_context    shabal;
    sph_whirlpool_context    whirlpool;
-    SHA512_CTX               sha512;
+    sph_sha512_context       sha512;
    sph_haval256_5_context   haval;
    sph_gost512_context      gost;
    sph_radiogatun64_context radiogatun;
@@ -68,28 +68,6 @@ void x20r_hash(void* output, const void* input)
 {
   uint32_t _ALIGN(128) hash[64/4];
   x20r_context_overlay ctx;
-/*
-	sph_blake512_context     ctx_blake;
-	sph_bmw512_context       ctx_bmw;
-	sph_groestl512_context   ctx_groestl;
-	sph_skein512_context     ctx_skein;
-	sph_jh512_context        ctx_jh;
-	sph_keccak512_context    ctx_keccak;
-	sph_luffa512_context     ctx_luffa;
-	sph_cubehash512_context  ctx_cubehash;
-	sph_shavite512_context   ctx_shavite;
-	sph_simd512_context      ctx_simd;
-	sph_echo512_context      ctx_echo;
-	sph_hamsi512_context     ctx_hamsi;
-	sph_fugue512_context     ctx_fugue;
-	sph_shabal512_context    ctx_shabal;
-	sph_whirlpool_context    ctx_whirlpool;
-	sph_sha512_context       ctx_sha512;
-	sph_haval256_5_context   ctx_haval;
-	sph_gost512_context      ctx_gost;
-	sph_radiogatun64_context ctx_radiogatun;
-	sph_panama_context       ctx_panama;
-*/
   void *in = (void*) input;
   int size = 80;

@@ -194,9 +172,9 @@ void x20r_hash(void* output, const void* input)
 		sph_whirlpool_close(&ctx.whirlpool, hash);
 		break;
 	   case SHA_512:
-                SHA512_Init( &ctx.sha512 );
-                SHA512_Update( &ctx.sha512, in, size );
-                SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+                sph_sha512_Init( &ctx.sha512 );
+                sph_sha512( &ctx.sha512, in, size );
+                sph_sha512_close( &ctx.sha512, hash );
 		break;
 	   case HAVAL:
 		sph_haval256_5_init(&ctx.haval);
--- a/algo/x22/x22i-4way.c
+++ b/algo/x22/x22i-4way.c
@@ -11,7 +11,7 @@
 #include "algo/shavite/shavite-hash-2way.h"
 #include "algo/shavite/sph_shavite.h"
 #include "algo/simd/simd-hash-2way.h"
-#include "algo/shavite/sph_shavite.h"
+#include "algo/shavite/shavite-hash-2way.h"
 #include "algo/hamsi/hamsi-hash-4way.h"
 #include "algo/fugue/fugue-aesni.h"
 #include "algo/shabal/shabal-hash-4way.h"
@@ -27,7 +27,9 @@
  #include "algo/shavite/shavite-hash-4way.h"
  #include "algo/echo/echo-hash-4way.h"
 #endif
-
+#if defined(__SHA__)
+  #include "algo/sha/sph_sha2.h"
+#endif

 #if defined(X22I_8WAY)

@@ -49,7 +51,11 @@ union _x22i_8way_ctx_overlay
    haval256_5_8way_context haval;
    sph_tiger_context       tiger;
    sph_gost512_context     gost;
+#if defined(X22I_8WAY_SHA)
+    sph_sha256_context      sha256;
+#else
    sha256_8way_context     sha256;
+#endif
 #if defined(__VAES__)
    groestl512_4way_context groestl;
    shavite512_4way_context shavite;
@@ -383,6 +389,35 @@ int x22i_8way_hash( void *output, const void *input, int thrid )
   sph_gost512 ( &ctx.gost, (const void*) hash7, 64 );
   sph_gost512_close( &ctx.gost, (void*) hash7 );

+#if defined(X22I_8WAY_SHA)
+
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash0, 64 );
+   sph_sha256_close( &ctx.sha256, output );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash1, 64 );
+   sph_sha256_close( &ctx.sha256, output+32 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash2, 64 );
+   sph_sha256_close( &ctx.sha256, output+64 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash3, 64 );
+   sph_sha256_close( &ctx.sha256, output+96 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash4, 64 );
+   sph_sha256_close( &ctx.sha256, output+128 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash5, 64 );
+   sph_sha256_close( &ctx.sha256, output+160 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash6, 64 );
+   sph_sha256_close( &ctx.sha256, output+192 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash7, 64 );
+   sph_sha256_close( &ctx.sha256, output+224 );
+   
+#else
+
   intrlv_8x32_512( vhash, hash0, hash1, hash2, hash3,
                           hash4, hash5, hash6, hash7 );

@@ -390,9 +425,55 @@ int x22i_8way_hash( void *output, const void *input, int thrid )
   sha256_8way_update( &ctx.sha256, vhash, 64 );
   sha256_8way_close( &ctx.sha256, output );

+#endif
+
   return 1;
 }

+#if defined(X22I_8WAY_SHA)
+
+int scanhash_x22i_8way_sha( struct work *work, uint32_t max_nonce,
+                       uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t hash[8*8] __attribute__ ((aligned (128)));
+   uint32_t vdata[20*8] __attribute__ ((aligned (64)));
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
+   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 8;
+   __m512i  *noncev = (__m512i*)vdata + 9;
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+
+   if ( bench )  ptarget[7] = 0x08ff;
+
+   InitializeSWIFFTX();
+
+   mm512_bswap32_intrlv80_8x64( vdata, pdata );
+   *noncev = mm512_intrlv_blend_32(
+              _mm512_set_epi32( n+7, 0, n+6, 0, n+5, 0, n+4, 0,
+                                n+3, 0, n+2, 0, n+1, 0, n,   0 ), *noncev );
+   do
+   {
+      if ( x22i_8way_hash( hash, vdata, thr_id ) )
+      for ( int i = 0; i < 8; i++ )
+      if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
+      {
+         pdata[19] = bswap_32( n+i );
+         submit_solution( work, hash+(i<<3), mythr );
+      }
+      *noncev = _mm512_add_epi32( *noncev,
+                                  m512_const1_64( 0x0000000800000000 ) );
+      n += 8;
+   } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+
+#else
+
 int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
                       uint64_t *hashes_done, struct thr_info *mythr )
 {
@@ -440,53 +521,7 @@ int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
   return 0;
 }

-/*
-int scanhash_x22i_8way( struct work* work, uint32_t max_nonce,
-                   uint64_t *hashes_done, struct thr_info *mythr )
-{
-   uint32_t hash[8*16] __attribute__ ((aligned (128)));
-   uint32_t vdata[24*8] __attribute__ ((aligned (64)));
-   uint32_t lane_hash[8] __attribute__ ((aligned (64)));
-   uint32_t *hash7 = &(hash[7<<3]);
-   uint32_t *pdata = work->data;
-   uint32_t *ptarget = work->target;
-   const uint32_t first_nonce = pdata[19];
-   __m512i  *noncev = (__m512i*)vdata + 9;   // aligned
-   uint32_t n = first_nonce;
-   const uint32_t last_nonce = max_nonce - 8;
-   const int thr_id = mythr->id;
-   const uint32_t Htarg = ptarget[7];
-
-   if (opt_benchmark)
-      ((uint32_t*)ptarget)[7] = 0x08ff;
-
-   InitializeSWIFFTX();
-
-   mm512_bswap32_intrlv80_8x64( vdata, pdata );
-   do
-   {
-      *noncev = mm512_intrlv_blend_32( mm512_bswap_32(
-               _mm512_set_epi32( n+7, 0, n+6, 0, n+5, 0, n+4, 0,
-                                 n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
-      x22i_8way_hash( hash, vdata );
-
-      for ( int lane = 0; lane < 8; lane++ )
-      if unlikely( ( hash7[ lane ] <= Htarg ) )
-      {
-         extr_lane_8x32( lane_hash, hash, lane, 256 );
-         if ( likely( fulltest( lane_hash, ptarget ) && !opt_benchmark ) )
-         {
-            pdata[19] = n + lane;
-            submit_solution( work, lane_hash, mythr );
-         }
-      }
-      n += 8;
-   } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
-
-   *hashes_done = n - first_nonce;
-   return 0;
-}
-*/
+#endif

 #elif defined(X22I_4WAY)

@@ -494,14 +529,19 @@ union _x22i_4way_ctx_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    echo_2way_context       echo;
+#else
    hashState_groestl       groestl;
    hashState_echo          echo;
+#endif
+    shavite512_2way_context shavite;
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
    luffa_2way_context      luffa;
    cube_2way_context       cube;
-    shavite512_2way_context shavite;
    simd_2way_context       simd;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
@@ -511,7 +551,11 @@ union _x22i_4way_ctx_overlay
    haval256_5_4way_context haval;
    sph_tiger_context       tiger;
    sph_gost512_context     gost;
+#if defined(X22I_4WAY_SHA)
+    sph_sha256_context      sha256;
+#else
    sha256_4way_context     sha256;
+#endif
 };
 typedef union _x22i_4way_ctx_overlay x22i_ctx_overlay;

@@ -535,15 +579,29 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   bmw512_4way_init( &ctx.bmw );
   bmw512_4way_update( &ctx.bmw, vhash, 64 );
   bmw512_4way_close( &ctx.bmw, vhash );
+
+#if defined(__VAES__)
+
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+   groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+   groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
   dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );

-   groestl512_full( &ctx.groestl, (char*)hash0, (const char*)hash0, 512 );
-   groestl512_full( &ctx.groestl, (char*)hash1, (const char*)hash1, 512 );
-   groestl512_full( &ctx.groestl, (char*)hash2, (const char*)hash2, 512 );
-   groestl512_full( &ctx.groestl, (char*)hash3, (const char*)hash3, 512 );
+   groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
+   groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
+   groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
+   groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );

   intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
   skein512_4way_full( &ctx.skein, vhash, vhash, 64 );

   jh512_4way_init( &ctx.jh );
@@ -570,6 +628,15 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
   simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );

+#if defined(__VAES__)
+
+   echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+   echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
   dintrlv_2x128_512( hash0, hash1, vhashA );
   dintrlv_2x128_512( hash2, hash3, vhashB );
   
@@ -584,6 +651,8 @@ int x22i_4way_hash( void *output, const void *input, int thrid )

   intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );

+#endif
+
   if ( work_restart[thrid].restart ) return false;
   
   hamsi512_4way_init( &ctx.hamsi );
@@ -625,7 +694,7 @@ int x22i_4way_hash( void *output, const void *input, int thrid )

   if ( work_restart[thrid].restart ) return false;
   
-	ComputeSingleSWIFFTX((unsigned char*)hash0, (unsigned char*)hashA0);
+   ComputeSingleSWIFFTX((unsigned char*)hash0, (unsigned char*)hashA0);
   ComputeSingleSWIFFTX((unsigned char*)hash1, (unsigned char*)hashA1);
   ComputeSingleSWIFFTX((unsigned char*)hash2, (unsigned char*)hashA2);
   ComputeSingleSWIFFTX((unsigned char*)hash3, (unsigned char*)hashA3);
@@ -639,7 +708,7 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   haval256_5_4way_close( &ctx.haval, vhash );
   dintrlv_4x32_512( hash0, hash1, hash2, hash3, vhash );
     
-	memset( hashA0, 0, 64 );
+   memset( hashA0, 0, 64 );
   memset( hashA1, 0, 64 );
   memset( hashA2, 0, 64 );
   memset( hashA3, 0, 64 );
@@ -654,8 +723,8 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   sph_tiger (&ctx.tiger, (const void*) hash2, 64);
   sph_tiger_close(&ctx.tiger, (void*) hashA2);
   sph_tiger_init(&ctx.tiger);
-	sph_tiger (&ctx.tiger, (const void*) hash3, 64);
-	sph_tiger_close(&ctx.tiger, (void*) hashA3);
+   sph_tiger (&ctx.tiger, (const void*) hash3, 64);
+   sph_tiger_close(&ctx.tiger, (void*) hashA3);

   if ( work_restart[thrid].restart ) return false;

@@ -682,9 +751,26 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   sph_gost512_init(&ctx.gost);
   sph_gost512 (&ctx.gost, (const void*) hash2, 64);
   sph_gost512_close(&ctx.gost, (void*) hash2);
-	sph_gost512_init(&ctx.gost);
-	sph_gost512 (&ctx.gost, (const void*) hash3, 64);
-	sph_gost512_close(&ctx.gost, (void*) hash3);
+   sph_gost512_init(&ctx.gost);
+   sph_gost512 (&ctx.gost, (const void*) hash3, 64);
+   sph_gost512_close(&ctx.gost, (void*) hash3);
+
+#if defined(X22I_4WAY_SHA)
+
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash0, 64 );
+   sph_sha256_close( &ctx.sha256, output );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash1, 64 );
+   sph_sha256_close( &ctx.sha256, output+32 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash2, 64 );
+   sph_sha256_close( &ctx.sha256, output+64 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash3, 64 );
+   sph_sha256_close( &ctx.sha256, output+96 );
+
+#else

   intrlv_4x32_512( vhash, hash0, hash1, hash2, hash3 );

@@ -692,11 +778,56 @@ int x22i_4way_hash( void *output, const void *input, int thrid )
   sha256_4way_update( &ctx.sha256, vhash, 64 );
   sha256_4way_close( &ctx.sha256, output );

+#endif
+
   return 1;
 }

+#if defined(X22I_4WAY_SHA)
+
+int scanhash_x22i_4way_sha( struct work* work, uint32_t max_nonce,
+                            uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t hash[8*4] __attribute__ ((aligned (64)));
+   uint32_t vdata[20*4] __attribute__ ((aligned (64)));
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
+   const uint32_t first_nonce = pdata[19];
+   const uint32_t last_nonce = max_nonce - 4;
+   __m256i  *noncev = (__m256i*)vdata + 9;
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+
+   if ( bench ) ptarget[7] = 0x08ff;
+
+   InitializeSWIFFTX();
+
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+   *noncev = mm256_intrlv_blend_32(
+                   _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
+   do
+   {
+      if ( x22i_4way_hash( hash, vdata, thr_id ) )
+      for ( int i = 0; i < 4; i++ )
+      if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
+      {
+         pdata[19] = bswap_32( n+i );
+         submit_solution( work, hash+(i<<3), mythr );
+      }
+      *noncev = _mm256_add_epi32( *noncev,
+                                  m256_const1_64( 0x0000000400000000 ) );
+      n += 4;
+   } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+
+#else
+
 int scanhash_x22i_4way( struct work* work, uint32_t max_nonce,
-                   uint64_t *hashes_done, struct thr_info *mythr )
+                        uint64_t *hashes_done, struct thr_info *mythr )
 {
   uint32_t hash[8*4] __attribute__ ((aligned (64)));
   uint32_t vdata[20*4] __attribute__ ((aligned (64)));
@@ -741,4 +872,6 @@ int scanhash_x22i_4way( struct work* work, uint32_t max_nonce,
   return 0;
 }

+#endif
+
 #endif  // X22I_4WAY
--- a/algo/x22/x22i-gate.c
+++ b/algo/x22/x22i-gate.c
@@ -7,21 +7,32 @@
 bool register_x22i_algo( algo_gate_t* gate )
 {
 #if defined (X22I_8WAY)
-  gate->scanhash  = (void*)&scanhash_x22i_8way;
-  gate->hash      = (void*)&x22i_8way_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT
-                      | AVX512_OPT | VAES_OPT;
-#elif defined (X22I_4WAY)
-  gate->scanhash  = (void*)&scanhash_x22i_4way;
-  gate->hash      = (void*)&x22i_4way_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                      | AVX512_OPT | VAES_OPT;
+
+#if defined(X22I_8WAY_SHA)
+  gate->scanhash  = (void*)&scanhash_x22i_8way_sha;
 #else
+  gate->scanhash  = (void*)&scanhash_x22i_8way;
+#endif
+  gate->hash      = (void*)&x22i_8way_hash;
+
+#elif defined (X22I_4WAY)
+
+#if defined(X22I_4WAY_SHA)
+  gate->scanhash  = (void*)&scanhash_x22i_4way_sha;
+#else
+  gate->scanhash  = (void*)&scanhash_x22i_4way;
+#endif
+  gate->hash      = (void*)&x22i_4way_hash;
+
+#else
+
  gate->scanhash  = (void*)&scanhash_x22i;
  gate->hash      = (void*)&x22i_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                      | AVX512_OPT | VAES_OPT;
+
 #endif
+
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
+                      | AVX512_OPT | VAES_OPT | VAES256_OPT;
  return true;
 };

@@ -30,21 +41,15 @@ bool register_x25x_algo( algo_gate_t* gate )
 #if defined (X25X_8WAY)
  gate->scanhash  = (void*)&scanhash_x25x_8way;
  gate->hash      = (void*)&x25x_8way_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT
-                      | AVX512_OPT | VAES_OPT;
 #elif defined (X25X_4WAY)
  gate->scanhash  = (void*)&scanhash_x25x_4way;
  gate->hash      = (void*)&x25x_4way_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                      | AVX512_OPT | VAES_OPT;
 #else
  gate->scanhash  = (void*)&scanhash_x25x;
  gate->hash      = (void*)&x25x_hash;
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
-                      | AVX512_OPT | VAES_OPT;
 #endif
-//  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT;
-
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT |
+                        AVX512_OPT | VAES_OPT | VAES256_OPT;
  return true;
 };

--- a/algo/x22/x22i-gate.h
+++ b/algo/x22/x22i-gate.h
@@ -12,19 +12,34 @@
  #define X22I_4WAY 1
 #endif

+#if defined(__SHA__)
+//  #define X22I_8WAY_SHA 1
+  #define X22I_4WAY_SHA 1
+#endif
+
 bool register_x22i_algo( algo_gate_t* gate );

 #if defined(X22I_8WAY)

 int x22i_8way_hash( void *state, const void *input, int thrid );
+#if defined(X22I_8WAY_SHA)
+int scanhash_x22i_8way_sha( struct work *work, uint32_t max_nonce,
+                        uint64_t *hashes_done, struct thr_info *mythr );
+#else
 int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
                        uint64_t *hashes_done, struct thr_info *mythr );
+#endif

 #elif defined(X22I_4WAY)

 int x22i_4way_hash( void *state, const void *input, int thrid );
+#if defined(X22I_4WAY_SHA)
+int scanhash_x22i_4way_sha( struct work *work, uint32_t max_nonce,
+                            uint64_t *hashes_done, struct thr_info *mythr );
+#else
 int scanhash_x22i_4way( struct work *work, uint32_t max_nonce,
                        uint64_t *hashes_done, struct thr_info *mythr );
+#endif

 #else

@@ -40,6 +55,11 @@ int scanhash_x22i( struct work *work, uint32_t max_nonce,
  #define X25X_4WAY 1
 #endif

+#if defined(__SHA__)
+//  #define X25X_8WAY_SHA 1
+  #define X25X_4WAY_SHA 1
+#endif
+
 bool register_x25i_algo( algo_gate_t* gate );

 #if defined(X25X_8WAY)
--- a/algo/x22/x22i.c
+++ b/algo/x22/x22i.c
@@ -23,7 +23,7 @@
 #include "algo/hamsi/sph_hamsi.h"
 #include "algo/shabal/sph_shabal.h"
 #include "algo/whirlpool/sph_whirlpool.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #include "algo/haval/sph-haval.h"
 #include "algo/tiger/sph_tiger.h"
 #include "algo/lyra2/lyra2.h"
@@ -53,11 +53,11 @@ union _x22i_context_overlay
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_haval256_5_context  haval;
        sph_tiger_context       tiger;
        sph_gost512_context     gost;
-        SHA256_CTX              sha256;
+        sph_sha256_context      sha256;
 };
 typedef union _x22i_context_overlay x22i_context_overlay;

@@ -67,13 +67,13 @@ int x22i_hash( void *output, const void *input, int thrid )
   unsigned char hash2[65]    __attribute__((aligned(64))) = {0};
   x22i_context_overlay ctx;

-	sph_blake512_init(&ctx.blake);
-	sph_blake512(&ctx.blake, input, 80);
-	sph_blake512_close(&ctx.blake, hash);
+   sph_blake512_init(&ctx.blake);
+   sph_blake512(&ctx.blake, input, 80);
+   sph_blake512_close(&ctx.blake, hash);

-	sph_bmw512_init(&ctx.bmw);
-	sph_bmw512(&ctx.bmw, (const void*) hash, 64);
-	sph_bmw512_close(&ctx.bmw, hash);
+   sph_bmw512_init(&ctx.bmw);
+   sph_bmw512(&ctx.bmw, (const void*) hash, 64);
+   sph_bmw512_close(&ctx.bmw, hash);

 #if defined(__AES__)
   init_groestl( &ctx.groestl, 64 );
@@ -85,17 +85,17 @@ int x22i_hash( void *output, const void *input, int thrid )
   sph_groestl512_close( &ctx.groestl, hash );
 #endif
   
-	sph_skein512_init(&ctx.skein);
-	sph_skein512(&ctx.skein, (const void*) hash, 64);
-	sph_skein512_close(&ctx.skein, hash);
+   sph_skein512_init(&ctx.skein);
+   sph_skein512(&ctx.skein, (const void*) hash, 64);
+   sph_skein512_close(&ctx.skein, hash);

-	sph_jh512_init(&ctx.jh);
-	sph_jh512(&ctx.jh, (const void*) hash, 64);
-	sph_jh512_close(&ctx.jh, hash);
+   sph_jh512_init(&ctx.jh);
+   sph_jh512(&ctx.jh, (const void*) hash, 64);
+   sph_jh512_close(&ctx.jh, hash);

-	sph_keccak512_init(&ctx.keccak);
-	sph_keccak512(&ctx.keccak, (const void*) hash, 64);
-	sph_keccak512_close(&ctx.keccak, hash);
+   sph_keccak512_init(&ctx.keccak);
+   sph_keccak512(&ctx.keccak, (const void*) hash, 64);
+   sph_keccak512_close(&ctx.keccak, hash);

   if ( work_restart[thrid].restart ) return 0;
   
@@ -107,9 +107,9 @@ int x22i_hash( void *output, const void *input, int thrid )
   cubehashUpdateDigest( &ctx.cube, (byte*) hash,
                              (const byte*)hash, 64 );

-	sph_shavite512_init(&ctx.shavite);
-	sph_shavite512(&ctx.shavite, (const void*) hash, 64);
-	sph_shavite512_close(&ctx.shavite, hash);
+   sph_shavite512_init(&ctx.shavite);
+   sph_shavite512(&ctx.shavite, (const void*) hash, 64);
+   sph_shavite512_close(&ctx.shavite, hash);

   init_sd( &ctx.simd, 512 );
   update_final_sd( &ctx.simd, (BitSequence*)hash,
@@ -127,56 +127,56 @@ int x22i_hash( void *output, const void *input, int thrid )

   if ( work_restart[thrid].restart ) return 0;
   
-	sph_hamsi512_init(&ctx.hamsi);
-	sph_hamsi512(&ctx.hamsi, (const void*) hash, 64);
-	sph_hamsi512_close(&ctx.hamsi, hash);
+   sph_hamsi512_init(&ctx.hamsi);
+   sph_hamsi512(&ctx.hamsi, (const void*) hash, 64);
+   sph_hamsi512_close(&ctx.hamsi, hash);

 #if defined(__AES__)
-        fugue512_full( &ctx.fugue, hash, hash, 64 );
+   fugue512_full( &ctx.fugue, hash, hash, 64 );
 #else
-	sph_fugue512_init(&ctx.fugue);
-	sph_fugue512(&ctx.fugue, (const void*) hash, 64);
-	sph_fugue512_close(&ctx.fugue, hash);
+   sph_fugue512_init(&ctx.fugue);
+   sph_fugue512(&ctx.fugue, (const void*) hash, 64);
+   sph_fugue512_close(&ctx.fugue, hash);
 #endif

-	sph_shabal512_init(&ctx.shabal);
-	sph_shabal512(&ctx.shabal, (const void*) hash, 64);
-	sph_shabal512_close(&ctx.shabal, &hash[64]);
+   sph_shabal512_init(&ctx.shabal);
+   sph_shabal512(&ctx.shabal, (const void*) hash, 64);
+   sph_shabal512_close(&ctx.shabal, &hash[64]);

-	sph_whirlpool_init(&ctx.whirlpool);
-	sph_whirlpool (&ctx.whirlpool, (const void*) &hash[64], 64);
-	sph_whirlpool_close(&ctx.whirlpool, &hash[128]);
+   sph_whirlpool_init(&ctx.whirlpool);
+   sph_whirlpool (&ctx.whirlpool, (const void*) &hash[64], 64);
+   sph_whirlpool_close(&ctx.whirlpool, &hash[128]);

-   SHA512_Init( &ctx.sha512 );
-   SHA512_Update(  &ctx.sha512, (const void*) &hash[128], 64);
-   SHA512_Final( (void*) &hash[192], &ctx.sha512 );
-
-	ComputeSingleSWIFFTX((unsigned char*)hash, (unsigned char*)hash2);
+   sph_sha512_init( &ctx.sha512 );
+   sph_sha512( &ctx.sha512, &hash[128], 64 );
+   sph_sha512_close( &ctx.sha512, &hash[192] );
+   
+   ComputeSingleSWIFFTX((unsigned char*)hash, (unsigned char*)hash2);

   if ( work_restart[thrid].restart ) return 0;
   
-	memset(hash, 0, 64);
-	sph_haval256_5_init(&ctx.haval);
-	sph_haval256_5(&ctx.haval,(const void*) hash2, 64);
-	sph_haval256_5_close(&ctx.haval,hash);
+   memset(hash, 0, 64);
+   sph_haval256_5_init(&ctx.haval);
+   sph_haval256_5(&ctx.haval,(const void*) hash2, 64);
+   sph_haval256_5_close(&ctx.haval,hash);

-	memset(hash2, 0, 64);
-	sph_tiger_init(&ctx.tiger);
-	sph_tiger (&ctx.tiger, (const void*) hash, 64);
-	sph_tiger_close(&ctx.tiger, (void*) hash2);
+   memset(hash2, 0, 64);
+   sph_tiger_init(&ctx.tiger);
+   sph_tiger (&ctx.tiger, (const void*) hash, 64);
+   sph_tiger_close(&ctx.tiger, (void*) hash2);

-	memset(hash, 0, 64);
-	LYRA2RE((void*) hash, 32, (const void*) hash2, 32, (const void*) hash2, 32, 1, 4, 4);
+   memset(hash, 0, 64);
+   LYRA2RE((void*) hash, 32, (const void*) hash2, 32, (const void*) hash2, 32, 1, 4, 4);

-	sph_gost512_init(&ctx.gost);
-	sph_gost512 (&ctx.gost, (const void*) hash, 64);
-	sph_gost512_close(&ctx.gost, (void*) hash);
+   sph_gost512_init(&ctx.gost);
+   sph_gost512 (&ctx.gost, (const void*) hash, 64);
+   sph_gost512_close(&ctx.gost, (void*) hash);

-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update(  &ctx.sha256, (const void*) hash, 64 );
-   SHA256_Final( (unsigned char*) hash, &ctx.sha256 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash, 64 );
+   sph_sha256_close( &ctx.sha256, hash );

-	memcpy(output, hash, 32);
+   memcpy(output, hash, 32);

   return 1;
 }
--- a/algo/x22/x25x-4way.c
+++ b/algo/x22/x25x-4way.c
@@ -15,6 +15,7 @@
 #include "algo/cubehash/cubehash_sse2.h"
 #include "algo/luffa/luffa-hash-2way.h"
 #include "algo/cubehash/cube-hash-2way.h"
+#include "algo/shavite/shavite-hash-2way.h"
 #include "algo/shavite/sph_shavite.h"
 #include "algo/simd/nist.h"
 #include "algo/simd/simd-hash-2way.h"
@@ -31,6 +32,9 @@
  #include "algo/shavite/shavite-hash-4way.h"
  #include "algo/echo/echo-hash-4way.h"
 #endif
+#if defined(__SHA__)
+  #include "algo/sha/sph_sha2.h"
+#endif

 void x25x_shuffle( void *hash )
 {
@@ -79,7 +83,11 @@ union _x25x_8way_ctx_overlay
    haval256_5_8way_context haval;
    sph_tiger_context       tiger;
    sph_gost512_context     gost;
+#if defined(X25X_8WAY_SHA)
+    sph_sha256_context      sha256;
+#else
    sha256_8way_context     sha256;
+#endif
    panama_8way_context     panama;
    blake2s_8way_state      blake2s;
 #if defined(__VAES__)
@@ -215,9 +223,9 @@ int x25x_8way_hash( void *output, const void *input, int thrid )

 #else

-	sph_shavite512_init(&ctx.shavite);
-	sph_shavite512(&ctx.shavite, (const void*) hash0[7], 64);
-	sph_shavite512_close(&ctx.shavite, hash0[8]);
+   sph_shavite512_init(&ctx.shavite);
+   sph_shavite512(&ctx.shavite, (const void*) hash0[7], 64);
+   sph_shavite512_close(&ctx.shavite, hash0[8]);
   sph_shavite512_init(&ctx.shavite);
   sph_shavite512(&ctx.shavite, (const void*) hash1[7], 64);
   sph_shavite512_close(&ctx.shavite, hash1[8]);
@@ -321,9 +329,9 @@ int x25x_8way_hash( void *output, const void *input, int thrid )
   dintrlv_8x32_512( hash0[13], hash1[13], hash2[13], hash3[13],
                     hash4[13], hash5[13], hash6[13], hash7[13], vhash );

-	sph_whirlpool_init(&ctx.whirlpool);
-	sph_whirlpool (&ctx.whirlpool, (const void*) hash0[13], 64);
-	sph_whirlpool_close(&ctx.whirlpool, hash0[14]);
+   sph_whirlpool_init(&ctx.whirlpool);
+   sph_whirlpool (&ctx.whirlpool, (const void*) hash0[13], 64);
+   sph_whirlpool_close(&ctx.whirlpool, hash0[14]);
   sph_whirlpool_init(&ctx.whirlpool);
   sph_whirlpool (&ctx.whirlpool, (const void*) hash1[13], 64);
   sph_whirlpool_close(&ctx.whirlpool, hash1[14]);
@@ -372,9 +380,9 @@ int x25x_8way_hash( void *output, const void *input, int thrid )
   dintrlv_8x32_512( hash0[17], hash1[17], hash2[17], hash3[17],
                     hash4[17], hash5[17], hash6[17], hash7[17], vhash );

-	sph_tiger_init(&ctx.tiger);
-	sph_tiger (&ctx.tiger, (const void*) hash0[17], 64);
-	sph_tiger_close(&ctx.tiger, (void*) hash0[18]);
+   sph_tiger_init(&ctx.tiger);
+   sph_tiger (&ctx.tiger, (const void*) hash0[17], 64);
+   sph_tiger_close(&ctx.tiger, (void*) hash0[18]);
   sph_tiger_init(&ctx.tiger);
   sph_tiger (&ctx.tiger, (const void*) hash1[17], 64);
   sph_tiger_close(&ctx.tiger, (void*) hash1[18]);
@@ -412,9 +420,9 @@ int x25x_8way_hash( void *output, const void *input, int thrid )
   LYRA2X_2WAY( vhash, 32, vhash, 32, 1, 4, 4 );
   dintrlv_2x256( hash6[19], hash7[19], vhash, 256 );

-	sph_gost512_init(&ctx.gost);
-	sph_gost512 (&ctx.gost, (const void*) hash0[19], 64);
-	sph_gost512_close(&ctx.gost, (void*) hash0[20]);
+   sph_gost512_init(&ctx.gost);
+   sph_gost512 (&ctx.gost, (const void*) hash0[19], 64);
+   sph_gost512_close(&ctx.gost, (void*) hash0[20]);
   sph_gost512_init(&ctx.gost);
   sph_gost512 (&ctx.gost, (const void*) hash1[19], 64);
   sph_gost512_close(&ctx.gost, (void*) hash1[20]);
@@ -436,6 +444,39 @@ int x25x_8way_hash( void *output, const void *input, int thrid )
   sph_gost512_init(&ctx.gost);
   sph_gost512 (&ctx.gost, (const void*) hash7[19], 64);
   sph_gost512_close(&ctx.gost, (void*) hash7[20]);
+
+#if defined(X25X_8WAY_SHA)
+
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash0[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash0[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash1[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash1[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash2[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash2[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash3[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash3[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash4[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash4[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash5[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash5[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash6[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash6[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash7[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash7[21] );
+
+   intrlv_8x32_512( vhash, hash0[21], hash1[21], hash2[21], hash3[21],
+                           hash4[21], hash5[21], hash6[21], hash7[21] );
+   
+#else
+
   intrlv_8x32_512( vhashA, hash0[20], hash1[20], hash2[20], hash3[20],
                            hash4[20], hash5[20], hash6[20], hash7[20] );

@@ -445,6 +486,8 @@ int x25x_8way_hash( void *output, const void *input, int thrid )
   dintrlv_8x32_512( hash0[21], hash1[21], hash2[21], hash3[21],
                     hash4[21], hash5[21], hash6[21], hash7[21], vhash );

+#endif
+
   panama_8way_init( &ctx.panama );
   panama_8way_update( &ctx.panama, vhash, 64 );
   panama_8way_close( &ctx.panama, vhash );
@@ -574,68 +617,26 @@ int scanhash_x25x_8way( struct work *work, uint32_t max_nonce,
   return 0;
 }

-/*
-int scanhash_x25x_8way( struct work* work, uint32_t max_nonce,
-                   uint64_t *hashes_done, struct thr_info *mythr )
-{
-   uint32_t hash[8*16] __attribute__ ((aligned (128)));
-   uint32_t vdata[24*8] __attribute__ ((aligned (64)));
-   uint32_t lane_hash[8] __attribute__ ((aligned (64)));
-   uint32_t *hash7 = &(hash[7<<3]);
-   uint32_t *pdata = work->data;
-   uint32_t *ptarget = work->target;
-   const uint32_t first_nonce = pdata[19];
-   __m512i  *noncev = (__m512i*)vdata + 9;   // aligned
-   uint32_t n = first_nonce;
-   const uint32_t last_nonce = max_nonce - 4;
-   const int thr_id = mythr->id;
-   const uint32_t Htarg = ptarget[7];
-
-   if (opt_benchmark)
-      ((uint32_t*)ptarget)[7] = 0x08ff;
-
-   InitializeSWIFFTX();
-
-   mm512_bswap32_intrlv80_8x64( vdata, pdata );
-   do
-   {
-      *noncev = mm512_intrlv_blend_32( mm512_bswap_32(
-              _mm512_set_epi32( n+7, 0, n+6, 0, n+5, 0, n+4, 0,
-                                n+3, 0, n+2, 0, n+1, 0, n,   0 ) ), *noncev );
-      x25x_8way_hash( hash, vdata );
-
-      for ( int lane = 0; lane < 8; lane++ ) if ( hash7[lane] <= Htarg )
-      {
-         extr_lane_8x32( lane_hash, hash, lane, 256 );
-         if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
-         {
-              pdata[19] = n + lane;
-              submit_solution( work, lane_hash, mythr );
-         }
-      }
-      n += 8;
-   } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
-
-   *hashes_done = n - first_nonce;
-   return 0;
-}
-*/
-
 #elif defined(X25X_4WAY)

 union _x25x_4way_ctx_overlay
 {
    blake512_4way_context   blake;
    bmw512_4way_context     bmw;
+#if defined(__VAES__)
+    groestl512_2way_context groestl;
+    echo_2way_context       echo;
+#else
    hashState_groestl       groestl;
    hashState_echo          echo;
+#endif
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
-    hashState_luffa         luffa;
-    cubehashParam           cube;
-    sph_shavite512_context  shavite;
-    hashState_sd            simd;
+    luffa_2way_context      luffa;
+    cube_2way_context       cube;
+    shavite512_2way_context shavite;
+    simd_2way_context       simd;
    hamsi512_4way_context   hamsi;
    hashState_fugue         fugue;
    shabal512_4way_context  shabal;
@@ -644,7 +645,11 @@ union _x25x_4way_ctx_overlay
    haval256_5_4way_context haval;
    sph_tiger_context       tiger;
    sph_gost512_context     gost;
+#if defined(X25X_4WAY_SHA)
+    sph_sha256_context      sha256;
+#else
    sha256_4way_context     sha256;
+#endif
    panama_4way_context     panama;
    blake2s_4way_state      blake2s;
 };
@@ -658,6 +663,8 @@ int x25x_4way_hash( void *output, const void *input, int thrid )
   unsigned char hash2[25][64] __attribute__((aligned(64))) = {0};
   unsigned char hash3[25][64] __attribute__((aligned(64))) = {0};
   unsigned char vhashX[24][64*4] __attribute__ ((aligned (64)));
+   uint64_t vhashA[8*4] __attribute__ ((aligned (64)));
+   uint64_t vhashB[8*4] __attribute__ ((aligned (64)));
   x25x_4way_ctx_overlay ctx __attribute__ ((aligned (64)));

   blake512_4way_full( &ctx.blake, vhash, input, 80 );
@@ -668,11 +675,25 @@ int x25x_4way_hash( void *output, const void *input, int thrid )
   bmw512_4way_close( &ctx.bmw, vhash );
   dintrlv_4x64_512( hash0[1], hash1[1], hash2[1], hash3[1], vhash );

+#if defined(__VAES__)
+
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+   groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
+   groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
+
+   dintrlv_2x128_512( hash0[2], hash1[2], vhashA );
+   dintrlv_2x128_512( hash2[2], hash3[2], vhashB );
+
+#else
+
   groestl512_full( &ctx.groestl, (char*)hash0[2], (const char*)hash0[1], 512 );
   groestl512_full( &ctx.groestl, (char*)hash1[2], (const char*)hash1[1], 512 );
   groestl512_full( &ctx.groestl, (char*)hash2[2], (const char*)hash2[1], 512 );
   groestl512_full( &ctx.groestl, (char*)hash3[2], (const char*)hash3[1], 512 );

+#endif
+
   intrlv_4x64_512( vhash, hash0[2], hash1[2], hash2[2], hash3[2] );
   skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
   dintrlv_4x64_512( hash0[3], hash1[3], hash2[3], hash3[3], vhash );
@@ -689,41 +710,38 @@ int x25x_4way_hash( void *output, const void *input, int thrid )
   keccak512_4way_close( &ctx.keccak, vhash );
   dintrlv_4x64_512( hash0[5], hash1[5], hash2[5], hash3[5], vhash );

-   luffa_full( &ctx.luffa, (BitSequence*)hash0[6], 512,
-                     (const BitSequence*)hash0[5], 64 );
-   luffa_full( &ctx.luffa, (BitSequence*)hash1[6], 512,
-                     (const BitSequence*)hash1[5], 64 );
-   luffa_full( &ctx.luffa, (BitSequence*)hash2[6], 512,
-                     (const BitSequence*)hash2[5], 64 );
-   luffa_full( &ctx.luffa, (BitSequence*)hash3[6], 512,
-                     (const BitSequence*)hash3[5], 64 );
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );

-   cubehash_full( &ctx.cube, (byte*)hash0[7], 512, (const byte*)hash0[6], 64 );
-   cubehash_full( &ctx.cube, (byte*)hash1[7], 512, (const byte*)hash1[6], 64 );
-   cubehash_full( &ctx.cube, (byte*)hash2[7], 512, (const byte*)hash2[6], 64 );
-   cubehash_full( &ctx.cube, (byte*)hash3[7], 512, (const byte*)hash3[6], 64 );
+   luffa512_2way_full( &ctx.luffa, vhashA, vhashA, 64 );
+   luffa512_2way_full( &ctx.luffa, vhashB, vhashB, 64 );
+   dintrlv_2x128_512( hash0[6], hash1[6], vhashA );
+   dintrlv_2x128_512( hash2[6], hash3[6], vhashB );
   
-   sph_shavite512_init(&ctx.shavite);
-   sph_shavite512(&ctx.shavite, (const void*) hash0[7], 64);
-   sph_shavite512_close(&ctx.shavite, hash0[8]);
-   sph_shavite512_init(&ctx.shavite);
-   sph_shavite512(&ctx.shavite, (const void*) hash1[7], 64);
-   sph_shavite512_close(&ctx.shavite, hash1[8]);
-   sph_shavite512_init(&ctx.shavite);
-   sph_shavite512(&ctx.shavite, (const void*) hash2[7], 64);
-   sph_shavite512_close(&ctx.shavite, hash2[8]);
-   sph_shavite512_init(&ctx.shavite);
-   sph_shavite512(&ctx.shavite, (const void*) hash3[7], 64);
-   sph_shavite512_close(&ctx.shavite, hash3[8]);
+   cube_2way_full( &ctx.cube, vhashA, 512, vhashA, 64 );
+   cube_2way_full( &ctx.cube, vhashB, 512, vhashB, 64 );
+   dintrlv_2x128_512( hash0[7], hash1[7], vhashA );
+   dintrlv_2x128_512( hash2[7], hash3[7], vhashB );

-   simd_full( &ctx.simd, (BitSequence*)hash0[9],
-                   (const BitSequence*)hash0[8], 512 );
-   simd_full( &ctx.simd, (BitSequence*)hash1[9],
-                   (const BitSequence*)hash1[8], 512 );
-   simd_full( &ctx.simd, (BitSequence*)hash2[9],
-                   (const BitSequence*)hash2[8], 512 );
-   simd_full( &ctx.simd, (BitSequence*)hash3[9],
-                   (const BitSequence*)hash3[8], 512 );
+   shavite512_2way_full( &ctx.shavite, vhashA, vhashA, 64 );
+   shavite512_2way_full( &ctx.shavite, vhashB, vhashB, 64 );
+   dintrlv_2x128_512( hash0[8], hash1[8], vhashA );
+   dintrlv_2x128_512( hash2[8], hash3[8], vhashB );
+
+   simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
+   simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
+   dintrlv_2x128_512( hash0[9], hash1[9], vhashA );
+   dintrlv_2x128_512( hash2[9], hash3[9], vhashB );
+
+#if defined(__VAES__)
+
+   echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
+   echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
+   dintrlv_2x128_512( hash0[10], hash1[10], vhashA );
+   dintrlv_2x128_512( hash2[10], hash3[10], vhashB );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else

   echo_full( &ctx.echo, (BitSequence *)hash0[10], 512,
                   (const BitSequence *)hash0[ 9], 64 );
@@ -736,6 +754,8 @@ int x25x_4way_hash( void *output, const void *input, int thrid )

   intrlv_4x64_512( vhash, hash0[10], hash1[10], hash2[10], hash3[10] );

+#endif
+
   if ( work_restart[thrid].restart ) return 0;
   
   hamsi512_4way_init( &ctx.hamsi );
@@ -826,6 +846,25 @@ int x25x_4way_hash( void *output, const void *input, int thrid )
   sph_gost512 (&ctx.gost, (const void*) hash3[19], 64);
   sph_gost512_close(&ctx.gost, (void*) hash3[20]);

+#if defined(X25X_4WAY_SHA)
+
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash0[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash0[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash1[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash1[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash2[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash2[21] );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, hash3[20], 64 );
+   sph_sha256_close( &ctx.sha256, hash3[21] );
+
+   intrlv_4x32_512( vhash, hash0[21], hash1[21], hash2[21], hash3[21] );
+
+#else   
+
   intrlv_4x32_512( vhashX[0], hash0[20], hash1[20], hash2[20], hash3[20] );
   memset( vhash, 0, 64*4 );

@@ -834,6 +873,8 @@ int x25x_4way_hash( void *output, const void *input, int thrid )
   sha256_4way_close( &ctx.sha256, vhash );
   dintrlv_4x32_512( hash0[21], hash1[21], hash2[21], hash3[21], vhash );

+#endif
+
   panama_4way_init( &ctx.panama );
   panama_4way_update( &ctx.panama, vhash, 64 );
   panama_4way_close( &ctx.panama, vhash );
--- a/algo/x22/x25x.c
+++ b/algo/x22/x25x.c
@@ -23,7 +23,7 @@
 #include "algo/hamsi/sph_hamsi.h"
 #include "algo/shabal/sph_shabal.h"
 #include "algo/whirlpool/sph_whirlpool.h"
-#include <openssl/sha.h>
+#include "algo/sha/sph_sha2.h"
 #include "algo/haval/sph-haval.h"
 #include "algo/tiger/sph_tiger.h"
 #include "algo/lyra2/lyra2.h"
@@ -56,11 +56,11 @@ union _x25x_context_overlay
        sph_hamsi512_context    hamsi;
        sph_shabal512_context   shabal;
        sph_whirlpool_context   whirlpool;
-        SHA512_CTX              sha512;
+        sph_sha512_context      sha512;
        sph_haval256_5_context  haval;
        sph_tiger_context       tiger;
        sph_gost512_context     gost;
-        SHA256_CTX              sha256;
+        sph_sha256_context      sha256;
        sph_panama_context      panama;
        blake2s_state           blake2s;
 };
@@ -71,13 +71,13 @@ int x25x_hash( void *output, const void *input, int thrid )
   unsigned char hash[25][64] __attribute__((aligned(64))) = {0};
   x25x_context_overlay ctx;

-	sph_blake512_init(&ctx.blake);
-	sph_blake512(&ctx.blake, input, 80);
-	sph_blake512_close(&ctx.blake, &hash[0] );
+   sph_blake512_init(&ctx.blake);
+   sph_blake512(&ctx.blake, input, 80);
+   sph_blake512_close(&ctx.blake, &hash[0] );

-	sph_bmw512_init(&ctx.bmw);
-	sph_bmw512(&ctx.bmw, (const void*) &hash[0], 64);
-	sph_bmw512_close(&ctx.bmw, &hash[1]);
+   sph_bmw512_init(&ctx.bmw);
+   sph_bmw512(&ctx.bmw, (const void*) &hash[0], 64);
+   sph_bmw512_close(&ctx.bmw, &hash[1]);

 #if defined(__AES__)
   init_groestl( &ctx.groestl, 64 );
@@ -89,17 +89,17 @@ int x25x_hash( void *output, const void *input, int thrid )
   sph_groestl512_close( &ctx.groestl, &hash[2] );
 #endif
   
-	sph_skein512_init(&ctx.skein);
-	sph_skein512(&ctx.skein, (const void*) &hash[2], 64);
-	sph_skein512_close(&ctx.skein, &hash[3]);
+   sph_skein512_init(&ctx.skein);
+   sph_skein512(&ctx.skein, (const void*) &hash[2], 64);
+   sph_skein512_close(&ctx.skein, &hash[3]);

-	sph_jh512_init(&ctx.jh);
-	sph_jh512(&ctx.jh, (const void*) &hash[3], 64);
-	sph_jh512_close(&ctx.jh, &hash[4]);
+   sph_jh512_init(&ctx.jh);
+   sph_jh512(&ctx.jh, (const void*) &hash[3], 64);
+   sph_jh512_close(&ctx.jh, &hash[4]);

-	sph_keccak512_init(&ctx.keccak);
-	sph_keccak512(&ctx.keccak, (const void*) &hash[4], 64);
-	sph_keccak512_close(&ctx.keccak, &hash[5]);
+   sph_keccak512_init(&ctx.keccak);
+   sph_keccak512(&ctx.keccak, (const void*) &hash[4], 64);
+   sph_keccak512_close(&ctx.keccak, &hash[5]);

   if ( work_restart[thrid].restart ) return 0;
   
@@ -111,9 +111,9 @@ int x25x_hash( void *output, const void *input, int thrid )
   cubehashUpdateDigest( &ctx.cube, (byte*) &hash[7],
                              (const byte*)&hash[6], 64 );

-	sph_shavite512_init(&ctx.shavite);
-	sph_shavite512(&ctx.shavite, (const void*) &hash[7], 64);
-	sph_shavite512_close(&ctx.shavite, &hash[8]);
+   sph_shavite512_init(&ctx.shavite);
+   sph_shavite512(&ctx.shavite, (const void*) &hash[7], 64);
+   sph_shavite512_close(&ctx.shavite, &hash[8]);

   init_sd( &ctx.simd, 512 );
   update_final_sd( &ctx.simd, (BitSequence*)&hash[9],
@@ -132,51 +132,51 @@ int x25x_hash( void *output, const void *input, int thrid )
   if ( work_restart[thrid].restart ) return 0;

   sph_hamsi512_init(&ctx.hamsi);
-	sph_hamsi512(&ctx.hamsi, (const void*) &hash[10], 64);
-	sph_hamsi512_close(&ctx.hamsi, &hash[11]);
+   sph_hamsi512(&ctx.hamsi, (const void*) &hash[10], 64);
+   sph_hamsi512_close(&ctx.hamsi, &hash[11]);

 #if defined(__AES__)
-        fugue512_full( &ctx.fugue, &hash[12], &hash[11], 64 );
+   fugue512_full( &ctx.fugue, &hash[12], &hash[11], 64 );
 #else
-	sph_fugue512_init(&ctx.fugue);
-	sph_fugue512(&ctx.fugue, (const void*) &hash[11], 64);
-	sph_fugue512_close(&ctx.fugue, &hash[12]);
+   sph_fugue512_init(&ctx.fugue);
+   sph_fugue512(&ctx.fugue, (const void*) &hash[11], 64);
+   sph_fugue512_close(&ctx.fugue, &hash[12]);
 #endif

-	sph_shabal512_init(&ctx.shabal);
-	sph_shabal512(&ctx.shabal, (const void*) &hash[12], 64);
-	sph_shabal512_close(&ctx.shabal, &hash[13]);
+   sph_shabal512_init(&ctx.shabal);
+   sph_shabal512(&ctx.shabal, (const void*) &hash[12], 64);
+   sph_shabal512_close(&ctx.shabal, &hash[13]);

-	sph_whirlpool_init(&ctx.whirlpool);
-	sph_whirlpool (&ctx.whirlpool, (const void*) &hash[13], 64);
-	sph_whirlpool_close(&ctx.whirlpool, &hash[14]);
+   sph_whirlpool_init(&ctx.whirlpool);
+   sph_whirlpool (&ctx.whirlpool, (const void*) &hash[13], 64);
+   sph_whirlpool_close(&ctx.whirlpool, &hash[14]);

-   SHA512_Init( &ctx.sha512 );
-   SHA512_Update(  &ctx.sha512, (const void*) &hash[14], 64);
-   SHA512_Final( (void*) &hash[15], &ctx.sha512 );
+   sph_sha512_init( &ctx.sha512 );
+   sph_sha512( &ctx.sha512, &hash[14], 64 );
+   sph_sha512_close( &ctx.sha512, &hash[15] );

   ComputeSingleSWIFFTX((unsigned char*)&hash[12], (unsigned char*)&hash[16]);

-	sph_haval256_5_init(&ctx.haval);
-	sph_haval256_5(&ctx.haval,(const void*) &hash[16], 64);
-	sph_haval256_5_close(&ctx.haval,&hash[17]);
+   sph_haval256_5_init(&ctx.haval);
+   sph_haval256_5(&ctx.haval,(const void*) &hash[16], 64);
+   sph_haval256_5_close(&ctx.haval,&hash[17]);

   if ( work_restart[thrid].restart ) return 0;
   
-	sph_tiger_init(&ctx.tiger);
-	sph_tiger (&ctx.tiger, (const void*) &hash[17], 64);
-	sph_tiger_close(&ctx.tiger, (void*) &hash[18]);
+   sph_tiger_init(&ctx.tiger);
+   sph_tiger (&ctx.tiger, (const void*) &hash[17], 64);
+   sph_tiger_close(&ctx.tiger, (void*) &hash[18]);

-	LYRA2RE( (void*)&hash[19], 32, (const void*)&hash[18], 32,
+   LYRA2RE( (void*)&hash[19], 32, (const void*)&hash[18], 32,
            (const void*)&hash[18], 32, 1, 4, 4 );

-	sph_gost512_init(&ctx.gost);
-	sph_gost512 (&ctx.gost, (const void*) &hash[19], 64);
-	sph_gost512_close(&ctx.gost, (void*) &hash[20]);
+   sph_gost512_init(&ctx.gost);
+   sph_gost512 (&ctx.gost, (const void*) &hash[19], 64);
+   sph_gost512_close(&ctx.gost, (void*) &hash[20]);

-   SHA256_Init( &ctx.sha256 );
-   SHA256_Update(  &ctx.sha256, (const void*) &hash[20], 64 );
-   SHA256_Final( (unsigned char*) &hash[21], &ctx.sha256 );
+   sph_sha256_init( &ctx.sha256 );
+   sph_sha256( &ctx.sha256, &hash[20], 64 );
+   sph_sha256_close( &ctx.sha256, &hash[21] );

   sph_panama_init(&ctx.panama);
   sph_panama (&ctx.panama, (const void*) &hash[21], 64 );
--- a/algo/yescrypt/yescrypt-simd.c
+++ b/algo/yescrypt/yescrypt-simd.c
@@ -1302,10 +1302,7 @@ yescrypt_kdf(const yescrypt_shared_t * shared, yescrypt_local_t * local,
 		S = (uint8_t *)XY + XY_size;

 	if (t || flags) {
-		SHA256_CTX ctx;
-		SHA256_Init(&ctx);
-		SHA256_Update(&ctx, passwd, passwdlen);
-		SHA256_Final(sha256, &ctx);
+		SHA256_Buf( passwd, passwdlen, sha256 );
 		passwd = sha256;
 		passwdlen = sizeof(sha256);
 	}
@@ -1382,10 +1379,7 @@ yescrypt_kdf(const yescrypt_shared_t * shared, yescrypt_local_t * local,
 	   }
 	   /* Compute StoredKey */
 	   {
-		SHA256_CTX ctx;
-		SHA256_Init(&ctx);
-		SHA256_Update(&ctx, sha256, sizeof(sha256));
-		SHA256_Final(buf, &ctx);
+		SHA256_Buf( sha256, sizeof(sha256), buf );
 	   }
 	}

--- a/algo/yespower/yescrypt-r8g.c
+++ b/algo/yespower/yescrypt-r8g.c
@@ -52,8 +52,8 @@ int scanhash_yespower_r8g( struct work *work, uint32_t max_nonce,
    endiandata[19] = n;

 // do sha256 prehash
-   SHA256_Init( &sha256_prehash_ctx );
-   SHA256_Update( &sha256_prehash_ctx, endiandata, 64 );
+   sph_sha256_init( &sha256_prehash_ctx );
+   sph_sha256( &sha256_prehash_ctx, endiandata, 64 );
    
    do {
       yespower_tls( (unsigned char *)endiandata, params.perslen,
--- a/algo/yespower/yespower-blake2b.c
+++ b/algo/yespower/yespower-blake2b.c
@@ -259,15 +259,24 @@ static inline void salsa20_simd_unshuffle(const salsa20_blk_t *Bin,
 #define WRITE_X(out) \
    (out).q[0] = X0; (out).q[1] = X1; (out).q[2] = X2; (out).q[3] = X3;

-#ifdef __XOP__
+#if defined(__AVX512VL__)
+
+#define ARX(out, in1, in2, s) \
+   out = _mm_xor_si128(out, _mm_rol_epi32(_mm_add_epi32(in1, in2), s));
+
+#elif defined(__XOP__)
+
 #define ARX(out, in1, in2, s) \
    out = _mm_xor_si128(out, _mm_roti_epi32(_mm_add_epi32(in1, in2), s));
+
 #else
+
 #define ARX(out, in1, in2, s) { \
    __m128i tmp = _mm_add_epi32(in1, in2); \
    out = _mm_xor_si128(out, _mm_slli_epi32(tmp, s)); \
    out = _mm_xor_si128(out, _mm_srli_epi32(tmp, 32 - s)); \
 }
+
 #endif

 #define SALSA20_2ROUNDS \
--- a/algo/yespower/yespower-gate.c
+++ b/algo/yespower/yespower-gate.c
@@ -33,7 +33,8 @@
 yespower_params_t yespower_params;

 //SHA256_CTX sha256_prehash_ctx;
-__thread SHA256_CTX sha256_prehash_ctx;
+__thread sph_sha256_context sha256_prehash_ctx;
+//__thread SHA256_CTX sha256_prehash_ctx;

 // YESPOWER

@@ -59,9 +60,9 @@ int scanhash_yespower( struct work *work, uint32_t max_nonce,
      be32enc( &endiandata[k], pdata[k] );
   endiandata[19] = n;

-// do sha256 prehash
-   SHA256_Init( &sha256_prehash_ctx );
-   SHA256_Update( &sha256_prehash_ctx, endiandata, 64 );
+   // do sha256 prehash
+   sph_sha256_init( &sha256_prehash_ctx );
+   sph_sha256( &sha256_prehash_ctx, endiandata, 64 );

   do {
      if ( yespower_hash( (char*)endiandata, (char*)vhash, 80, thr_id ) )
@@ -100,9 +101,9 @@ int scanhash_yespower_b2b( struct work *work, uint32_t max_nonce,
      be32enc( &endiandata[k], pdata[k] );
   endiandata[19] = n;

-// do sha256 prehash
-   SHA256_Init( &sha256_prehash_ctx );
-   SHA256_Update( &sha256_prehash_ctx, endiandata, 64 );
+   // do sha256 prehash
+   sph_sha256_init( &sha256_prehash_ctx );
+   sph_sha256( &sha256_prehash_ctx, endiandata, 64 );

   do {
      if (yespower_b2b_hash( (char*) endiandata, (char*) vhash, 80, thr_id ) )
@@ -165,25 +166,14 @@ bool register_yespowerr16_algo( algo_gate_t* gate )
  return true;
 };

-/* not used, doesn't work
-bool register_yescrypt_05_algo( algo_gate_t* gate )
-{
-   gate->optimizations = SSE2_OPT | SHA_OPT;
-   gate->scanhash   = (void*)&scanhash_yespower;
-   yespower_params.version = YESPOWER_0_5;
-   yespower_params.N       = 2048;
-   yespower_params.r       = 8;
-   yespower_params.pers    = NULL;
-   yespower_params.perslen = 0;
-   opt_target_factor = 65536.0;
-   return true;
-}
+// Legacy Yescrypt (yespower v0.5)

 bool register_yescrypt_05_algo( algo_gate_t* gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yespower;
   yespower_params.version = YESPOWER_0_5;
+   opt_target_factor = 65536.0;

   if ( opt_param_n )  yespower_params.N = opt_param_n;
   else                yespower_params.N = 2048;
@@ -202,8 +192,6 @@ bool register_yescrypt_05_algo( algo_gate_t* gate )
     yespower_params.perslen = 0;
   }

-//   YESCRYPT_P = 1;
-
   applog( LOG_NOTICE,"Yescrypt parameters: N= %d, R= %d.",
                                      yespower_params.N, yespower_params.r );
   if ( yespower_params.pers )
@@ -233,8 +221,8 @@ bool register_yescryptr16_05_algo( algo_gate_t* gate )
   yespower_params.version = YESPOWER_0_5;
   yespower_params.N       = 4096;
   yespower_params.r       = 16;
-   yespower_params.pers    = NULL;
-   yespower_params.perslen = 0;
+   yespower_params.pers    = "Client Key";
+   yespower_params.perslen = 10;
   opt_target_factor = 65536.0;
   return true;
 }
@@ -251,7 +239,6 @@ bool register_yescryptr32_05_algo( algo_gate_t* gate )
   opt_target_factor = 65536.0;
   return true;
 }
-*/

 // POWER2B

--- a/algo/yespower/yespower-opt.c
+++ b/algo/yespower/yespower-opt.c
@@ -165,6 +165,7 @@ static inline void salsa20_simd_unshuffle(const salsa20_blk_t *Bin,
 }

 #ifdef __SSE2__
+
 #define DECL_X \
 	__m128i X0, X1, X2, X3;
 #define DECL_Y \
@@ -174,15 +175,24 @@ static inline void salsa20_simd_unshuffle(const salsa20_blk_t *Bin,
 #define WRITE_X(out) \
 	(out).q[0] = X0; (out).q[1] = X1; (out).q[2] = X2; (out).q[3] = X3;

-#ifdef __XOP__
+#if defined(__AVX512VL__)
+
+#define ARX(out, in1, in2, s) \
+   out = _mm_xor_si128(out, _mm_rol_epi32(_mm_add_epi32(in1, in2), s));
+
+#elif defined(__XOP__)
+
 #define ARX(out, in1, in2, s) \
 	out = _mm_xor_si128(out, _mm_roti_epi32(_mm_add_epi32(in1, in2), s));
+
 #else
+
 #define ARX(out, in1, in2, s) { \
 	__m128i tmp = _mm_add_epi32(in1, in2); \
 	out = _mm_xor_si128(out, _mm_slli_epi32(tmp, s)); \
 	out = _mm_xor_si128(out, _mm_srli_epi32(tmp, 32 - s)); \
 }
+
 #endif

 #define SALSA20_2ROUNDS \
@@ -1029,72 +1039,72 @@ int yespower(yespower_local_t *local,
    const yespower_params_t *params,
    yespower_binary_t *dst, int thrid )
 {
-	yespower_version_t version = params->version;
-	uint32_t N = params->N;
-	uint32_t r = params->r;
-	const uint8_t *pers = params->pers;
-	size_t perslen = params->perslen;
-	uint32_t Swidth;
-	size_t B_size, V_size, XY_size, need;
-	uint8_t *B, *S;
-	salsa20_blk_t *V, *XY;
-	pwxform_ctx_t ctx;
-	uint8_t sha256[32];
-   SHA256_CTX sha256_ctx;
+   yespower_version_t version = params->version;
+   uint32_t N = params->N;
+   uint32_t r = params->r;
+   const uint8_t *pers = params->pers;
+   size_t perslen = params->perslen;
+   uint32_t Swidth;
+   size_t B_size, V_size, XY_size, need;
+   uint8_t *B, *S;
+   salsa20_blk_t *V, *XY;
+   pwxform_ctx_t ctx;
+   uint8_t sha256[32];
+   sph_sha256_context sha256_ctx;

-	/* Sanity-check parameters */
-	if ( (version != YESPOWER_0_5 && version != YESPOWER_1_0)
+   /* Sanity-check parameters */
+   if ( (version != YESPOWER_0_5 && version != YESPOWER_1_0)
      || N < 1024 || N > 512 * 1024 || r < 8 || r > 32
      || (N & (N - 1)) != 0 || ( !pers && perslen ) )
   {
-		errno = EINVAL;
-		return -1;
-	}
+      errno = EINVAL;
+      return -1;
+   }

-	/* Allocate memory */
-	B_size = (size_t)128 * r;
-	V_size = B_size * N;
-	if ( version == YESPOWER_0_5 )
+   /* Allocate memory */
+   B_size = (size_t)128 * r;
+   V_size = B_size * N;
+   if ( version == YESPOWER_0_5 )
   {
-		XY_size = B_size * 2;
-		Swidth = Swidth_0_5;
-		ctx.Sbytes = 2 * Swidth_to_Sbytes1( Swidth );
-	} else {
-		XY_size = B_size + 64;
-		Swidth = Swidth_1_0;
-		ctx.Sbytes = 3 * Swidth_to_Sbytes1( Swidth );
-	}
-	need = B_size + V_size + XY_size + ctx.Sbytes;
-	if ( local->aligned_size < need )
+      XY_size = B_size * 2;
+      Swidth = Swidth_0_5;
+      ctx.Sbytes = 2 * Swidth_to_Sbytes1( Swidth );
+   }
+   else
   {
-		if ( free_region( local ) )
-			return -1;
-		if ( !alloc_region( local, need ) )
-			return -1;
-	}
-	B = (uint8_t *)local->aligned;
-	V = (salsa20_blk_t *)((uint8_t *)B + B_size);
-	XY = (salsa20_blk_t *)((uint8_t *)V + V_size);
-	S = (uint8_t *)XY + XY_size;
-	ctx.S0 = S;
-	ctx.S1 = S + Swidth_to_Sbytes1( Swidth );
+      XY_size = B_size + 64;
+      Swidth = Swidth_1_0;
+      ctx.Sbytes = 3 * Swidth_to_Sbytes1( Swidth );
+   }
+   need = B_size + V_size + XY_size + ctx.Sbytes;
+   if ( local->aligned_size < need )
+   {
+      if ( free_region( local ) )
+         return -1;
+      if ( !alloc_region( local, need ) )
+         return -1;
+   }
+   B = (uint8_t *)local->aligned;
+   V = (salsa20_blk_t *)((uint8_t *)B + B_size);
+   XY = (salsa20_blk_t *)((uint8_t *)V + V_size);
+   S = (uint8_t *)XY + XY_size;
+   ctx.S0 = S;
+   ctx.S1 = S + Swidth_to_Sbytes1( Swidth );

-
-// copy prehash, do tail   
+   // copy prehash, do tail   
   memcpy( &sha256_ctx, &sha256_prehash_ctx, sizeof sha256_ctx );
-   SHA256_Update( &sha256_ctx, src+64, srclen-64 );
-   SHA256_Final( sha256, &sha256_ctx );

-//   SHA256_Buf(src, srclen, sha256);
+   sph_sha256( &sha256_ctx, src+64, srclen-64 );
+   sph_sha256_close( &sha256_ctx, sha256 );

-	if ( version == YESPOWER_0_5 )
+   if ( version == YESPOWER_0_5 )
   {
      PBKDF2_SHA256( sha256, sizeof(sha256), src, srclen, 1, B, B_size );

      if ( work_restart[thrid].restart ) return 0;
   
      memcpy( sha256, B, sizeof(sha256) );
-		smix( B, r, N, V, XY, &ctx );
+      smix( B, r, N, V, XY, &ctx );

      if ( work_restart[thrid].restart ) return 0;

@@ -1108,54 +1118,36 @@ int yespower(yespower_local_t *local,
         src = pers;
         srclen = perslen;
      }
-      else
-         srclen = 0;
+
+      HMAC_SHA256_Buf( dst, sizeof(*dst), src, srclen, sha256 );
+      SHA256_Buf( sha256, sizeof(sha256), (uint8_t *)dst );
      
-      HMAC_SHA256_CTX ctx;
-      HMAC_SHA256_Init( &ctx, dst, sizeof(*dst) );
-      HMAC_SHA256_Update( &ctx, src, srclen );
-      HMAC_SHA256_Final( sha256, &ctx );
-
-//      SHA256_CTX ctx;
-      SHA256_Init( &sha256_ctx );
-      SHA256_Update( &sha256_ctx, sha256, sizeof(sha256) );
-      SHA256_Final( (unsigned char*)dst, &sha256_ctx );
-
-
-/*
-      if ( pers )
-      {
-			HMAC_SHA256_Buf( dst, sizeof(*dst), pers, perslen, sha256 );
-         SHA256_Buf( sha256, sizeof(sha256), (uint8_t *)dst );
-		}
-*/
   }
   else
   {
-		ctx.S2 = S + 2 * Swidth_to_Sbytes1( Swidth );
-		ctx.w = 0;
-
-		if ( pers )
+      ctx.S2 = S + 2 * Swidth_to_Sbytes1( Swidth );
+      ctx.w = 0;
+      if ( pers )
      {
-			src = pers;
-			srclen = perslen;
-		}
+         src = pers;
+         srclen = perslen;
+      }
      else
-			srclen = 0;
+         srclen = 0;

-		PBKDF2_SHA256( sha256, sizeof(sha256), src, srclen, 1, B, 128 );
-		memcpy( sha256, B, sizeof(sha256) );
+      PBKDF2_SHA256( sha256, sizeof(sha256), src, srclen, 1, B, 128 );
+      memcpy( sha256, B, sizeof(sha256) );

      if ( work_restart[thrid].restart ) return 0;

      smix_1_0( B, r, N, V, XY, &ctx );
-
+      
      HMAC_SHA256_Buf( B + B_size - 64, 64, sha256, sizeof(sha256),
                       (uint8_t *)dst );
-	}
+   }

-	/* Success! */
-	return 1;
+   /* Success! */
+   return 1;
 }

 /**
--- a/algo/yespower/yespower.h
+++ b/algo/yespower/yespower.h
@@ -34,6 +34,7 @@
 #include <stdlib.h> /* for size_t */
 #include "miner.h"
 #include "simd-utils.h"
+#include "algo/sha/sph_sha2.h"
 #include <openssl/sha.h>

 #ifdef __cplusplus
@@ -79,7 +80,8 @@ typedef struct {
 extern yespower_params_t yespower_params;

 //SHA256_CTX sha256_prehash_ctx;
-extern __thread SHA256_CTX sha256_prehash_ctx;
+extern __thread sph_sha256_context sha256_prehash_ctx;
+//extern __thread SHA256_CTX sha256_prehash_ctx;

 /**
 * yespower_init_local(local):
--- a/build-allarch.sh
+++ b/build-allarch.sh
@@ -4,8 +4,9 @@
 # during develpment. However the information contained may provide compilation
 # tips to users.

-rm cpuminer-avx512-sha-vaes cpuminer-avx512 cpuminer-avx2 cpuminer-aes-avx cpuminer-aes-sse42 cpuminer-sse42 cpuminer-ssse3 cpuminer-sse2 cpuminer-zen  > /dev/null
+rm cpuminer-avx512-sha-vaes cpuminer-avx512-sha cpuminer-avx512 cpuminer-avx2 cpuminer-aes-avx cpuminer-aes-sse42 cpuminer-sse42 cpuminer-ssse3 cpuminer-sse2 cpuminer-zen cpuminer-zen3  > /dev/null

+# Icelake AVX512 SHA VAES
 make distclean || echo clean
 rm -f config.status
 ./autogen.sh || echo done
@@ -16,6 +17,21 @@ mv cpuminer.exe cpuminer-avx512-sha-vaes.exe
 strip -s cpuminer
 mv cpuminer cpuminer-avx512-sha-vaes

+# Rocketlake AVX512 SHA AES
+make clean || echo clean
+rm -f config.status
+CFLAGS="-O3 -march=cascadelake -msha -Wall -fno-common" ./configure --with-curl
+#CFLAGS="-O3 -march=skylake-avx512 -msha -Wall -fno-common" ./configure --with-curl
+# CFLAGS="-O3 -march=rocketlake -Wall -fno-common" ./configure --with-curl
+make -j 8
+strip -s cpuminer.exe
+mv cpuminer.exe cpuminer-avx512-sha.exe
+strip -s cpuminer
+mv cpuminer cpuminer-avx512-sha
+
+# Slylake-X AVX512 AES
+make clean || echo clean
+rm -f config.status
 CFLAGS="-O3 -march=skylake-avx512 -Wall -fno-common" ./configure --with-curl
 make -j 8
 strip -s cpuminer.exe
@@ -23,6 +39,7 @@ mv cpuminer.exe cpuminer-avx512.exe
 strip -s cpuminer
 mv cpuminer cpuminer-avx512

+# Haswell AVX2 AES
 make clean || echo clean
 rm -f config.status
 # GCC 9 doesn't include AES with core-avx2
@@ -33,6 +50,7 @@ mv cpuminer.exe cpuminer-avx2.exe
 strip -s cpuminer
 mv cpuminer cpuminer-avx2

+# Sandybridge AVX AES
 make clean || echo clean
 rm -f config.status
 CFLAGS="-O3 -march=corei7-avx -maes -Wall -fno-common" ./configure --with-curl
@@ -42,15 +60,17 @@ mv cpuminer.exe cpuminer-avx.exe
 strip -s cpuminer
 mv cpuminer cpuminer-avx

+# Westmere SSE4.2 AES
 make clean || echo clean
 rm -f config.status
-CFLAGS="-O3 -maes -msse4.2 -Wall -fno-common" ./configure --with-curl
+CFLAGS="-O3 -march=westmere -Wall -fno-common" ./configure --with-curl
 make -j 8
 strip -s cpuminer.exe
 mv cpuminer.exe cpuminer-aes-sse42.exe
 strip -s cpuminer
 mv cpuminer cpuminer-aes-sse42

+# Nehalem SSE4.2
 make clean || echo clean
 rm -f config.status
 CFLAGS="-O3 -march=corei7 -Wall -fno-common" ./configure --with-curl
@@ -60,6 +80,7 @@ mv cpuminer.exe cpuminer-sse42.exe
 strip -s cpuminer
 mv cpuminer cpuminer-sse42

+# Core2 SSSE3
 make clean || echo clean
 rm -f config.status
 CFLAGS="-O3 -march=core2 -Wall -fno-common" ./configure --with-curl
@@ -69,6 +90,7 @@ mv cpuminer.exe cpuminer-ssse3.exe
 strip -s cpuminer
 mv cpuminer cpuminer-ssse3

+# Generic SSE2
 make clean || echo clean
 rm -f config.status
 CFLAGS="-O3 -msse2 -Wall -fno-common" ./configure --with-curl
@@ -78,6 +100,7 @@ mv cpuminer.exe cpuminer-sse2.exe
 strip -s cpuminer
 mv cpuminer cpuminer-sse2

+# AMD Zen1 AVX2 SHA
 make clean || echo done
 rm -f config.status
 CFLAGS="-O3 -march=znver1 -Wall -fno-common" ./configure --with-curl
@@ -87,6 +110,18 @@ mv cpuminer.exe cpuminer-zen.exe
 strip -s cpuminer
 mv cpuminer cpuminer-zen

+# AMD Zen3 AVX2 SHA VAES
+make clean || echo done
+rm -f config.status
+CFLAGS="-O3 -march=znver2 -mvaes -Wall -fno-common" ./configure --with-curl
+# CFLAGS="-O3 -march=znver3 -Wall -fno-common" ./configure --with-curl
+make -j 8
+strip -s cpuminer.exe
+mv cpuminer.exe cpuminer-zen3.exe
+strip -s cpuminer
+mv cpuminer cpuminer-zen3
+
+# Native to current CPU
 make clean || echo done
 rm -f config.status
 CFLAGS="-O3 -march=native -Wall -fno-common" ./configure --with-curl
--- a/build-no-common.sh
+++ b/build-no-common.sh
@@ -1,27 +0,0 @@
-#!/bin/bash
-
-#if [ "$OS" = "Windows_NT" ]; then
-#    ./mingw64.sh
-#    exit 0
-#fi
-
-# Linux build
-
-make distclean || echo clean
-
-rm -f config.status
-./autogen.sh || echo done
-
-# Ubuntu 10.04 (gcc 4.4)
-# extracflags="-O3 -march=native -Wall -D_REENTRANT -funroll-loops -fvariable-expansion-in-unroller -fmerge-all-constants -fbranch-target-load-optimize2 -fsched2-use-superblocks -falign-loops=16 -falign-functions=16 -falign-jumps=16 -falign-labels=16"
-
-# Debian 7.7 / Ubuntu 14.04 (gcc 4.7+)
-#extracflags="$extracflags -Ofast -flto -fuse-linker-plugin -ftree-loop-if-convert-stores"
-
-#CFLAGS="-O3 -march=native -Wall" ./configure --with-curl --with-crypto=$HOME/usr
-CFLAGS="-O3 -march=native -Wall -fno-common" ./configure --with-curl
-#CFLAGS="-O3 -march=native -Wall" CXXFLAGS="$CFLAGS -std=gnu++11" ./configure --with-curl
-
-make -j 4
-
-strip -s cpuminer
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Jay D Dee	3c5e8921b7	v3.16.3	2021-05-06 14:55:03 -04:00
Jay D Dee	f3333b0070	v3.16.2	2021-04-08 18:09:31 -04:00
Jay D Dee	902ec046dd	v3.16.1	2021-03-24 18:24:20 -04:00
Jay D Dee	d0b4941321	v3.16.0	2021-03-19 15:45:32 -04:00
Jay D Dee	40089428c5	v3.15.7	2021-03-08 22:44:44 -05:00
Jay D Dee	dc6b007a18	v3.15.6	2021-02-12 15:16:53 -05:00
Jay D Dee	06bfaa1249	v3.15.5	2020-12-21 13:25:33 -05:00
Jay D Dee	6566e99a13	v3.15.4	2020-12-15 13:15:02 -05:00
Jay D Dee	ccfccbadd5	v3.15.3	2020-12-10 18:23:49 -05:00
Jay D Dee	45ecd0de14	v3.15.2	2020-11-15 17:57:06 -05:00
Jay D Dee	4fa8fcea8b	v3.15.1	2020-11-09 13:19:05 -05:00