v3.23.2

v3.23.1
v3.23.0
2025-09-17 23:44:27 +00:00 · 2023-09-21 12:34:06 -04:00 · 2023-09-13 11:48:52 -04:00 · 2023-08-30 20:15:48 -04:00 · 2023-06-14 11:07:40 -04:00 · 2023-04-06 13:38:37 -04:00
217 changed files with 22536 additions and 23029 deletions
--- a/4
+++ b/4
@@ -1,4 +1,6 @@

+These instructions may be out of date, see the Wiki for the latest...
+https://github.com/JayDDee/cpuminer-opt/wiki/Compiling-from-source

 1. Requirements:
 ---------------
@@ -35,7 +37,7 @@ SHA support on AMD Ryzen CPUs requires gcc version 5 or higher and
 openssl 1.1.0e or higher.

 znver1 and znver2 should be recognized on most recent version of GCC and
-znver3 is expected with GCC 11. GCC 11 also includes rocketlake support.
+znver3 is available with GCC 11. GCC 11 also includes rocketlake support.
 In the meantime here are some suggestions to compile with new CPUs:

 "-march=native" is usually the best choice, used by build.sh.
--- a/156
+++ b/156
@@ -1,158 +1,4 @@
-Instructions for compiling cpuminer-opt for Windows.
-
-Thwaw intructions nay be out of date. Please consult the wiki for
-the latest:
+Please consult the wiki for Windows compile instructions.

 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
-cpuminer-opt executable binaries for Windows.
-
-These instructions were written for Debian and Ubuntu compatible distributions
-but should work on other major distributions as well. However some of the
-package names or file paths may be different.
-
-It is assumed a Linux system is already available and running. And the user
-has enough Linux knowledge to find and install packages and follow these
-instructions.
-
-First it is a good idea to create new user specifically for cross compiling.
-It keeps all mingw stuff contained and isolated from the rest of the system.
-
-Step by step...
-
-1. Install necessary packages from the distribution's repositories.
-
-Refer to Linux compile instructions and install required packages.
-
-Additionally, 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 $HOME/usr/lib/ and uncompress them. 
-
-openssl: https://github.com/openssl/openssl/releases
-
-curl: https://github.com/curl/curl/releases
-
-gmp: https://gmplib.org/download/gmp/
-
-In most cases the latest version is ok but it's safest to download the same major and minor version as included in your distribution. The following uses versions from Ubuntu 20.04. Change version numbers as required.
-
-Run the following commands or follow the supplied instructions. Do not run "make install" unless you are using /usr/lib, which isn't recommended.
-
-Some instructions insist on running "make check". If make check fails it may still work, YMMV.
-
-You can speed up "make" by using all CPU cores available with "-j n" where n is the number of CPU threads you want to use.
-
-openssl:
-
-$ ./Configure mingw64 shared --cross-compile-prefix=x86_64-w64-mingw32-
-$ make
-
-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
-
-gmp:
-
-$ ./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.
-
-Define some local variables to point to local library.
-
-$ export LOCAL_LIB="$HOME/usr/lib"
-
-$ 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"
-
-Adjust for gcc version:
-
-$ 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 $LOCAL_LIB/curl/lib/.libs/libcurl-4.dll release/
-
-The following steps need to be done every time a new source package is
-opened.
-
-5. Download cpuminer-opt
-
-Download the latest source code package of cpumuner-opt to your desired
-location. .zip or .tar.gz, your choice.
-
-https://github.com/JayDDee/cpuminer-opt/releases
-
-Decompress and change to the cpuminer-opt directory.
-
-6. compile
-
-Create a link to the locally compiled version of gmp.h
-
-$ 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.
-
-
--- a/Makefile.am
+++ b/Makefile.am
@@ -55,9 +55,6 @@ cpuminer_SOURCES = \
  algo/blake/mod_blakecoin.c \
  algo/blake/blakecoin.c \
  algo/blake/blakecoin-4way.c \
-  algo/blake/decred-gate.c \
-  algo/blake/decred.c \
-  algo/blake/decred-4way.c \
  algo/blake/pentablake-gate.c \
  algo/blake/pentablake-4way.c \
  algo/blake/pentablake.c \
@@ -166,8 +163,6 @@ cpuminer_SOURCES = \
  algo/sha/sph_sha2big.c \
  algo/sha/sha256-hash-4way.c \
  algo/sha/sha512-hash-4way.c \
-  algo/sha/sha256-hash-opt.c \
-  algo/sha/sha256-hash-2way-ni.c \
  algo/sha/hmac-sha256-hash.c \
  algo/sha/hmac-sha256-hash-4way.c \
  algo/sha/sha256d.c \
@@ -175,9 +170,10 @@ cpuminer_SOURCES = \
  algo/sha/sha256d-4way.c \
  algo/sha/sha256t-gate.c \
  algo/sha/sha256t-4way.c \
-  algo/sha/sha256t.c \
  algo/sha/sha256q-4way.c \
  algo/sha/sha256q.c \
+  algo/sha/sha512256d-4way.c \
+  algo/sha/sha256dt.c \
  algo/shabal/sph_shabal.c \
  algo/shabal/shabal-hash-4way.c \
  algo/shavite/sph_shavite.c \
@@ -205,7 +201,6 @@ cpuminer_SOURCES = \
  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 \
  algo/whirlpool/whirlpool.c \
  algo/whirlpool/whirlpoolx.c \
@@ -285,8 +280,6 @@ cpuminer_SOURCES = \
  algo/x22/x22i-gate.c \
  algo/x22/x25x.c \
  algo/x22/x25x-4way.c \
-  algo/yescrypt/yescrypt.c \
-  algo/yescrypt/yescrypt-best.c \
  algo/yespower/yespower-gate.c \
  algo/yespower/yespower-blake2b.c \
  algo/yespower/crypto/hmac-blake2b.c \
@@ -298,10 +291,10 @@ disable_flags =
 if USE_ASM
   cpuminer_SOURCES += asm/neoscrypt_asm.S
 if ARCH_x86
-   cpuminer_SOURCES += asm/sha2-x86.S asm/scrypt-x86.S asm/aesb-x86.S
+   cpuminer_SOURCES += asm/sha2-x86.S asm/scrypt-x86.S
 endif
 if ARCH_x86_64
-   cpuminer_SOURCES += asm/sha2-x64.S asm/scrypt-x64.S asm/aesb-x64.S
+   cpuminer_SOURCES += asm/sha2-x64.S asm/scrypt-x64.S
 endif
 if ARCH_ARM
   cpuminer_SOURCES += asm/sha2-arm.S asm/scrypt-arm.S
--- a/README.md
+++ b/README.md
@@ -40,17 +40,25 @@ Requirements
 Intel Core2 and newer and AMD equivalents. Further optimizations are available
 on some algoritms for CPUs with AES, AVX, AVX2, SHA, AVX512 and VAES.

-Older CPUs are supported by cpuminer-multi by TPruvot but at reduced
-performance.
+32 bit CPUs are not supported.
+Other CPU architectures such as ARM, Raspberry Pi, RISC-V, Xeon Phi, etc,
+are not supported.

-ARM and Aarch64 CPUs are not supported.
+Mobile CPUs like laptop computers are not recommended because they aren't
+designed for extreme heat of operating at full load for extended periods of
+time.
+
+Older CPUs and ARM architecture may be supported by cpuminer-multi by TPruvot.

 2. 64 bit Linux or Windows OS. Ubuntu and Fedora based distributions,
 including Mint and Centos, are known to work and have all dependencies
 in their repositories. Others may work but may require more effort. Older
 versions such as Centos 6 don't work due to missing features. 
-64 bit Windows OS is supported with mingw_w64 and msys or pre-built binaries.

+Windows 7 or newer is supported with mingw_w64 and msys or using the pre-built
+binaries. WindowsXP 64 bit is YMMV.
+
+FreeBSD is not actively tested but should work, YMMV.
 MacOS, OSx and Android are not supported.

 3. Stratum pool supporting stratum+tcp:// or stratum+ssl:// protocols or
@@ -66,53 +74,50 @@ Supported Algorithms
                          argon2d250    argon2d-crds, Credits (CRDS)
                          argon2d500    argon2d-dyn,  Dynamic (DYN)
                          argon2d4096   argon2d-uis, Unitus, (UIS)
-                          axiom         Shabal-256 MemoHash
-                          blake         Blake-256 (SFR)
-                          blake2b       Blake2b 256
-                          blake2s       Blake-2 S
+                          blake         Blake-256
+                          blake2b       Blake2-512
+                          blake2s       Blake2-256
                          blakecoin     blake256r8
                          bmw           BMW 256
                          bmw512        BMW 512
-                          c11           Chaincoin
+                          c11           
                          decred
                          deep          Deepcoin (DCN)
                          dmd-gr        Diamond-Groestl
                          groestl       Groestl coin
                          hex           x16r-hex
-                          hmq1725       Espers
+                          hmq1725       
                          hodl          Hodlcoin
                          jha           Jackpotcoin
                          keccak        Maxcoin
                          keccakc       Creative coin
                          lbry          LBC, LBRY Credits
-                          luffa         Luffa
-                          lyra2h        Hppcoin
+                          lyra2h        
                          lyra2re       lyra2
                          lyra2rev2     lyra2v2
                          lyra2rev3     lyrav2v3
                          lyra2z        
-                          lyra2z330     Lyra2 330 rows, Zoin (ZOI)
-                          m7m           Magi (XMG)
-                          minotaur      Ringcoin (RNG)
+                          lyra2z330     
+                          m7m           
+                          minotaur 
+                          minotaurx
                          myr-gr        Myriad-Groestl
                          neoscrypt     NeoScrypt(128, 2, 1)
                          nist5         Nist5
                          pentablake    Pentablake
                          phi1612       phi
-                          phi2          Luxcoin (LUX)
-                          phi2-lux      identical to phi2
-                          pluck         Pluck:128 (Supcoin)
+                          phi2          
                          polytimos     Ninja
                          power2b       MicroBitcoin (MBC)
                          quark         Quark
                          qubit         Qubit
                          scrypt        scrypt(1024, 1, 1) (default)
                          scrypt:N      scrypt(N, 1, 1)
+                          scryptn2      scrypt(1048576, 1, 1)
                          sha256d       Double SHA-256
-                          sha256q       Quad SHA-256, Pyrite (PYE)
-                          sha256t       Triple SHA-256, Onecoin (OC)
+                          sha256q       Quad SHA-256
+                          sha256t       Triple SHA-256
                          sha3d         Double keccak256 (BSHA3)
-                          shavite3      Shavite3
                          skein         Skein+Sha (Skeincoin)
                          skein2        Double Skein (Woodcoin)
                          skunk         Signatum (SIGT)
@@ -128,17 +133,17 @@ Supported Algorithms
                          x11           Dash
                          x11evo        Revolvercoin
                          x11gost       sib (SibCoin)
-                          x12           Galaxie Cash (GCH)
-                          x13           X13
+                          x12           
+                          x13           
                          x13bcd        bcd
                          x13sm3        hsr (Hshare)
-                          x14           X14
-                          x15           X15
+                          x14           
+                          x15           
                          x16r          
                          x16rv2        
-                          x16rt         Gincoin (GIN)
-                          x16rt-veil    Veil (VEIL)
-                          x16s          Pigeoncoin (PGN)
+                          x16rt         
+                          x16rt-veil    veil
+                          x16s          
                          x17
                          x21s
                          x22i
--- a/README.txt
+++ b/README.txt
@@ -1,12 +1,22 @@
 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:
+cpuminer-opt is open source and free of any fees. Many forks exist that are
+closed source and contain usage fees. support open source free software.
+
+This package is officially avalaible 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.
+command prompt. There is no GUI and no mouse support.
+
+New users are encouraged to consult the cpuminer-opt Wiki for detailed
+information on usage:
+
+https://github.com/JayDDee/cpuminer-opt/wiki

 Miner programs are often flagged as malware by antivirus programs. This is
 a false positive, they are flagged simply because they are cryptocurrency 
@@ -43,12 +53,11 @@ cpuminer-avx2.exe              Haswell, Skylake, Kabylake, Coffeelake, Cometlake
 cpuminer-avx2-sha.exe          AMD Zen1, Zen2
 cpuminer-avx2-sha-vaes.exe     Intel Alderlake*, AMD Zen3
 cpuminer-avx512.exe            Intel HEDT Skylake-X, Cascadelake
-cpuminer-avx512-sha-vaes.exe   Icelake, Tigerlake, Rocketlake
+cpuminer-avx512-sha-vaes.exe   AMD Zen4, Intel Rocketlake, Icelake

-* Alderlake is a hybrid architecture. With the E-cores disabled it may be
-  possible to enable AVX512 on the the P-cores and use the avx512-sha-vaes
-  build. This is not officially supported by Intel at time of writing.
-  Check for current information.
+* Alderlake is a hybrid architecture with a mix of E-cores & P-cores. Although
+  the P-cores can support AVX512 the E-cores can't so Intel decided to disable
+  AVX512 on the the P-cores.

 Notes about included DLL files:

@@ -59,9 +68,10 @@ 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

-Some DLL filess may already be installed on the system by Windows or third
-party packages. They often will work and may be used instead of the included
-file. 
+Some included DLL files may already be installed on the system by Windows or
+third party packages. They often will work and may be used instead of the
+included version of the files.
+

 If you like this software feel free to donate:

--- a/151
+++ b/151
@@ -65,12 +65,131 @@ If not what makes it happen or not happen?
 Change Log
 ----------

+v3.23.2
+
+sha256dt, sha256t & sha256d +10% with SHA, small improvement with AVX2.
+Other small improvements and code cleanup.
+
+v3.23.1
+
+#349: Fix sha256t low difficulty shares and low effective hash rate.
+Faster sha256dt:  AVX512 +7%, SHA +200%, AVX2 +5%.
+Faster blakecoin & vanilla: AVX2 +30%, AVX512 +110%.
+Other small improvements and code cleanup.
+
+v3.23.0
+
+#398: Prevent GBT fallback to Getwork on network error.
+#398: Prevent excessive logs when conditional mining is paused when mining solo.
+Fix a false start if stratum doesn't immediately send a new job after connecting.
+Tweak diagonal shuffle in Blake2b & Blake256 1-way SIMD to reduce latency.
+CPUID support for AVX10.
+Initial changes to AVX2 targeted code in preparation for AVX10.
+Code cleanup and miscellaneous small improvements.
+
+v3.22.3
+
+Data interleaving and byte swap optimizations with AVX2, AVX512 & AVX512VBMI.
+Faster Luffa with AVX2 & AVX512.
+Other small optimizations.
+Some code cleanup.
+
+v3.22.2
+
+Added sha512256d & sha256dt algos.
+Fixed intermittant invalid shares lyra2v2 AVX512.
+Removed application limits on the number of CPUs and threads, HW and OS limits still apply.
+Added a log warning if more threads are defined than active CPUs in affinity mask.
+Improved merkle tree memory management for stratum.
+Added transaction count to New Work log.
+Other small improvements.
+
+v3.22.1
+
+#393 fixed segfault in GBT, regression from v3.22.0.
+More efficient 32 bit data interleaving.
+
+v3.22.0
+
+Stratum: faster netdiff calculation.
+Merged a few updates from Pooler/cpuminer:
+   Use CURLOPT_POSTFIELDS in json_rpc_call,
+   Use CURLINFO_ACTIVESOCKET when supported,
+   JSONRPC speedup,
+   Speed up hex2bin function.  
+Small log improvements, notably more frequent hash rate reports.
+Removed decred algo.
+
+v3.21.5
+
+All issues with v3.21.3 & v3.21.4 should be resolved.
+Changes since v3.21.2:
+#392 #379 #389 Fixed misaligned address segfault solo mining.
+#392 Fixed stats for myr-gr algo, and a few others, for CPUs without AVX2.
+#392 Fixed conditional mining.
+#392 Fixed cpu affinity on Ryzen CPUs using Windows binaries,
+     Windows binaries no longer support CPU groups,
+     Windows binaries support CPUs with up to 64 threads.
+Small optimizations to serialized vectoring.
+
+v3.21.4 CANCELLED
+
+Reapply selected changes from v3.21.3.
+#392 #379 #389 Fixed misaligned address segfault solo mining.
+#392 Fixed conditional mining.
+#392 Fixed cpu affinity on Ryzen CPUs using Windows binaries,
+     Windows binaries no longer support CPU groups,
+     Windows binaries support CPUs with up to 64 threads.
+
+v3.21.3.1 UNRELEASED
+
+Revert to 3.21.2
+
+v3.21.3 CANCELLED
+
+#392 #379 #389 Fixed misaligned address segfault solo mining.
+#392 Fixed stats for myr-gr algo, and a few others, for CPUs without AVX2.
+#392 Fixed conditional mining.
+#392 Fixed cpu affinity on Ryzen CPUs using Windows binaries,
+     Windows binaries no longer support CPU groups,
+     Windows binaries support CPUs with up to 64 threads.
+Midstate prehash is now centralized, done only once instead of by every thread
+for selected algos.
+Small optimizations to serialized vectoring.
+
+v3.21.2 
+
+Faster SALSA SIMD shuffle for yespower, yescrypt & scryptn2.
+Fixed a couple of compiler warnings with gcc-12.
+
+v3.21.1
+
+Fixed a segfault in some obsolete algos.
+Small optimizations to Hamsi & Shabal AVX2 & AVX512.
+
+v3.21.0
+
+Added minotaurx algo for stratum only.
+Blake256 & sha256 prehash optimized to ignore zero-padded data for AVX2 & AVX512.
+Other small improvements.
+
+v3.20.3
+
+Faster c11 algo: AVX512 6%, AVX2 4%, AVX2+VAES 15%.
+Faster AVX2+VAES for anime 14%, hmq1725 6%.
+Small optimizations to Luffa AVX2 & AVX512.
+
+v3.20.2
+
+Bit rotation optimizations to Blake256, Blake512, Blake2b, Blake2s & Lyra2-blake2b for SSE2 & AVX2.
+Removed old unused yescrypt library and other unused code.
+
 v3.20.1

 sph_blake2b optimized 1-way SSSE3 & AVX2.
 Removed duplicate Blake2b used by Power2b algo, will now use optimized sph_blake2b.
 Removed imprecise hash & target display from rejected share log.
-Share and target difficulty is now displayed only for low diificulty shares.
+Share and target difficulty is now displayed only for low difficulty shares.
 Updated configure.ac to check for AVX512 asm support.
 Small optimization to Lyra2 SSE2.

@@ -87,12 +206,9 @@ v3.19.8

 #370 "stratum+ssl", in addition to "stratum+tcps", is now recognized as a valid
 url protocol specifier for requesting a secure stratum connection.
-
 The full url, including the protocol, is now displayed in the stratum connect
 log and the periodic summary log.
-
 Small optimizations to Cubehash, AVX2 & AVX512.
-
 Byte order and prehash optimizations for Blake256 & Blake512, AVX2 & AVX512.

 v3.19.7
@@ -110,40 +226,29 @@ v3.19.5

 Enhanced stratum-keepalive preemptively resets the stratum connection
 before the server to avoid lost shares.
-
 Added build-msys2.sh shell script for easier compiling on Windows, see Wiki for details.
-
 X16RT: eliminate unnecessary recalculations of the hash order.
-
 Fix a few compiler warnings.
-
 Fixed log colour error when a block is solved.

 v3.19.4

 #359: Fix verthash memory allocation for non-hugepages, broken in v3.19.3.
-
 New option stratum-keepalive prevents stratum timeouts when no shares are
 submitted for several minutes due to high difficulty.
-
 Fixed a bug displaying optimizations for some algos.

 v3.19.3

 Linux: Faster verthash (+25%), scryptn2 (+2%) when huge pages are available.
-
 Small speed up for Hamsi AVX2 & AVX512, Keccak AVX512.

 v3.19.2

 Fixed log displaying incorrect memory usage for scrypt, broken in v3.19.1.
-
 Reduce log noise when replies to submitted shares are lost due to stratum errors.
-
 Fugue prehash optimization for X16r family AVX2 & AVX512.
-
 Small speed improvement for Hamsi AVX2 & AVX512.
-
 Win: With CPU groups enabled the number of CPUs displayed in the ASCII art
 affinity map is the number of CPUs in a CPU group, was number of CPUs up to 64.

@@ -155,7 +260,6 @@ Changes to Windows binaries package:
 - zen build renamed to avx2-sha, supports Zen1 & Zen2,
 - avx512-sha build removed, Rocketlake CPUs can use avx512-sha-vaes,
 - see README.txt for compatibility details.
-
 Fixed a few compiler warnings that are new in GCC 11.
 Other minor fixes.

@@ -169,22 +273,17 @@ Changes to cpu-affinity:
  - streamlined code for more efficient initialization of miner threads,
  - precise affining of each miner thread to a specific CPU,
  - added an option to disable CPU affinity with "--cpu-affinity 0"
-
 Faster sha256t with AVX512 & AVX2.
-
 Added stratum error count to stats log, reported only when non-zero.

 v3.18.2

 Issue #342, fixed Groestl AES on Windows, broken in v3.18.0.
-
 AVX512 for sha256d.
-
 SSE42 and AVX may now be displayed as mining features at startup.
 This is hard coded for each algo, and is only implemented for scrypt
 at this time as it is the only algo with significant performance differences
 with those features.
-
 Fixed an issue where a high hashrate algo could cause excessive invalid hash
 rate log reports when starting up in benchmark mode.

@@ -195,9 +294,7 @@ More speed for scrypt:
 - AVX2 is now used by default on CPUS with SHA but not AVX512,
 - scrypt:1024 performance lost in v3.18.0 is restored,
 - AVX512 & AVX2 improvements to scrypt:1024.
-
 Big speedup for SwiFFTx AVX2 & SSE4.1: x22i +55%, x25x +22%.
-
 Issue #337: fixed a problem that could display negative stats values in the
 first summary report if the report was forced prematurely due to a stratum
 diff change. The stats will still be invalid but should display zeros.
@@ -210,26 +307,19 @@ Complete rewrite of Scrypt code, optimized for large N factor (scryptn2):
  - memory requirements reduced 30-60% depending on CPU architecture,
  - memory usage displayed at startup,
  - scrypt, default N=1024 (LTC), will likely perform slower.
-
 Improved stale share detection and handling for Scrypt with large N factor:
  - abort and discard partially computed hash when new work is detected,
  - quicker response to new job, less time wasted mining stale job.
-
 Improved stale share handling for all algorithms:
  - report possible stale share when new work received with a previously
    submitted share still pending,
  - when new work is detected report the submission of an already completed,
    otherwise valid, but likely stale, share,
  - fixed incorrect block height in stale share log.
-
 Small performance improvements to sha, bmw, cube & hamsi for AVX512 & AVX2.
-
 When stratum disconnects miner threads go to idle until reconnected.
-
 Colour changes to some logs.
-
 Some low level function name changes for clarity and consistency.
-
 The reference hashrate in the summary log and the benchmark total hashrate
 are now the mean hashrate for the session. 

@@ -342,7 +432,6 @@ Fixed neoscrypt BUG log.
 v3.14.3

 #265: more mutex changes to reduce blocking with high thread count.
-
 #267: fixed hodl algo potential memory alignment issue,
      add warning when thread count is not valid for mining hodl algo.

--- a/aclocal.m4
+++ b/aclocal.m4
@@ -1,6 +1,6 @@
-# generated automatically by aclocal 1.16.1 -*- Autoconf -*-
+# generated automatically by aclocal 1.16.5 -*- Autoconf -*-

-# Copyright (C) 1996-2018 Free Software Foundation, Inc.
+# Copyright (C) 1996-2021 Free Software Foundation, Inc.

 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -14,13 +14,13 @@
 m4_ifndef([AC_CONFIG_MACRO_DIRS], [m4_defun([_AM_CONFIG_MACRO_DIRS], [])m4_defun([AC_CONFIG_MACRO_DIRS], [_AM_CONFIG_MACRO_DIRS($@)])])
 m4_ifndef([AC_AUTOCONF_VERSION],
  [m4_copy([m4_PACKAGE_VERSION], [AC_AUTOCONF_VERSION])])dnl
-m4_if(m4_defn([AC_AUTOCONF_VERSION]), [2.69],,
-[m4_warning([this file was generated for autoconf 2.69.
+m4_if(m4_defn([AC_AUTOCONF_VERSION]), [2.71],,
+[m4_warning([this file was generated for autoconf 2.71.
 You have another version of autoconf.  It may work, but is not guaranteed to.
 If you have problems, you may need to regenerate the build system entirely.
 To do so, use the procedure documented by the package, typically 'autoreconf'.])])

-# Copyright (C) 2002-2018 Free Software Foundation, Inc.
+# Copyright (C) 2002-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -35,7 +35,7 @@ AC_DEFUN([AM_AUTOMAKE_VERSION],
 [am__api_version='1.16'
 dnl Some users find AM_AUTOMAKE_VERSION and mistake it for a way to
 dnl require some minimum version.  Point them to the right macro.
-m4_if([$1], [1.16.1], [],
+m4_if([$1], [1.16.5], [],
      [AC_FATAL([Do not call $0, use AM_INIT_AUTOMAKE([$1]).])])dnl
 ])

@@ -51,14 +51,14 @@ m4_define([_AM_AUTOCONF_VERSION], [])
 # Call AM_AUTOMAKE_VERSION and AM_AUTOMAKE_VERSION so they can be traced.
 # This function is AC_REQUIREd by AM_INIT_AUTOMAKE.
 AC_DEFUN([AM_SET_CURRENT_AUTOMAKE_VERSION],
-[AM_AUTOMAKE_VERSION([1.16.1])dnl
+[AM_AUTOMAKE_VERSION([1.16.5])dnl
 m4_ifndef([AC_AUTOCONF_VERSION],
  [m4_copy([m4_PACKAGE_VERSION], [AC_AUTOCONF_VERSION])])dnl
 _AM_AUTOCONF_VERSION(m4_defn([AC_AUTOCONF_VERSION]))])

 # Figure out how to run the assembler.                      -*- Autoconf -*-

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -78,7 +78,7 @@ _AM_IF_OPTION([no-dependencies],, [_AM_DEPENDENCIES([CCAS])])dnl

 # AM_AUX_DIR_EXPAND                                         -*- Autoconf -*-

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -130,7 +130,7 @@ am_aux_dir=`cd "$ac_aux_dir" && pwd`

 # AM_CONDITIONAL                                            -*- Autoconf -*-

-# Copyright (C) 1997-2018 Free Software Foundation, Inc.
+# Copyright (C) 1997-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -161,7 +161,7 @@ AC_CONFIG_COMMANDS_PRE(
 Usually this means the macro was only invoked conditionally.]])
 fi])])

-# Copyright (C) 1999-2018 Free Software Foundation, Inc.
+# Copyright (C) 1999-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -352,7 +352,7 @@ _AM_SUBST_NOTMAKE([am__nodep])dnl

 # Generate code to set up dependency tracking.              -*- Autoconf -*-

-# Copyright (C) 1999-2018 Free Software Foundation, Inc.
+# Copyright (C) 1999-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -391,7 +391,9 @@ AC_DEFUN([_AM_OUTPUT_DEPENDENCY_COMMANDS],
  done
  if test $am_rc -ne 0; then
    AC_MSG_FAILURE([Something went wrong bootstrapping makefile fragments
-    for automatic dependency tracking.  Try re-running configure with the
+    for automatic dependency tracking.  If GNU make was not used, consider
+    re-running the configure script with MAKE="gmake" (or whatever is
+    necessary).  You can also try re-running configure with the
    '--disable-dependency-tracking' option to at least be able to build
    the package (albeit without support for automatic dependency tracking).])
  fi
@@ -418,7 +420,7 @@ AC_DEFUN([AM_OUTPUT_DEPENDENCY_COMMANDS],

 # Do all the work for Automake.                             -*- Autoconf -*-

-# Copyright (C) 1996-2018 Free Software Foundation, Inc.
+# Copyright (C) 1996-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -446,6 +448,10 @@ m4_defn([AC_PROG_CC])
 # release and drop the old call support.
 AC_DEFUN([AM_INIT_AUTOMAKE],
 [AC_PREREQ([2.65])dnl
+m4_ifdef([_$0_ALREADY_INIT],
+  [m4_fatal([$0 expanded multiple times
+]m4_defn([_$0_ALREADY_INIT]))],
+  [m4_define([_$0_ALREADY_INIT], m4_expansion_stack)])dnl
 dnl Autoconf wants to disallow AM_ names.  We explicitly allow
 dnl the ones we care about.
 m4_pattern_allow([^AM_[A-Z]+FLAGS$])dnl
@@ -482,7 +488,7 @@ m4_ifval([$3], [_AM_SET_OPTION([no-define])])dnl
 [_AM_SET_OPTIONS([$1])dnl
 dnl Diagnose old-style AC_INIT with new-style AM_AUTOMAKE_INIT.
 m4_if(
-  m4_ifdef([AC_PACKAGE_NAME], [ok]):m4_ifdef([AC_PACKAGE_VERSION], [ok]),
+  m4_ifset([AC_PACKAGE_NAME], [ok]):m4_ifset([AC_PACKAGE_VERSION], [ok]),
  [ok:ok],,
  [m4_fatal([AC_INIT should be called with package and version arguments])])dnl
 AC_SUBST([PACKAGE], ['AC_PACKAGE_TARNAME'])dnl
@@ -534,6 +540,20 @@ AC_PROVIDE_IFELSE([AC_PROG_OBJCXX],
 		  [m4_define([AC_PROG_OBJCXX],
 			     m4_defn([AC_PROG_OBJCXX])[_AM_DEPENDENCIES([OBJCXX])])])dnl
 ])
+# Variables for tags utilities; see am/tags.am
+if test -z "$CTAGS"; then
+  CTAGS=ctags
+fi
+AC_SUBST([CTAGS])
+if test -z "$ETAGS"; then
+  ETAGS=etags
+fi
+AC_SUBST([ETAGS])
+if test -z "$CSCOPE"; then
+  CSCOPE=cscope
+fi
+AC_SUBST([CSCOPE])
+
 AC_REQUIRE([AM_SILENT_RULES])dnl
 dnl The testsuite driver may need to know about EXEEXT, so add the
 dnl 'am__EXEEXT' conditional if _AM_COMPILER_EXEEXT was seen.  This
@@ -615,7 +635,7 @@ for _am_header in $config_headers :; do
 done
 echo "timestamp for $_am_arg" >`AS_DIRNAME(["$_am_arg"])`/stamp-h[]$_am_stamp_count])

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -636,7 +656,7 @@ if test x"${install_sh+set}" != xset; then
 fi
 AC_SUBST([install_sh])])

-# Copyright (C) 2003-2018 Free Software Foundation, Inc.
+# Copyright (C) 2003-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -658,7 +678,7 @@ AC_SUBST([am__leading_dot])])
 # Add --enable-maintainer-mode option to configure.         -*- Autoconf -*-
 # From Jim Meyering

-# Copyright (C) 1996-2018 Free Software Foundation, Inc.
+# Copyright (C) 1996-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -693,7 +713,7 @@ AC_MSG_CHECKING([whether to enable maintainer-specific portions of Makefiles])

 # Check to see how 'make' treats includes.	            -*- Autoconf -*-

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -736,7 +756,7 @@ AC_SUBST([am__quote])])

 # Fake the existence of programs that GNU maintainers use.  -*- Autoconf -*-

-# Copyright (C) 1997-2018 Free Software Foundation, Inc.
+# Copyright (C) 1997-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -757,12 +777,7 @@ AC_DEFUN([AM_MISSING_HAS_RUN],
 [AC_REQUIRE([AM_AUX_DIR_EXPAND])dnl
 AC_REQUIRE_AUX_FILE([missing])dnl
 if test x"${MISSING+set}" != xset; then
-  case $am_aux_dir in
-  *\ * | *\	*)
-    MISSING="\${SHELL} \"$am_aux_dir/missing\"" ;;
-  *)
-    MISSING="\${SHELL} $am_aux_dir/missing" ;;
-  esac
+  MISSING="\${SHELL} '$am_aux_dir/missing'"
 fi
 # Use eval to expand $SHELL
 if eval "$MISSING --is-lightweight"; then
@@ -775,7 +790,7 @@ fi

 # Helper functions for option handling.                     -*- Autoconf -*-

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -804,7 +819,7 @@ AC_DEFUN([_AM_SET_OPTIONS],
 AC_DEFUN([_AM_IF_OPTION],
 [m4_ifset(_AM_MANGLE_OPTION([$1]), [$2], [$3])])

-# Copyright (C) 1999-2018 Free Software Foundation, Inc.
+# Copyright (C) 1999-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -851,7 +866,7 @@ AC_LANG_POP([C])])
 # For backward compatibility.
 AC_DEFUN_ONCE([AM_PROG_CC_C_O], [AC_REQUIRE([AC_PROG_CC])])

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -870,7 +885,7 @@ AC_DEFUN([AM_RUN_LOG],

 # Check to make sure that the build environment is sane.    -*- Autoconf -*-

-# Copyright (C) 1996-2018 Free Software Foundation, Inc.
+# Copyright (C) 1996-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -951,7 +966,7 @@ AC_CONFIG_COMMANDS_PRE(
 rm -f conftest.file
 ])

-# Copyright (C) 2009-2018 Free Software Foundation, Inc.
+# Copyright (C) 2009-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -1011,7 +1026,7 @@ AC_SUBST([AM_BACKSLASH])dnl
 _AM_SUBST_NOTMAKE([AM_BACKSLASH])dnl
 ])

-# Copyright (C) 2001-2018 Free Software Foundation, Inc.
+# Copyright (C) 2001-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -1039,7 +1054,7 @@ fi
 INSTALL_STRIP_PROGRAM="\$(install_sh) -c -s"
 AC_SUBST([INSTALL_STRIP_PROGRAM])])

-# Copyright (C) 2006-2018 Free Software Foundation, Inc.
+# Copyright (C) 2006-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
@@ -1058,7 +1073,7 @@ AC_DEFUN([AM_SUBST_NOTMAKE], [_AM_SUBST_NOTMAKE($@)])

 # Check how to create a tarball.                            -*- Autoconf -*-

-# Copyright (C) 2004-2018 Free Software Foundation, Inc.
+# Copyright (C) 2004-2021 Free Software Foundation, Inc.
 #
 # This file is free software; the Free Software Foundation
 # gives unlimited permission to copy and/or distribute it,
--- a/algo-gate-api.c
+++ b/algo-gate-api.c
@@ -67,7 +67,6 @@ void do_nothing   () {}
 bool return_true  () { return true;  }
 bool return_false () { return false; }
 void *return_null () { return NULL;  }
-void call_error   () { printf("ERR: Uninitialized function pointer\n"); }

 void algo_not_tested()
 {
@@ -95,7 +94,8 @@ int null_scanhash()
   return 0;
 }

-// Default generic scanhash can be used in many cases.
+// Default generic scanhash can be used in many cases. Not to be used when
+// prehashing can be done or when byte swapping the data can be avoided.
 int scanhash_generic( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr )
 {
@@ -152,6 +152,9 @@ int scanhash_4way_64in_32out( struct work *work, uint32_t max_nonce,
   const bool bench = opt_benchmark;

   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+   // overwrite byte swapped nonce with original byte order for proper
+   // incrementing. The nonce only needs to byte swapped if it is to be
+   // sumbitted.
   *noncev = mm256_intrlv_blend_32(
                   _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
   do
@@ -168,7 +171,7 @@ int scanhash_4way_64in_32out( struct work *work, uint32_t max_nonce,
         }
      }
      *noncev = _mm256_add_epi32( *noncev,
-                                  m256_const1_64( 0x0000000400000000 ) );
+                                  _mm256_set1_epi64x( 0x0000000400000000 ) );
      n += 4;
   } while ( likely( ( n <= last_nonce ) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
@@ -224,7 +227,7 @@ int scanhash_8way_64in_32out( struct work *work, uint32_t max_nonce,
         }
      }
      *noncev = _mm512_add_epi32( *noncev,
-                                  m512_const1_64( 0x0000000800000000 ) );
+                                  _mm512_set1_epi64( 0x0000000800000000 ) );
      n += 8;
   } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
@@ -245,7 +248,7 @@ int null_hash()
   return 0;
 };

-void init_algo_gate( algo_gate_t* gate )
+static void init_algo_gate( algo_gate_t* gate )
 {
   gate->miner_thread_init       = (void*)&return_true;
   gate->scanhash                = (void*)&scanhash_generic;
@@ -260,8 +263,6 @@ void init_algo_gate( algo_gate_t* gate )
   gate->build_block_header      = (void*)&std_build_block_header;
   gate->build_extraheader       = (void*)&std_build_extraheader;
   gate->set_work_data_endian    = (void*)&do_nothing;
-   gate->calc_network_diff       = (void*)&std_calc_network_diff;
-   gate->ready_to_mine           = (void*)&std_ready_to_mine;
   gate->resync_threads          = (void*)&do_nothing;
   gate->do_this_thread          = (void*)&return_true;
   gate->longpoll_rpc_call       = (void*)&std_longpoll_rpc_call;
@@ -305,7 +306,6 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    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;
@@ -324,6 +324,7 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    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_MINOTAURX:    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;
@@ -336,9 +337,11 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    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_SHA256DT:     rc = register_sha256dt_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_SHA512256D:   rc = register_sha512256d_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;
@@ -371,15 +374,11 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    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_YESCRYPT:     rc = register_yescrypt_algo      ( gate ); break;
+    case ALGO_YESCRYPTR8:   rc = 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_YESCRYPTR16:  rc = register_yescryptr16_algo   ( gate ); break;
+    case ALGO_YESCRYPTR32:  rc = 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;
@@ -427,7 +426,6 @@ const char* const algo_alias_map[][2] =
  { "blake256r8",        "blakecoin"      },
  { "blake256r8vnl",     "vanilla"        },
  { "blake256r14",       "blake"          },
-  { "blake256r14dcr",    "decred"         },
  { "diamond",           "dmd-gr"         },
  { "espers",            "hmq1725"        },
  { "flax",              "c11"            },
--- a/algo-gate-api.h
+++ b/algo-gate-api.h
@@ -94,10 +94,13 @@ typedef  uint32_t set_t;
 #define SSE42_OPT        4
 #define AVX_OPT          8   // Sandybridge
 #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 SHA_OPT       0x20   // Zen1, Icelake (deprecated)
+#define AVX512_OPT    0x40   // Skylake-X, Zen4 (AVX512[F,VL,DQ,BW])
+#define VAES_OPT      0x80   // Icelake, Zen3

+// AVX10 does not have explicit algo features:
+//  AVX10_512 is compatible with AVX512 + VAES
+//  AVX10_256 is compatible with AVX2 + VAES

 // return set containing all elements from sets a & b
 inline set_t set_union ( set_t a, set_t b ) { return a | b; }
@@ -155,19 +158,13 @@ char* ( *malloc_txs_request )   ( struct work* );
 // Big endian or little endian
 void ( *set_work_data_endian )  ( struct work* );

-double ( *calc_network_diff )   ( struct work* );
-
-// Wait for first work
-bool ( *ready_to_mine )         ( struct work*, struct stratum_ctx*, int );
-
 // Diverge mining threads
 bool ( *do_this_thread )        ( int );

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

-// No longer needed
-json_t* (*longpoll_rpc_call)      ( CURL*, int*, char* );
+json_t* ( *longpoll_rpc_call )  ( CURL*, int*, char* );

 set_t optimizations;
 int  ( *get_work_data_size )     ();
@@ -270,7 +267,9 @@ void std_get_new_work( struct work *work, struct work *g_work, int thr_id,
                       uint32_t* end_nonce_ptr );

 void sha256d_gen_merkle_root( char *merkle_root, struct stratum_ctx *sctx );
-void SHA256_gen_merkle_root ( char *merkle_root, struct stratum_ctx *sctx );
+void sha256_gen_merkle_root ( char *merkle_root, struct stratum_ctx *sctx );
+// OpenSSL sha256 deprecated
+//void SHA256_gen_merkle_root ( char *merkle_root, struct stratum_ctx *sctx );

 bool std_le_work_decode( struct work *work );
 bool std_be_work_decode( struct work *work );
@@ -286,8 +285,6 @@ char* std_malloc_txs_request( struct work *work );
 // 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 );
-
 void std_build_block_header( struct work* g_work, uint32_t version,
 	                          uint32_t *prevhash,  uint32_t *merkle_root,
   	                       uint32_t ntime,      uint32_t nbits,
@@ -297,9 +294,6 @@ void std_build_extraheader( struct work *work, struct stratum_ctx *sctx );

 json_t* std_longpoll_rpc_call( CURL *curl, int *err, char *lp_url );

-bool std_ready_to_mine( struct work* work, struct stratum_ctx* stratum,
-                        int thr_id );
-
 int std_get_work_data_size();

 // Gate admin functions
--- a/algo/argon2/argon2a/argon2a.c
+++ b/algo/argon2/argon2a/argon2a.c
@@ -77,7 +77,7 @@ bool register_argon2_algo( algo_gate_t* gate )
  gate->optimizations = SSE2_OPT | AVX_OPT | AVX2_OPT;
  gate->scanhash        = (void*)&scanhash_argon2;
  gate->hash            = (void*)&argon2hash;
-  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
+  gate->gen_merkle_root = (void*)&sha256_gen_merkle_root;
  opt_target_factor = 65536.0;

  return true;
--- a/algo/blake/blake-hash-4way.h
+++ b/algo/blake/blake-hash-4way.h
@@ -1,60 +1,15 @@
-/* $Id: sph_blake.h 252 2011-06-07 17:55:14Z tp $ */
-/**
- * BLAKE interface. BLAKE is a family of functions which differ by their
- * output size; this implementation defines BLAKE for output sizes 224,
- * 256, 384 and 512 bits. This implementation conforms to the "third
- * round" specification.
- *
- * ==========================(LICENSE BEGIN)============================
- *
- * Copyright (c) 2007-2010  Projet RNRT SAPHIR
- * 
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- * 
- * The above copyright notice and this permission notice shall be
- * included in all copies or substantial portions of the Software.
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * ===========================(LICENSE END)=============================
- *
- * @file     sph_blake.h
- * @author   Thomas Pornin <thomas.pornin@cryptolog.com>
- */
-
-#ifndef __BLAKE_HASH_4WAY__
-#define __BLAKE_HASH_4WAY__ 1
-
-#ifdef __cplusplus
-extern "C"{
-#endif
+#ifndef BLAKE_HASH_4WAY__
+#define BLAKE_HASH_4WAY__ 1

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
 #include "simd-utils.h"

-#define SPH_SIZE_blake256   256
-
-#define SPH_SIZE_blake512   512
-
 /////////////////////////
 //
 //  Blake-256 1 way SSE2

 void  blake256_transform_le( uint32_t *H, const uint32_t *buf,
-                             const uint32_t T0, const uint32_t T1 );
+                             const uint32_t T0, const uint32_t T1, int rounds );

 /////////////////////////
 //
@@ -75,13 +30,13 @@ typedef struct {
   int rounds;   // 14 for blake, 8 for blakecoin & vanilla
 } blake_4way_small_context __attribute__ ((aligned (64)));

-// Default, 14 rounds, blake, decred
+// Default, 14 rounds
 typedef blake_4way_small_context blake256_4way_context;
 void blake256_4way_init(void *ctx);
 void blake256_4way_update(void *ctx, const void *data, size_t len);
 void blake256_4way_close(void *ctx, void *dst);

-// 14 rounds, blake, decred
+// 14 rounds
 typedef blake_4way_small_context blake256r14_4way_context;
 void blake256r14_4way_init(void *cc);
 void blake256r14_4way_update(void *cc, const void *data, size_t len);
@@ -103,7 +58,7 @@ typedef struct {
   __m256i buf[16] __attribute__ ((aligned (64)));
   __m256i H[8];
   size_t ptr;
-   sph_u32 T0, T1;
+   uint32_t T0, T1;
   int rounds;   // 14 for blake, 8 for blakecoin & vanilla
 } blake_8way_small_context;

@@ -115,9 +70,9 @@ void blake256_8way_close(void *cc, void *dst);
 void blake256_8way_update_le(void *cc, const void *data, size_t len);
 void blake256_8way_close_le(void *cc, void *dst);
 void blake256_8way_round0_prehash_le( void *midstate, const void *midhash,
-                                       const void *data );
+                                      void *data );
 void blake256_8way_final_rounds_le( void *final_hash, const void *midstate,
-                                     const void *midhash, const void *data );
+                    const void *midhash, const void *data, const int rounds );

 // 14 rounds, blake, decred
 typedef blake_8way_small_context blake256r14_8way_context;
@@ -138,7 +93,7 @@ typedef struct {
   __m256i H[8];
   __m256i S[4];   
   size_t ptr;
-   sph_u64 T0, T1;
+   uint64_t T0, T1;
 } blake_4way_big_context __attribute__ ((aligned (128)));

 typedef blake_4way_big_context blake512_4way_context;
@@ -178,9 +133,9 @@ void blake256_16way_close(void *cc, void *dst);
 void blake256_16way_update_le(void *cc, const void *data, size_t len);
 void blake256_16way_close_le(void *cc, void *dst);
 void blake256_16way_round0_prehash_le( void *midstate, const void *midhash,
-                                       const void *data );
+                                       void *data );
 void blake256_16way_final_rounds_le( void *final_hash, const void *midstate,
-                                     const void *midhash, const void *data );
+                     const void *midhash, const void *data, const int rounds );


 // 14 rounds, blake, decred
@@ -204,7 +159,7 @@ typedef struct {
   __m512i H[8];
   __m512i S[4];
   size_t ptr;
-   sph_u64 T0, T1;
+   uint64_t T0, T1;
 } blake_8way_big_context __attribute__ ((aligned (128)));

 typedef blake_8way_big_context blake512_8way_context;
@@ -224,8 +179,4 @@ void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,
 #endif  // AVX512
 #endif  // AVX2

-#ifdef __cplusplus
-}
-#endif
-
 #endif  // BLAKE_HASH_4WAY_H__
--- a/algo/blake/blake256-hash-4way.c
+++ b/algo/blake/blake256-hash-4way.c
--- a/algo/blake/blake2b-hash-4way.c
+++ b/algo/blake/blake2b-hash-4way.c
@@ -252,14 +252,14 @@ static void blake2b_8way_compress( blake2b_8way_ctx *ctx, int last )
   v[ 5] = ctx->h[5];
   v[ 6] = ctx->h[6];
   v[ 7] = ctx->h[7];
-   v[ 8] = m512_const1_64( 0x6A09E667F3BCC908 );
-   v[ 9] = m512_const1_64( 0xBB67AE8584CAA73B );
-   v[10] = m512_const1_64( 0x3C6EF372FE94F82B );
-   v[11] = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   v[12] = m512_const1_64( 0x510E527FADE682D1 );
-   v[13] = m512_const1_64( 0x9B05688C2B3E6C1F );
-   v[14] = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   v[15] = m512_const1_64( 0x5BE0CD19137E2179 );
+   v[ 8] = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   v[ 9] = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   v[10] = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   v[11] = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   v[12] = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   v[13] = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   v[14] = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   v[15] = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

   v[12] = _mm512_xor_si512( v[12], _mm512_set1_epi64( ctx->t[0] ) );
   v[13] = _mm512_xor_si512( v[13], _mm512_set1_epi64( ctx->t[1] ) );
@@ -310,16 +310,16 @@ int blake2b_8way_init( blake2b_8way_ctx *ctx )
 {
   size_t i;

-   ctx->h[0] = m512_const1_64( 0x6A09E667F3BCC908 );
-   ctx->h[1] = m512_const1_64( 0xBB67AE8584CAA73B );
-   ctx->h[2] = m512_const1_64( 0x3C6EF372FE94F82B );
-   ctx->h[3] = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   ctx->h[4] = m512_const1_64( 0x510E527FADE682D1 );
-   ctx->h[5] = m512_const1_64( 0x9B05688C2B3E6C1F );
-   ctx->h[6] = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   ctx->h[7] = m512_const1_64( 0x5BE0CD19137E2179 );
+   ctx->h[0] = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   ctx->h[1] = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   ctx->h[2] = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   ctx->h[3] = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   ctx->h[4] = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   ctx->h[5] = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   ctx->h[6] = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   ctx->h[7] = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

-   ctx->h[0] = _mm512_xor_si512( ctx->h[0], m512_const1_64( 0x01010020 ) );
+   ctx->h[0] = _mm512_xor_si512( ctx->h[0], _mm512_set1_epi64( 0x01010020 ) );

   ctx->t[0] = 0;
   ctx->t[1] = 0;
@@ -388,11 +388,11 @@ void blake2b_8way_final( blake2b_8way_ctx *ctx, void *out )
 #define B2B_G(a, b, c, d, x, y) \
 { \
   v[a] = _mm256_add_epi64( _mm256_add_epi64( v[a], v[b] ), x ); \
-	v[d] = mm256_ror_64( _mm256_xor_si256( v[d], v[a] ), 32 ); \
+	v[d] = mm256_swap64_32( _mm256_xor_si256( v[d], v[a] ) ); \
 	v[c] = _mm256_add_epi64( v[c], v[d] ); \
-	v[b] = mm256_ror_64( _mm256_xor_si256( v[b], v[c] ), 24 ); \
+	v[b] = mm256_shuflr64_24( _mm256_xor_si256( v[b], v[c] ) ); \
 	v[a] = _mm256_add_epi64( _mm256_add_epi64( v[a], v[b] ), y ); \
-	v[d] = mm256_ror_64( _mm256_xor_si256( v[d], v[a] ), 16 ); \
+	v[d] = mm256_shuflr64_16( _mm256_xor_si256( v[d], v[a] ) ); \
 	v[c] = _mm256_add_epi64( v[c], v[d] ); \
 	v[b] = mm256_ror_64( _mm256_xor_si256( v[b], v[c] ), 63 ); \
 }
@@ -419,14 +419,14 @@ static void blake2b_4way_compress( blake2b_4way_ctx *ctx, int last )
   v[ 5] = ctx->h[5];
   v[ 6] = ctx->h[6];
   v[ 7] = ctx->h[7];
-   v[ 8] = m256_const1_64( 0x6A09E667F3BCC908 );
-   v[ 9] = m256_const1_64( 0xBB67AE8584CAA73B );
-   v[10] = m256_const1_64( 0x3C6EF372FE94F82B );
-   v[11] = m256_const1_64( 0xA54FF53A5F1D36F1 );
-   v[12] = m256_const1_64( 0x510E527FADE682D1 );
-   v[13] = m256_const1_64( 0x9B05688C2B3E6C1F );
-   v[14] = m256_const1_64( 0x1F83D9ABFB41BD6B );
-   v[15] = m256_const1_64( 0x5BE0CD19137E2179 );
+   v[ 8] = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
+   v[ 9] = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
+   v[10] = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
+   v[11] = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
+   v[12] = _mm256_set1_epi64x( 0x510E527FADE682D1 );
+   v[13] = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
+   v[14] = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
+   v[15] = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

   v[12] = _mm256_xor_si256( v[12], _mm256_set1_epi64x( ctx->t[0] ) );
   v[13] = _mm256_xor_si256( v[13], _mm256_set1_epi64x( ctx->t[1] ) );
@@ -477,16 +477,16 @@ int blake2b_4way_init( blake2b_4way_ctx *ctx )
 {
 	size_t i;

-   ctx->h[0] = m256_const1_64( 0x6A09E667F3BCC908 );
-   ctx->h[1] = m256_const1_64( 0xBB67AE8584CAA73B );
-   ctx->h[2] = m256_const1_64( 0x3C6EF372FE94F82B );
-   ctx->h[3] = m256_const1_64( 0xA54FF53A5F1D36F1 );
-   ctx->h[4] = m256_const1_64( 0x510E527FADE682D1 );
-   ctx->h[5] = m256_const1_64( 0x9B05688C2B3E6C1F );
-   ctx->h[6] = m256_const1_64( 0x1F83D9ABFB41BD6B );
-   ctx->h[7] = m256_const1_64( 0x5BE0CD19137E2179 );
+   ctx->h[0] = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
+   ctx->h[1] = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
+   ctx->h[2] = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
+   ctx->h[3] = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
+   ctx->h[4] = _mm256_set1_epi64x( 0x510E527FADE682D1 );
+   ctx->h[5] = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
+   ctx->h[6] = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
+   ctx->h[7] = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

-   ctx->h[0] = _mm256_xor_si256( ctx->h[0], m256_const1_64( 0x01010020 ) );
+   ctx->h[0] = _mm256_xor_si256( ctx->h[0], _mm256_set1_epi64x( 0x01010020 ) );

 	ctx->t[0] = 0;
 	ctx->t[1] = 0;
--- a/algo/blake/blake2s-hash-4way.c
+++ b/algo/blake/blake2s-hash-4way.c
@@ -62,14 +62,14 @@ int blake2s_4way_init( blake2s_4way_state *S, const uint8_t outlen )

   memset( S, 0, sizeof( blake2s_4way_state ) );

-   S->h[0] = m128_const1_64( 0x6A09E6676A09E667ULL );
-   S->h[1] = m128_const1_64( 0xBB67AE85BB67AE85ULL );
-   S->h[2] = m128_const1_64( 0x3C6EF3723C6EF372ULL );
-   S->h[3] = m128_const1_64( 0xA54FF53AA54FF53AULL );
-   S->h[4] = m128_const1_64( 0x510E527F510E527FULL );
-   S->h[5] = m128_const1_64( 0x9B05688C9B05688CULL );
-   S->h[6] = m128_const1_64( 0x1F83D9AB1F83D9ABULL );
-   S->h[7] = m128_const1_64( 0x5BE0CD195BE0CD19ULL );
+   S->h[0] = _mm_set1_epi64x( 0x6A09E6676A09E667ULL );
+   S->h[1] = _mm_set1_epi64x( 0xBB67AE85BB67AE85ULL );
+   S->h[2] = _mm_set1_epi64x( 0x3C6EF3723C6EF372ULL );
+   S->h[3] = _mm_set1_epi64x( 0xA54FF53AA54FF53AULL );
+   S->h[4] = _mm_set1_epi64x( 0x510E527F510E527FULL );
+   S->h[5] = _mm_set1_epi64x( 0x9B05688C9B05688CULL );
+   S->h[6] = _mm_set1_epi64x( 0x1F83D9AB1F83D9ABULL );
+   S->h[7] = _mm_set1_epi64x( 0x5BE0CD195BE0CD19ULL );
   
 //   for( int i = 0; i < 8; ++i )
 //      S->h[i] = _mm_set1_epi32( blake2s_IV[i] );
@@ -90,29 +90,29 @@ int blake2s_4way_compress( blake2s_4way_state *S, const __m128i* block )
   memcpy_128( m, block, 16 );
   memcpy_128( v, S->h, 8 );

-   v[ 8] = m128_const1_64( 0x6A09E6676A09E667ULL );
-   v[ 9] = m128_const1_64( 0xBB67AE85BB67AE85ULL );
-   v[10] = m128_const1_64( 0x3C6EF3723C6EF372ULL );
-   v[11] = m128_const1_64( 0xA54FF53AA54FF53AULL );
+   v[ 8] = _mm_set1_epi64x( 0x6A09E6676A09E667ULL );
+   v[ 9] = _mm_set1_epi64x( 0xBB67AE85BB67AE85ULL );
+   v[10] = _mm_set1_epi64x( 0x3C6EF3723C6EF372ULL );
+   v[11] = _mm_set1_epi64x( 0xA54FF53AA54FF53AULL );
   v[12] = _mm_xor_si128( _mm_set1_epi32( S->t[0] ),
-                          m128_const1_64( 0x510E527F510E527FULL ) );
+                          _mm_set1_epi64x( 0x510E527F510E527FULL ) );
   v[13] = _mm_xor_si128( _mm_set1_epi32( S->t[1] ),
-                          m128_const1_64( 0x9B05688C9B05688CULL ) );
+                          _mm_set1_epi64x( 0x9B05688C9B05688CULL ) );
   v[14] = _mm_xor_si128( _mm_set1_epi32( S->f[0] ),
-                          m128_const1_64( 0x1F83D9AB1F83D9ABULL ) );
+                          _mm_set1_epi64x( 0x1F83D9AB1F83D9ABULL ) );
   v[15] = _mm_xor_si128( _mm_set1_epi32( S->f[1] ),
-                          m128_const1_64( 0x5BE0CD195BE0CD19ULL ) );
+                          _mm_set1_epi64x( 0x5BE0CD195BE0CD19ULL ) );

 #define G4W( sigma0, sigma1, a, b, c, d ) \
 do { \
   uint8_t s0 = sigma0; \
   uint8_t s1 = sigma1; \
   a = _mm_add_epi32( _mm_add_epi32( a, b ), m[ s0 ] ); \
-   d = mm128_ror_32( _mm_xor_si128( d, a ), 16 ); \
+   d = mm128_swap32_16( _mm_xor_si128( d, a ) ); \
   c = _mm_add_epi32( c, d ); \
   b = mm128_ror_32( _mm_xor_si128( b, c ), 12 ); \
   a = _mm_add_epi32( _mm_add_epi32( a, b ), m[ s1 ] ); \
-   d = mm128_ror_32( _mm_xor_si128( d, a ),  8 ); \
+   d = mm128_shuflr32_8( _mm_xor_si128( d, a ) ); \
   c = _mm_add_epi32( c, d ); \
   b = mm128_ror_32( _mm_xor_si128( b, c ),  7 ); \
 } while(0)
@@ -269,21 +269,21 @@ int blake2s_8way_compress( blake2s_8way_state *S, const __m256i *block )
   memcpy_256( m, block, 16 );
   memcpy_256( v, S->h, 8 );

-   v[ 8] = m256_const1_64( 0x6A09E6676A09E667ULL );
-   v[ 9] = m256_const1_64( 0xBB67AE85BB67AE85ULL );
-   v[10] = m256_const1_64( 0x3C6EF3723C6EF372ULL );
-   v[11] = m256_const1_64( 0xA54FF53AA54FF53AULL );
+   v[ 8] = _mm256_set1_epi64x( 0x6A09E6676A09E667ULL );
+   v[ 9] = _mm256_set1_epi64x( 0xBB67AE85BB67AE85ULL );
+   v[10] = _mm256_set1_epi64x( 0x3C6EF3723C6EF372ULL );
+   v[11] = _mm256_set1_epi64x( 0xA54FF53AA54FF53AULL );
   v[12] = _mm256_xor_si256( _mm256_set1_epi32( S->t[0] ),
-                          m256_const1_64( 0x510E527F510E527FULL ) );
+                          _mm256_set1_epi64x( 0x510E527F510E527FULL ) );

   v[13] = _mm256_xor_si256( _mm256_set1_epi32( S->t[1] ),
-                          m256_const1_64( 0x9B05688C9B05688CULL ) );
+                          _mm256_set1_epi64x( 0x9B05688C9B05688CULL ) );

   v[14] = _mm256_xor_si256( _mm256_set1_epi32( S->f[0] ),
-                          m256_const1_64( 0x1F83D9AB1F83D9ABULL ) );
+                          _mm256_set1_epi64x( 0x1F83D9AB1F83D9ABULL ) );

   v[15] = _mm256_xor_si256( _mm256_set1_epi32( S->f[1] ),
-                          m256_const1_64( 0x5BE0CD195BE0CD19ULL ) );
+                          _mm256_set1_epi64x( 0x5BE0CD195BE0CD19ULL ) );

 /*
   v[ 8] = _mm256_set1_epi32( blake2s_IV[0] );
@@ -320,11 +320,11 @@ do { \
   uint8_t s0 = sigma0; \
   uint8_t s1 = sigma1; \
   a = _mm256_add_epi32( _mm256_add_epi32( a, b ), m[ s0 ] ); \
-   d = mm256_ror_32( _mm256_xor_si256( d, a ), 16 ); \
+   d = mm256_swap32_16( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi32( c, d ); \
   b = mm256_ror_32( _mm256_xor_si256( b, c ), 12 ); \
   a = _mm256_add_epi32( _mm256_add_epi32( a, b ), m[ s1 ] ); \
-   d = mm256_ror_32( _mm256_xor_si256( d, a ),  8 ); \
+   d = mm256_shuflr32_8( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi32( c, d ); \
   b = mm256_ror_32( _mm256_xor_si256( b, c ),  7 ); \
 } while(0)
@@ -391,14 +391,14 @@ int blake2s_8way_init( blake2s_8way_state *S, const uint8_t outlen )
   memset( P->personal, 0, sizeof( P->personal ) );

   memset( S, 0, sizeof( blake2s_8way_state ) );
-   S->h[0] = m256_const1_64( 0x6A09E6676A09E667ULL );
-   S->h[1] = m256_const1_64( 0xBB67AE85BB67AE85ULL );
-   S->h[2] = m256_const1_64( 0x3C6EF3723C6EF372ULL );
-   S->h[3] = m256_const1_64( 0xA54FF53AA54FF53AULL );
-   S->h[4] = m256_const1_64( 0x510E527F510E527FULL );
-   S->h[5] = m256_const1_64( 0x9B05688C9B05688CULL );
-   S->h[6] = m256_const1_64( 0x1F83D9AB1F83D9ABULL );
-   S->h[7] = m256_const1_64( 0x5BE0CD195BE0CD19ULL );
+   S->h[0] = _mm256_set1_epi64x( 0x6A09E6676A09E667ULL );
+   S->h[1] = _mm256_set1_epi64x( 0xBB67AE85BB67AE85ULL );
+   S->h[2] = _mm256_set1_epi64x( 0x3C6EF3723C6EF372ULL );
+   S->h[3] = _mm256_set1_epi64x( 0xA54FF53AA54FF53AULL );
+   S->h[4] = _mm256_set1_epi64x( 0x510E527F510E527FULL );
+   S->h[5] = _mm256_set1_epi64x( 0x9B05688C9B05688CULL );
+   S->h[6] = _mm256_set1_epi64x( 0x1F83D9AB1F83D9ABULL );
+   S->h[7] = _mm256_set1_epi64x( 0x5BE0CD195BE0CD19ULL );


 //   for( int i = 0; i < 8; ++i )
@@ -510,21 +510,21 @@ int blake2s_16way_compress( blake2s_16way_state *S, const __m512i *block )
   memcpy_512( m, block, 16 );
   memcpy_512( v, S->h, 8 );

-   v[ 8] = m512_const1_64( 0x6A09E6676A09E667ULL );
-   v[ 9] = m512_const1_64( 0xBB67AE85BB67AE85ULL );
-   v[10] = m512_const1_64( 0x3C6EF3723C6EF372ULL );
-   v[11] = m512_const1_64( 0xA54FF53AA54FF53AULL );
+   v[ 8] = _mm512_set1_epi64( 0x6A09E6676A09E667ULL );
+   v[ 9] = _mm512_set1_epi64( 0xBB67AE85BB67AE85ULL );
+   v[10] = _mm512_set1_epi64( 0x3C6EF3723C6EF372ULL );
+   v[11] = _mm512_set1_epi64( 0xA54FF53AA54FF53AULL );
   v[12] = _mm512_xor_si512( _mm512_set1_epi32( S->t[0] ),
-                          m512_const1_64( 0x510E527F510E527FULL ) );
+                          _mm512_set1_epi64( 0x510E527F510E527FULL ) );

   v[13] = _mm512_xor_si512( _mm512_set1_epi32( S->t[1] ),
-                          m512_const1_64( 0x9B05688C9B05688CULL ) );
+                          _mm512_set1_epi64( 0x9B05688C9B05688CULL ) );

   v[14] = _mm512_xor_si512( _mm512_set1_epi32( S->f[0] ),
-                          m512_const1_64( 0x1F83D9AB1F83D9ABULL ) );
+                          _mm512_set1_epi64( 0x1F83D9AB1F83D9ABULL ) );

   v[15] = _mm512_xor_si512( _mm512_set1_epi32( S->f[1] ),
-                          m512_const1_64( 0x5BE0CD195BE0CD19ULL ) );
+                          _mm512_set1_epi64( 0x5BE0CD195BE0CD19ULL ) );


 #define G16W( sigma0, sigma1, a, b, c, d) \
@@ -589,14 +589,14 @@ int blake2s_16way_init( blake2s_16way_state *S, const uint8_t outlen )
   memset( P->personal, 0, sizeof( P->personal ) );

   memset( S, 0, sizeof( blake2s_16way_state ) );
-   S->h[0] = m512_const1_64( 0x6A09E6676A09E667ULL );
-   S->h[1] = m512_const1_64( 0xBB67AE85BB67AE85ULL );
-   S->h[2] = m512_const1_64( 0x3C6EF3723C6EF372ULL );
-   S->h[3] = m512_const1_64( 0xA54FF53AA54FF53AULL );
-   S->h[4] = m512_const1_64( 0x510E527F510E527FULL );
-   S->h[5] = m512_const1_64( 0x9B05688C9B05688CULL );
-   S->h[6] = m512_const1_64( 0x1F83D9AB1F83D9ABULL );
-   S->h[7] = m512_const1_64( 0x5BE0CD195BE0CD19ULL );
+   S->h[0] = _mm512_set1_epi64( 0x6A09E6676A09E667ULL );
+   S->h[1] = _mm512_set1_epi64( 0xBB67AE85BB67AE85ULL );
+   S->h[2] = _mm512_set1_epi64( 0x3C6EF3723C6EF372ULL );
+   S->h[3] = _mm512_set1_epi64( 0xA54FF53AA54FF53AULL );
+   S->h[4] = _mm512_set1_epi64( 0x510E527F510E527FULL );
+   S->h[5] = _mm512_set1_epi64( 0x9B05688C9B05688CULL );
+   S->h[6] = _mm512_set1_epi64( 0x1F83D9AB1F83D9ABULL );
+   S->h[7] = _mm512_set1_epi64( 0x5BE0CD195BE0CD19ULL );

   uint32_t *p = ( uint32_t * )( P );

--- a/algo/blake/blake512-hash-4way.c
+++ b/algo/blake/blake512-hash-4way.c
@@ -1,62 +1,22 @@
-/* $Id: blake.c 252 2011-06-07 17:55:14Z tp $ */
-/*
- * BLAKE implementation.
- *
- * ==========================(LICENSE BEGIN)============================
- *
- * Copyright (c) 2007-2010  Projet RNRT SAPHIR
- *
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice shall be
- * included in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * ===========================(LICENSE END)=============================
- *
- * @author   Thomas Pornin <thomas.pornin@cryptolog.com>
- */
-
 #if defined (__AVX2__)

 #include <stddef.h>
 #include <string.h>
 #include <limits.h>
-
 #include "blake-hash-4way.h"

-#ifdef __cplusplus
-extern "C"{
-#endif
-
-#ifdef _MSC_VER
-#pragma warning (disable: 4146)
-#endif
-
 // Blake-512 common
   
 /*
-static const sph_u64 IV512[8] = {
-	SPH_C64(0x6A09E667F3BCC908), SPH_C64(0xBB67AE8584CAA73B),
-	SPH_C64(0x3C6EF372FE94F82B), SPH_C64(0xA54FF53A5F1D36F1),
-	SPH_C64(0x510E527FADE682D1), SPH_C64(0x9B05688C2B3E6C1F),
-	SPH_C64(0x1F83D9ABFB41BD6B), SPH_C64(0x5BE0CD19137E2179)
+static const uint64_t IV512[8] =
+{
+  0x6A09E667F3BCC908, 0xBB67AE8584CAA73B,
+  0x3C6EF372FE94F82B, 0xA54FF53A5F1D36F1,
+  0x510E527FADE682D1, 0x9B05688C2B3E6C1F,
+  0x1F83D9ABFB41BD6B, 0x5BE0CD19137E2179
 };

-static const sph_u64 salt_zero_big[4] = { 0, 0, 0, 0 };
+static const uint64_t salt_zero_big[4] = { 0, 0, 0, 0 };

 static const unsigned sigma[16][16] = {
 	{  0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15 },
@@ -77,15 +37,15 @@ static const unsigned sigma[16][16] = {
 	{  2, 12,  6, 10,  0, 11,  8,  3,  4, 13,  7,  5, 15, 14,  1,  9 }
 };

-static const sph_u64 CB[16] = {
-   SPH_C64(0x243F6A8885A308D3), SPH_C64(0x13198A2E03707344),
-   SPH_C64(0xA4093822299F31D0), SPH_C64(0x082EFA98EC4E6C89),
-   SPH_C64(0x452821E638D01377), SPH_C64(0xBE5466CF34E90C6C),
-   SPH_C64(0xC0AC29B7C97C50DD), SPH_C64(0x3F84D5B5B5470917),
-   SPH_C64(0x9216D5D98979FB1B), SPH_C64(0xD1310BA698DFB5AC),
-   SPH_C64(0x2FFD72DBD01ADFB7), SPH_C64(0xB8E1AFED6A267E96),
-   SPH_C64(0xBA7C9045F12C7F99), SPH_C64(0x24A19947B3916CF7),
-   SPH_C64(0x0801F2E2858EFC16), SPH_C64(0x636920D871574E69)
+static const uint64_t CB[16] = {
+   0x243F6A8885A308D3, 0x13198A2E03707344,
+   0xA4093822299F31D0, 0x082EFA98EC4E6C89,
+   0x452821E638D01377, 0xBE5466CF34E90C6C,
+   0xC0AC29B7C97C50DD, 0x3F84D5B5B5470917,
+   0x9216D5D98979FB1B, 0xD1310BA698DFB5AC,
+   0x2FFD72DBD01ADFB7, 0xB8E1AFED6A267E96,
+   0xBA7C9045F12C7F99, 0x24A19947B3916CF7,
+   0x0801F2E2858EFC16, 0x636920D871574E69

 */

@@ -314,10 +274,11 @@ static const sph_u64 CB[16] = {

 // Blake-512 8 way AVX512

-#define GB_8WAY(m0, m1, c0, c1, a, b, c, d)   do { \
+#define GB_8WAY( m0, m1, c0, c1, a, b, c, d ) \
+{ \
   a = _mm512_add_epi64( _mm512_add_epi64( _mm512_xor_si512( \
                 _mm512_set1_epi64( c1 ), m0 ), b ), a ); \
-   d = mm512_ror_64( _mm512_xor_si512( d, a ), 32 ); \
+   d = mm512_swap64_32( _mm512_xor_si512( d, a ) ); \
   c = _mm512_add_epi64( c, d ); \
   b = mm512_ror_64( _mm512_xor_si512( b, c ), 25 ); \
   a = _mm512_add_epi64( _mm512_add_epi64( _mm512_xor_si512( \
@@ -325,9 +286,10 @@ static const sph_u64 CB[16] = {
   d = mm512_ror_64( _mm512_xor_si512( d, a ), 16 ); \
   c = _mm512_add_epi64( c, d ); \
   b = mm512_ror_64( _mm512_xor_si512( b, c ), 11 ); \
-} while (0)
+}

-#define ROUND_B_8WAY(r)   do { \
+#define ROUND_B_8WAY( r ) \
+{ \
   GB_8WAY(Mx(r, 0), Mx(r, 1), CBx(r, 0), CBx(r, 1), V0, V4, V8, VC); \
   GB_8WAY(Mx(r, 2), Mx(r, 3), CBx(r, 2), CBx(r, 3), V1, V5, V9, VD); \
   GB_8WAY(Mx(r, 4), Mx(r, 5), CBx(r, 4), CBx(r, 5), V2, V6, VA, VE); \
@@ -336,19 +298,18 @@ static const sph_u64 CB[16] = {
   GB_8WAY(Mx(r, A), Mx(r, B), CBx(r, A), CBx(r, B), V1, V6, VB, VC); \
   GB_8WAY(Mx(r, C), Mx(r, D), CBx(r, C), CBx(r, D), V2, V7, V8, VD); \
   GB_8WAY(Mx(r, E), Mx(r, F), CBx(r, E), CBx(r, F), V3, V4, V9, VE); \
-   } while (0)
+}

 #define DECL_STATE64_8WAY \
   __m512i H0, H1, H2, H3, H4, H5, H6, H7; \
   uint64_t T0, T1;

-#define COMPRESS64_8WAY( buf )   do \
+#define COMPRESS64_8WAY( buf ) \
 { \
  __m512i M0, M1, M2, M3, M4, M5, M6, M7; \
  __m512i M8, M9, MA, MB, MC, MD, ME, MF; \
  __m512i V0, V1, V2, V3, V4, V5, V6, V7; \
  __m512i V8, V9, VA, VB, VC, VD, VE, VF; \
-  __m512i shuf_bswap64; \
  V0 = H0; \
  V1 = H1; \
  V2 = H2; \
@@ -357,18 +318,16 @@ static const sph_u64 CB[16] = {
  V5 = H5; \
  V6 = H6; \
  V7 = H7; \
-  V8 = m512_const1_64( CB0 );  \
-  V9 = m512_const1_64( CB1 );  \
-  VA = m512_const1_64( CB2 );  \
-  VB = m512_const1_64( CB3 );  \
+  V8 = _mm512_set1_epi64( CB0 );  \
+  V9 = _mm512_set1_epi64( CB1 );  \
+  VA = _mm512_set1_epi64( CB2 );  \
+  VB = _mm512_set1_epi64( CB3 );  \
  VC = _mm512_set1_epi64( T0 ^ CB4 ); \
  VD = _mm512_set1_epi64( T0 ^ CB5 ); \
  VE = _mm512_set1_epi64( T1 ^ CB6 ); \
  VF = _mm512_set1_epi64( T1 ^ CB7 ); \
-  shuf_bswap64 = m512_const_64( 0x38393a3b3c3d3e3f, 0x3031323334353637, \
-                                0x28292a2b2c2d2e2f, 0x2021222324252627, \
-                                0x18191a1b1c1d1e1f, 0x1011121314151617, \
-                                0x08090a0b0c0d0e0f, 0x0001020304050607 ); \
+  const __m512i shuf_bswap64 = mm512_bcast_m128( _mm_set_epi64x( \
+                                   0x08090a0b0c0d0e0f, 0x0001020304050607 ) ); \
  M0 = _mm512_shuffle_epi8( *(buf+ 0), shuf_bswap64 ); \
  M1 = _mm512_shuffle_epi8( *(buf+ 1), shuf_bswap64 ); \
  M2 = _mm512_shuffle_epi8( *(buf+ 2), shuf_bswap64 ); \
@@ -409,7 +368,7 @@ static const sph_u64 CB[16] = {
  H5 = mm512_xor3( VD, V5, H5 ); \
  H6 = mm512_xor3( VE, V6, H6 ); \
  H7 = mm512_xor3( VF, V7, H7 ); \
-} while (0)
+}

 void blake512_8way_compress( blake_8way_big_context *sc )
 { 
@@ -417,7 +376,6 @@ void blake512_8way_compress( blake_8way_big_context *sc )
  __m512i M8, M9, MA, MB, MC, MD, ME, MF;
  __m512i V0, V1, V2, V3, V4, V5, V6, V7;
  __m512i V8, V9, VA, VB, VC, VD, VE, VF;
-  __m512i shuf_bswap64;

  V0 = sc->H[0];
  V1 = sc->H[1];
@@ -427,19 +385,17 @@ void blake512_8way_compress( blake_8way_big_context *sc )
  V5 = sc->H[5];
  V6 = sc->H[6];
  V7 = sc->H[7];
-  V8 = m512_const1_64( CB0 );
-  V9 = m512_const1_64( CB1 );
-  VA = m512_const1_64( CB2 );
-  VB = m512_const1_64( CB3 );
+  V8 = _mm512_set1_epi64( CB0 );
+  V9 = _mm512_set1_epi64( CB1 );
+  VA = _mm512_set1_epi64( CB2 );
+  VB = _mm512_set1_epi64( CB3 );
  VC = _mm512_set1_epi64( sc->T0 ^ CB4 );
  VD = _mm512_set1_epi64( sc->T0 ^ CB5 );
  VE = _mm512_set1_epi64( sc->T1 ^ CB6 );
  VF = _mm512_set1_epi64( sc->T1 ^ CB7 );

-  shuf_bswap64 = m512_const_64( 0x38393a3b3c3d3e3f, 0x3031323334353637,
-                                0x28292a2b2c2d2e2f, 0x2021222324252627,
-                                0x18191a1b1c1d1e1f, 0x1011121314151617,
-                                0x08090a0b0c0d0e0f, 0x0001020304050607 );
+  const __m512i shuf_bswap64 = mm512_bcast_m128( _mm_set_epi64x( 
+                                   0x08090a0b0c0d0e0f, 0x0001020304050607 ) );

  M0 = _mm512_shuffle_epi8( sc->buf[ 0], shuf_bswap64 );
  M1 = _mm512_shuffle_epi8( sc->buf[ 1], shuf_bswap64 );
@@ -501,10 +457,10 @@ void blake512_8way_compress_le( blake_8way_big_context *sc )
  V5 = sc->H[5];
  V6 = sc->H[6];
  V7 = sc->H[7];
-  V8 = m512_const1_64( CB0 );
-  V9 = m512_const1_64( CB1 );
-  VA = m512_const1_64( CB2 );
-  VB = m512_const1_64( CB3 );
+  V8 = _mm512_set1_epi64( CB0 );
+  V9 = _mm512_set1_epi64( CB1 );
+  VA = _mm512_set1_epi64( CB2 );
+  VB = _mm512_set1_epi64( CB3 );
  VC = _mm512_set1_epi64( sc->T0 ^ CB4 );
  VD = _mm512_set1_epi64( sc->T0 ^ CB5 );
  VE = _mm512_set1_epi64( sc->T1 ^ CB6 );
@@ -563,23 +519,23 @@ void blake512_8way_prehash_le( blake_8way_big_context *sc, __m512i *midstate,
   __m512i V8, V9, VA, VB, VC, VD, VE, VF;

   // initial hash
-   casti_m512i( sc->H, 0 ) = m512_const1_64( 0x6A09E667F3BCC908 );
-   casti_m512i( sc->H, 1 ) = m512_const1_64( 0xBB67AE8584CAA73B );
-   casti_m512i( sc->H, 2 ) = m512_const1_64( 0x3C6EF372FE94F82B );
-   casti_m512i( sc->H, 3 ) = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m512i( sc->H, 4 ) = m512_const1_64( 0x510E527FADE682D1 );
-   casti_m512i( sc->H, 5 ) = m512_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m512i( sc->H, 6 ) = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m512i( sc->H, 7 ) = m512_const1_64( 0x5BE0CD19137E2179 );
+   casti_m512i( sc->H, 0 ) = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   casti_m512i( sc->H, 1 ) = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   casti_m512i( sc->H, 2 ) = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   casti_m512i( sc->H, 3 ) = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   casti_m512i( sc->H, 4 ) = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   casti_m512i( sc->H, 5 ) = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   casti_m512i( sc->H, 6 ) = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   casti_m512i( sc->H, 7 ) = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

   // fill buffer
   memcpy_512( sc->buf, (__m512i*)data, 80>>3 );
-   sc->buf[10] = m512_const1_64( 0x8000000000000000ULL );
+   sc->buf[10] = _mm512_set1_epi64( 0x8000000000000000ULL );
   sc->buf[11] = 
   sc->buf[12] = m512_zero;
   sc->buf[13] = m512_one_64;
   sc->buf[14] = m512_zero;
-   sc->buf[15] = m512_const1_64( 80*8 );
+   sc->buf[15] = _mm512_set1_epi64( 80*8 );

   // build working variables
   V0 = sc->H[0];
@@ -590,10 +546,10 @@ void blake512_8way_prehash_le( blake_8way_big_context *sc, __m512i *midstate,
   V5 = sc->H[5];
   V6 = sc->H[6];
   V7 = sc->H[7];
-   V8 = m512_const1_64( CB0 );
-   V9 = m512_const1_64( CB1 );
-   VA = m512_const1_64( CB2 );
-   VB = m512_const1_64( CB3 );
+   V8 = _mm512_set1_epi64( CB0 );
+   V9 = _mm512_set1_epi64( CB1 );
+   VA = _mm512_set1_epi64( CB2 );
+   VB = _mm512_set1_epi64( CB3 );
   VC = _mm512_set1_epi64( CB4 ^ 0x280ULL );
   VD = _mm512_set1_epi64( CB5 ^ 0x280ULL );
   VE = _mm512_set1_epi64( CB6 );
@@ -610,7 +566,7 @@ void blake512_8way_prehash_le( blake_8way_big_context *sc, __m512i *midstate,

   V0 = _mm512_add_epi64( _mm512_add_epi64( _mm512_xor_si512( 
                       _mm512_set1_epi64( CB9 ), sc->buf[ 8] ), V5 ), V0 ); 
-   VF = mm512_ror_64( _mm512_xor_si512( VF, V0 ), 32 ); 
+   VF = mm512_swap64_32( _mm512_xor_si512( VF, V0 ) ); 
   VA = _mm512_add_epi64( VA, VF ); 
   V5 = mm512_ror_64( _mm512_xor_si512( V5, VA ), 25 );
   V0 = _mm512_add_epi64( V0, V5 );
@@ -714,7 +670,7 @@ void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,
 //   V1 = _mm512_add_epi64( V1, _mm512_xor_si512( _mm512_set1_epi64( c1 ), m0 );

   V1 = _mm512_add_epi64( V1, V5 );   
-   VD = mm512_ror_64( _mm512_xor_si512( VD, V1 ), 32 );
+   VD = mm512_swap64_32( _mm512_xor_si512( VD, V1 ) );
   V9 = _mm512_add_epi64( V9, VD );
   V5 = mm512_ror_64( _mm512_xor_si512( V5, V9 ), 25 );
   V1 = _mm512_add_epi64( V1, _mm512_add_epi64( _mm512_xor_si512(
@@ -728,7 +684,7 @@ void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,
 //   V2 = _mm512_add_epi64( V2, V6 );
   V2 = _mm512_add_epi64( V2, _mm512_xor_si512( 
                 _mm512_set1_epi64( CBF ), M9 ) );
-   VE = mm512_ror_64( _mm512_xor_si512( VE, V2 ), 32 );
+   VE = mm512_swap64_32( _mm512_xor_si512( VE, V2 ) );
   VA = _mm512_add_epi64( VA, VE );
   V6 = mm512_ror_64( _mm512_xor_si512( V6, VA ), 25 );
   V2 = _mm512_add_epi64( V2, _mm512_add_epi64( _mm512_xor_si512(
@@ -742,7 +698,7 @@ void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,
 //   V3 = _mm512_add_epi64( V3, _mm512_add_epi64( _mm512_xor_si512( 
 //                 _mm512_set1_epi64( CBx(1, 7) ), Mx(1, 6) ), V7 ) ); 

-   VF = mm512_ror_64( _mm512_xor_si512( VF, V3 ), 32 ); 
+   VF = mm512_swap64_32( _mm512_xor_si512( VF, V3 ) ); 
   VB = _mm512_add_epi64( VB, VF ); 
   V7 = mm512_ror_64( _mm512_xor_si512( V7, VB ), 25 );
   V3 = _mm512_add_epi64( V3, _mm512_add_epi64( _mm512_xor_si512(
@@ -788,14 +744,14 @@ void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,

 void blake512_8way_init( blake_8way_big_context *sc )
 {
-   casti_m512i( sc->H, 0 ) = m512_const1_64( 0x6A09E667F3BCC908 );
-   casti_m512i( sc->H, 1 ) = m512_const1_64( 0xBB67AE8584CAA73B );
-   casti_m512i( sc->H, 2 ) = m512_const1_64( 0x3C6EF372FE94F82B );
-   casti_m512i( sc->H, 3 ) = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m512i( sc->H, 4 ) = m512_const1_64( 0x510E527FADE682D1 );
-   casti_m512i( sc->H, 5 ) = m512_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m512i( sc->H, 6 ) = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m512i( sc->H, 7 ) = m512_const1_64( 0x5BE0CD19137E2179 );
+   casti_m512i( sc->H, 0 ) = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   casti_m512i( sc->H, 1 ) = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   casti_m512i( sc->H, 2 ) = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   casti_m512i( sc->H, 3 ) = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   casti_m512i( sc->H, 4 ) = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   casti_m512i( sc->H, 5 ) = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   casti_m512i( sc->H, 6 ) = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   casti_m512i( sc->H, 7 ) = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

   sc->T0 = sc->T1 = 0;
   sc->ptr = 0;
@@ -859,7 +815,7 @@ blake64_8way_close( blake_8way_big_context *sc, void *dst )

   ptr = sc->ptr;
   bit_len = ((unsigned)ptr << 3);
-   buf[ptr>>3] = m512_const1_64( 0x80 );
+   buf[ptr>>3] = _mm512_set1_epi64( 0x80 );
   tl = sc->T0 + bit_len;
   th = sc->T1;
   if (ptr == 0 )
@@ -880,9 +836,9 @@ blake64_8way_close( blake_8way_big_context *sc, void *dst )
   {
       memset_zero_512( buf + (ptr>>3) + 1, (104-ptr) >> 3 );
       buf[104>>3] = _mm512_or_si512( buf[104>>3],
-                                 m512_const1_64( 0x0100000000000000ULL ) );
-       buf[112>>3] = m512_const1_64( bswap_64( th ) );
-       buf[120>>3] = m512_const1_64( bswap_64( tl ) );
+                                 _mm512_set1_epi64( 0x0100000000000000ULL ) );
+       buf[112>>3] = _mm512_set1_epi64( bswap_64( th ) );
+       buf[120>>3] = _mm512_set1_epi64( bswap_64( tl ) );

       blake64_8way( sc, buf + (ptr>>3), 128 - ptr );
   }
@@ -894,9 +850,9 @@ blake64_8way_close( blake_8way_big_context *sc, void *dst )
       sc->T0 = 0xFFFFFFFFFFFFFC00ULL;
       sc->T1 = 0xFFFFFFFFFFFFFFFFULL;
       memset_zero_512( buf, 112>>3 );
-       buf[104>>3] = m512_const1_64( 0x0100000000000000ULL );
-       buf[112>>3] = m512_const1_64( bswap_64( th ) );
-       buf[120>>3] = m512_const1_64( bswap_64( tl ) );
+       buf[104>>3] = _mm512_set1_epi64( 0x0100000000000000ULL );
+       buf[112>>3] = _mm512_set1_epi64( bswap_64( th ) );
+       buf[120>>3] = _mm512_set1_epi64( bswap_64( tl ) );

       blake64_8way( sc, buf, 128 );
   }
@@ -910,14 +866,14 @@ void blake512_8way_full( blake_8way_big_context *sc, void * dst,
   
 // init

-   casti_m512i( sc->H, 0 ) = m512_const1_64( 0x6A09E667F3BCC908 );
-   casti_m512i( sc->H, 1 ) = m512_const1_64( 0xBB67AE8584CAA73B );
-   casti_m512i( sc->H, 2 ) = m512_const1_64( 0x3C6EF372FE94F82B );
-   casti_m512i( sc->H, 3 ) = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m512i( sc->H, 4 ) = m512_const1_64( 0x510E527FADE682D1 );
-   casti_m512i( sc->H, 5 ) = m512_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m512i( sc->H, 6 ) = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m512i( sc->H, 7 ) = m512_const1_64( 0x5BE0CD19137E2179 );
+   casti_m512i( sc->H, 0 ) = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   casti_m512i( sc->H, 1 ) = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   casti_m512i( sc->H, 2 ) = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   casti_m512i( sc->H, 3 ) = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   casti_m512i( sc->H, 4 ) = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   casti_m512i( sc->H, 5 ) = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   casti_m512i( sc->H, 6 ) = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   casti_m512i( sc->H, 7 ) = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

   sc->T0 = sc->T1 = 0;
   sc->ptr = 0;
@@ -941,7 +897,7 @@ void blake512_8way_full( blake_8way_big_context *sc, void * dst,
   uint64_t th, tl;

   bit_len = sc->ptr << 3;
-   sc->buf[ptr64] = m512_const1_64( 0x80 );
+   sc->buf[ptr64] = _mm512_set1_epi64( 0x80 );
   tl = sc->T0 + bit_len;
   th = sc->T1;

@@ -959,9 +915,9 @@ void blake512_8way_full( blake_8way_big_context *sc, void * dst,
      sc->T0 -= 1024 - bit_len;

   memset_zero_512( sc->buf + ptr64 + 1, 13 - ptr64 );
-   sc->buf[13] = m512_const1_64( 0x0100000000000000ULL );
-   sc->buf[14] = m512_const1_64( bswap_64( th ) );
-   sc->buf[15] = m512_const1_64( bswap_64( tl ) );
+   sc->buf[13] = _mm512_set1_epi64( 0x0100000000000000ULL );
+   sc->buf[14] = _mm512_set1_epi64( bswap_64( th ) );
+   sc->buf[15] = _mm512_set1_epi64( bswap_64( tl ) );

   if ( ( sc->T0 = sc->T0 + 1024 ) < 1024 )
       sc->T1 = sc->T1 + 1;
@@ -977,14 +933,14 @@ void blake512_8way_full_le( blake_8way_big_context *sc, void * dst,

 // init

-   casti_m512i( sc->H, 0 ) = m512_const1_64( 0x6A09E667F3BCC908 );
-   casti_m512i( sc->H, 1 ) = m512_const1_64( 0xBB67AE8584CAA73B );
-   casti_m512i( sc->H, 2 ) = m512_const1_64( 0x3C6EF372FE94F82B );
-   casti_m512i( sc->H, 3 ) = m512_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m512i( sc->H, 4 ) = m512_const1_64( 0x510E527FADE682D1 );
-   casti_m512i( sc->H, 5 ) = m512_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m512i( sc->H, 6 ) = m512_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m512i( sc->H, 7 ) = m512_const1_64( 0x5BE0CD19137E2179 );
+   casti_m512i( sc->H, 0 ) = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
+   casti_m512i( sc->H, 1 ) = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
+   casti_m512i( sc->H, 2 ) = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
+   casti_m512i( sc->H, 3 ) = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
+   casti_m512i( sc->H, 4 ) = _mm512_set1_epi64( 0x510E527FADE682D1 );
+   casti_m512i( sc->H, 5 ) = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
+   casti_m512i( sc->H, 6 ) = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
+   casti_m512i( sc->H, 7 ) = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

   sc->T0 = sc->T1 = 0;
   sc->ptr = 0;
@@ -1008,7 +964,7 @@ void blake512_8way_full_le( blake_8way_big_context *sc, void * dst,
   uint64_t th, tl;

   bit_len = sc->ptr << 3;
-   sc->buf[ptr64] = m512_const1_64( 0x8000000000000000ULL );
+   sc->buf[ptr64] = _mm512_set1_epi64( 0x8000000000000000ULL );
   tl = sc->T0 + bit_len;
   th = sc->T1;

@@ -1027,8 +983,8 @@ void blake512_8way_full_le( blake_8way_big_context *sc, void * dst,

   memset_zero_512( sc->buf + ptr64 + 1, 13 - ptr64 );
   sc->buf[13] = m512_one_64;
-   sc->buf[14] = m512_const1_64( th );
-   sc->buf[15] = m512_const1_64( tl );
+   sc->buf[14] = _mm512_set1_epi64( th );
+   sc->buf[15] = _mm512_set1_epi64( tl );

   if ( ( sc->T0 = sc->T0 + 1024 ) < 1024 )
       sc->T1 = sc->T1 + 1;
@@ -1054,20 +1010,22 @@ blake512_8way_close(void *cc, void *dst)

 // Blake-512 4 way

-#define GB_4WAY(m0, m1, c0, c1, a, b, c, d)   do { \
+#define GB_4WAY(m0, m1, c0, c1, a, b, c, d) \
+{ \
   a = _mm256_add_epi64( _mm256_add_epi64( _mm256_xor_si256( \
                 _mm256_set1_epi64x( c1 ), m0 ), b ), a ); \
-   d = mm256_ror_64( _mm256_xor_si256( d, a ), 32 ); \
+   d = mm256_swap64_32( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi64( c, d ); \
   b = mm256_ror_64( _mm256_xor_si256( b, c ), 25 ); \
   a = _mm256_add_epi64( _mm256_add_epi64( _mm256_xor_si256( \
                 _mm256_set1_epi64x( c0 ), m1 ), b ), a ); \
-   d = mm256_ror_64( _mm256_xor_si256( d, a ), 16 ); \
+   d = mm256_shuflr64_16( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi64( c, d ); \
   b = mm256_ror_64( _mm256_xor_si256( b, c ), 11 ); \
-} while (0)
+}

-#define ROUND_B_4WAY(r)   do { \
+#define ROUND_B_4WAY(r) \
+{ \
 	GB_4WAY(Mx(r, 0), Mx(r, 1), CBx(r, 0), CBx(r, 1), V0, V4, V8, VC); \
 	GB_4WAY(Mx(r, 2), Mx(r, 3), CBx(r, 2), CBx(r, 3), V1, V5, V9, VD); \
 	GB_4WAY(Mx(r, 4), Mx(r, 5), CBx(r, 4), CBx(r, 5), V2, V6, VA, VE); \
@@ -1076,19 +1034,18 @@ blake512_8way_close(void *cc, void *dst)
 	GB_4WAY(Mx(r, A), Mx(r, B), CBx(r, A), CBx(r, B), V1, V6, VB, VC); \
 	GB_4WAY(Mx(r, C), Mx(r, D), CBx(r, C), CBx(r, D), V2, V7, V8, VD); \
 	GB_4WAY(Mx(r, E), Mx(r, F), CBx(r, E), CBx(r, F), V3, V4, V9, VE); \
-	} while (0)
+}

 #define DECL_STATE64_4WAY \
 	__m256i H0, H1, H2, H3, H4, H5, H6, H7; \
 	uint64_t T0, T1;

-#define COMPRESS64_4WAY   do \
+#define COMPRESS64_4WAY \
 { \
  __m256i M0, M1, M2, M3, M4, M5, M6, M7; \
  __m256i M8, M9, MA, MB, MC, MD, ME, MF; \
  __m256i V0, V1, V2, V3, V4, V5, V6, V7; \
  __m256i V8, V9, VA, VB, VC, VD, VE, VF; \
-  __m256i shuf_bswap64; \
  V0 = H0; \
  V1 = H1; \
  V2 = H2; \
@@ -1097,16 +1054,16 @@ blake512_8way_close(void *cc, void *dst)
  V5 = H5; \
  V6 = H6; \
  V7 = H7; \
-  V8 = m256_const1_64( CB0 );  \
-  V9 = m256_const1_64( CB1 );  \
-  VA = m256_const1_64( CB2 );  \
-  VB = m256_const1_64( CB3 );  \
+  V8 = _mm256_set1_epi64x( CB0 );  \
+  V9 = _mm256_set1_epi64x( CB1 );  \
+  VA = _mm256_set1_epi64x( CB2 );  \
+  VB = _mm256_set1_epi64x( CB3 );  \
  VC = _mm256_set1_epi64x( T0 ^ CB4 ); \
  VD = _mm256_set1_epi64x( T0 ^ CB5 ); \
  VE = _mm256_set1_epi64x( T1 ^ CB6 ); \
  VF = _mm256_set1_epi64x( T1 ^ CB7 ); \
-  shuf_bswap64 = m256_const_64( 0x18191a1b1c1d1e1f, 0x1011121314151617, \
-                                0x08090a0b0c0d0e0f, 0x0001020304050607 ); \
+  const __m256i shuf_bswap64 = mm256_bcast_m128( _mm_set_epi64x( \
+                             0x08090a0b0c0d0e0f, 0x0001020304050607 ) ); \
  M0 = _mm256_shuffle_epi8( *(buf+ 0), shuf_bswap64 ); \
  M1 = _mm256_shuffle_epi8( *(buf+ 1), shuf_bswap64 ); \
  M2 = _mm256_shuffle_epi8( *(buf+ 2), shuf_bswap64 ); \
@@ -1147,7 +1104,7 @@ blake512_8way_close(void *cc, void *dst)
  H5 = mm256_xor3( VD, V5, H5 ); \
  H6 = mm256_xor3( VE, V6, H6 ); \
  H7 = mm256_xor3( VF, V7, H7 ); \
-} while (0)
+}


 void blake512_4way_compress( blake_4way_big_context *sc )
@@ -1156,7 +1113,6 @@ void blake512_4way_compress( blake_4way_big_context *sc )
  __m256i M8, M9, MA, MB, MC, MD, ME, MF;
  __m256i V0, V1, V2, V3, V4, V5, V6, V7;
  __m256i V8, V9, VA, VB, VC, VD, VE, VF;
-  __m256i shuf_bswap64;

  V0 = sc->H[0];
  V1 = sc->H[1];
@@ -1166,20 +1122,20 @@ void blake512_4way_compress( blake_4way_big_context *sc )
  V5 = sc->H[5];
  V6 = sc->H[6];
  V7 = sc->H[7];
-  V8 = m256_const1_64( CB0 );
-  V9 = m256_const1_64( CB1 );
-  VA = m256_const1_64( CB2 );
-  VB = m256_const1_64( CB3 );
+  V8 = _mm256_set1_epi64x( CB0 );
+  V9 = _mm256_set1_epi64x( CB1 );
+  VA = _mm256_set1_epi64x( CB2 );
+  VB = _mm256_set1_epi64x( CB3 );
  VC = _mm256_xor_si256( _mm256_set1_epi64x( sc->T0 ),
-                             m256_const1_64( CB4 ) );
+                             _mm256_set1_epi64x( CB4 ) );
  VD = _mm256_xor_si256( _mm256_set1_epi64x( sc->T0 ),
-                             m256_const1_64( CB5 ) );
+                             _mm256_set1_epi64x( CB5 ) );
  VE = _mm256_xor_si256( _mm256_set1_epi64x( sc->T1 ),
-                             m256_const1_64( CB6 ) );
+                             _mm256_set1_epi64x( CB6 ) );
  VF = _mm256_xor_si256( _mm256_set1_epi64x( sc->T1 ),
-                             m256_const1_64( CB7 ) );
-  shuf_bswap64 = m256_const_64( 0x18191a1b1c1d1e1f, 0x1011121314151617,
-                                0x08090a0b0c0d0e0f, 0x0001020304050607 );
+                             _mm256_set1_epi64x( CB7 ) );
+  const __m256i shuf_bswap64 = mm256_bcast_m128( _mm_set_epi64x(
+                                    0x08090a0b0c0d0e0f, 0x0001020304050607 ) );

  M0 = _mm256_shuffle_epi8( sc->buf[ 0], shuf_bswap64 );
  M1 = _mm256_shuffle_epi8( sc->buf[ 1], shuf_bswap64 );
@@ -1232,23 +1188,23 @@ void blake512_4way_prehash_le( blake_4way_big_context *sc, __m256i *midstate,
   __m256i V8, V9, VA, VB, VC, VD, VE, VF;

   // initial hash
-   casti_m256i( sc->H, 0 ) = m256_const1_64( 0x6A09E667F3BCC908 );
-   casti_m256i( sc->H, 1 ) = m256_const1_64( 0xBB67AE8584CAA73B );
-   casti_m256i( sc->H, 2 ) = m256_const1_64( 0x3C6EF372FE94F82B );
-   casti_m256i( sc->H, 3 ) = m256_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m256i( sc->H, 4 ) = m256_const1_64( 0x510E527FADE682D1 );
-   casti_m256i( sc->H, 5 ) = m256_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m256i( sc->H, 6 ) = m256_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m256i( sc->H, 7 ) = m256_const1_64( 0x5BE0CD19137E2179 );
+   casti_m256i( sc->H, 0 ) = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
+   casti_m256i( sc->H, 1 ) = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
+   casti_m256i( sc->H, 2 ) = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
+   casti_m256i( sc->H, 3 ) = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
+   casti_m256i( sc->H, 4 ) = _mm256_set1_epi64x( 0x510E527FADE682D1 );
+   casti_m256i( sc->H, 5 ) = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
+   casti_m256i( sc->H, 6 ) = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
+   casti_m256i( sc->H, 7 ) = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );
   
   // fill buffer
   memcpy_256( sc->buf, (__m256i*)data, 80>>3 );
-   sc->buf[10] = m256_const1_64( 0x8000000000000000ULL );
+   sc->buf[10] = _mm256_set1_epi64x( 0x8000000000000000ULL );
   sc->buf[11] = m256_zero;
   sc->buf[12] = m256_zero;
   sc->buf[13] = m256_one_64;
   sc->buf[14] = m256_zero;
-   sc->buf[15] = m256_const1_64( 80*8 );
+   sc->buf[15] = _mm256_set1_epi64x( 80*8 );

   // build working variables
   V0 = sc->H[0];
@@ -1259,10 +1215,10 @@ void blake512_4way_prehash_le( blake_4way_big_context *sc, __m256i *midstate,
   V5 = sc->H[5];
   V6 = sc->H[6];
   V7 = sc->H[7];
-   V8 = m256_const1_64( CB0 );
-   V9 = m256_const1_64( CB1 );
-   VA = m256_const1_64( CB2 );
-   VB = m256_const1_64( CB3 );
+   V8 = _mm256_set1_epi64x( CB0 );
+   V9 = _mm256_set1_epi64x( CB1 );
+   VA = _mm256_set1_epi64x( CB2 );
+   VB = _mm256_set1_epi64x( CB3 );
   VC = _mm256_set1_epi64x( CB4 ^ 0x280ULL );
   VD = _mm256_set1_epi64x( CB5 ^ 0x280ULL );
   VE = _mm256_set1_epi64x( CB6 );
@@ -1277,7 +1233,7 @@ void blake512_4way_prehash_le( blake_4way_big_context *sc, __m256i *midstate,
   // G4 skip nonce
   V0 = _mm256_add_epi64( _mm256_add_epi64( _mm256_xor_si256(
                       _mm256_set1_epi64x( CB9 ), sc->buf[ 8] ), V5 ), V0 );
-   VF = mm256_ror_64( _mm256_xor_si256( VF, V0 ), 32 );
+   VF = mm256_swap64_32( _mm256_xor_si256( VF, V0 ) );
   VA = _mm256_add_epi64( VA, VF );
   V5 = mm256_ror_64( _mm256_xor_si256( V5, VA ), 25 );
   V0 = _mm256_add_epi64( V0, V5 );
@@ -1364,7 +1320,7 @@ void blake512_4way_final_le( blake_4way_big_context *sc, void *hash,
   // finish round 0, with the nonce now available 
   V0 = _mm256_add_epi64( V0, _mm256_xor_si256(
                                       _mm256_set1_epi64x( CB8 ), M9 ) );
-   VF = mm256_ror_64( _mm256_xor_si256( VF, V0 ), 16 );
+   VF = mm256_shuflr64_16( _mm256_xor_si256( VF, V0 ) );
   VA = _mm256_add_epi64( VA, VF );
   V5 = mm256_ror_64( _mm256_xor_si256( V5, VA ), 11 );

@@ -1374,34 +1330,34 @@ void blake512_4way_final_le( blake_4way_big_context *sc, void *hash,

   // G1
   V1 = _mm256_add_epi64( V1, V5 );
-   VD = mm256_ror_64( _mm256_xor_si256( VD, V1 ), 32 );
+   VD = mm256_swap64_32( _mm256_xor_si256( VD, V1 ) );
   V9 = _mm256_add_epi64( V9, VD );
   V5 = mm256_ror_64( _mm256_xor_si256( V5, V9 ), 25 );
   V1 = _mm256_add_epi64( V1, _mm256_add_epi64( _mm256_xor_si256(
                 _mm256_set1_epi64x( CBx(1,2) ), Mx(1,3) ), V5 ) );
-   VD = mm256_ror_64( _mm256_xor_si256( VD, V1 ), 16 );
+   VD = mm256_shuflr64_16( _mm256_xor_si256( VD, V1 ) );
   V9 = _mm256_add_epi64( V9, VD );
   V5 = mm256_ror_64( _mm256_xor_si256( V5, V9 ), 11 );

   // G2
   V2 = _mm256_add_epi64( V2, _mm256_xor_si256(
                 _mm256_set1_epi64x( CBF ), M9 ) );
-   VE = mm256_ror_64( _mm256_xor_si256( VE, V2 ), 32 );
+   VE = mm256_swap64_32( _mm256_xor_si256( VE, V2 ) );
   VA = _mm256_add_epi64( VA, VE );
   V6 = mm256_ror_64( _mm256_xor_si256( V6, VA ), 25 );
   V2 = _mm256_add_epi64( V2, _mm256_add_epi64( _mm256_xor_si256(
                 _mm256_set1_epi64x( CB9 ), MF ), V6 ) );
-   VE = mm256_ror_64( _mm256_xor_si256( VE, V2 ), 16 );
+   VE = mm256_shuflr64_16( _mm256_xor_si256( VE, V2 ) );
   VA = _mm256_add_epi64( VA, VE );
   V6 = mm256_ror_64( _mm256_xor_si256( V6, VA ), 11 );

   // G3
-   VF = mm256_ror_64( _mm256_xor_si256( VF, V3 ), 32 );
+   VF = mm256_swap64_32( _mm256_xor_si256( VF, V3 ) );
   VB = _mm256_add_epi64( VB, VF );
   V7 = mm256_ror_64( _mm256_xor_si256( V7, VB ), 25 );
   V3 = _mm256_add_epi64( V3, _mm256_add_epi64( _mm256_xor_si256(
                 _mm256_set1_epi64x( CBx(1, 6) ), Mx(1, 7) ), V7 ) );
-   VF = mm256_ror_64( _mm256_xor_si256( VF, V3 ), 16 );
+   VF = mm256_shuflr64_16( _mm256_xor_si256( VF, V3 ) );
   VB = _mm256_add_epi64( VB, VF );
   V7 = mm256_ror_64( _mm256_xor_si256( V7, VB ), 11 );

@@ -1442,14 +1398,14 @@ void blake512_4way_final_le( blake_4way_big_context *sc, void *hash,

 void blake512_4way_init( blake_4way_big_context *sc )
 {
-   casti_m256i( sc->H, 0 ) = m256_const1_64( 0x6A09E667F3BCC908 );
-   casti_m256i( sc->H, 1 ) = m256_const1_64( 0xBB67AE8584CAA73B );
-   casti_m256i( sc->H, 2 ) = m256_const1_64( 0x3C6EF372FE94F82B );
-   casti_m256i( sc->H, 3 ) = m256_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m256i( sc->H, 4 ) = m256_const1_64( 0x510E527FADE682D1 );
-   casti_m256i( sc->H, 5 ) = m256_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m256i( sc->H, 6 ) = m256_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m256i( sc->H, 7 ) = m256_const1_64( 0x5BE0CD19137E2179 );
+   casti_m256i( sc->H, 0 ) = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
+   casti_m256i( sc->H, 1 ) = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
+   casti_m256i( sc->H, 2 ) = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
+   casti_m256i( sc->H, 3 ) = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
+   casti_m256i( sc->H, 4 ) = _mm256_set1_epi64x( 0x510E527FADE682D1 );
+   casti_m256i( sc->H, 5 ) = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
+   casti_m256i( sc->H, 6 ) = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
+   casti_m256i( sc->H, 7 ) = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

   sc->T0 = sc->T1 = 0;
   sc->ptr = 0;
@@ -1490,7 +1446,7 @@ blake64_4way( blake_4way_big_context *sc, const void *data, size_t len)
 	   if ( ptr == buf_size )
      {
 		   if ( (T0 = T0 + 1024 ) < 1024 )
-			   T1 = SPH_T64(T1 + 1);
+			   T1 = T1 + 1;
 	   	COMPRESS64_4WAY;
 		   ptr = 0;
 	   }
@@ -1509,7 +1465,7 @@ blake64_4way_close( blake_4way_big_context *sc, void *dst )

   ptr = sc->ptr;
   bit_len = ((unsigned)ptr << 3);
-   buf[ptr>>3] = m256_const1_64( 0x80 );
+   buf[ptr>>3] = _mm256_set1_epi64x( 0x80 );
   tl = sc->T0 + bit_len;
   th = sc->T1;
   if (ptr == 0 )
@@ -1531,9 +1487,9 @@ blake64_4way_close( blake_4way_big_context *sc, void *dst )
   {
       memset_zero_256( buf + (ptr>>3) + 1, (104-ptr) >> 3 );
       buf[104>>3] = _mm256_or_si256( buf[104>>3],
-                                 m256_const1_64( 0x0100000000000000ULL ) );
-       buf[112>>3] = m256_const1_64( bswap_64( th ) );
-       buf[120>>3] = m256_const1_64( bswap_64( tl ) );
+                                 _mm256_set1_epi64x( 0x0100000000000000ULL ) );
+       buf[112>>3] = _mm256_set1_epi64x( bswap_64( th ) );
+       buf[120>>3] = _mm256_set1_epi64x( bswap_64( tl ) );

       blake64_4way( sc, buf + (ptr>>3), 128 - ptr );
   }
@@ -1542,12 +1498,12 @@ blake64_4way_close( blake_4way_big_context *sc, void *dst )
       memset_zero_256( buf + (ptr>>3) + 1, (120 - ptr) >> 3 );

       blake64_4way( sc, buf + (ptr>>3), 128 - ptr );
-       sc->T0 = SPH_C64(0xFFFFFFFFFFFFFC00ULL);
-       sc->T1 = SPH_C64(0xFFFFFFFFFFFFFFFFULL);
+       sc->T0 = 0xFFFFFFFFFFFFFC00ULL;
+       sc->T1 = 0xFFFFFFFFFFFFFFFFULL;
       memset_zero_256( buf, 112>>3 ); 
-       buf[104>>3] = m256_const1_64( 0x0100000000000000ULL );
-       buf[112>>3] = m256_const1_64( bswap_64( th ) );
-       buf[120>>3] = m256_const1_64( bswap_64( tl ) );
+       buf[104>>3] = _mm256_set1_epi64x( 0x0100000000000000ULL );
+       buf[112>>3] = _mm256_set1_epi64x( bswap_64( th ) );
+       buf[120>>3] = _mm256_set1_epi64x( bswap_64( tl ) );

       blake64_4way( sc, buf, 128 );
   }
@@ -1561,14 +1517,14 @@ void blake512_4way_full( blake_4way_big_context *sc, void * dst,

 // init

-   casti_m256i( sc->H, 0 ) = m256_const1_64( 0x6A09E667F3BCC908 );
-   casti_m256i( sc->H, 1 ) = m256_const1_64( 0xBB67AE8584CAA73B );
-   casti_m256i( sc->H, 2 ) = m256_const1_64( 0x3C6EF372FE94F82B );
-   casti_m256i( sc->H, 3 ) = m256_const1_64( 0xA54FF53A5F1D36F1 );
-   casti_m256i( sc->H, 4 ) = m256_const1_64( 0x510E527FADE682D1 );
-   casti_m256i( sc->H, 5 ) = m256_const1_64( 0x9B05688C2B3E6C1F );
-   casti_m256i( sc->H, 6 ) = m256_const1_64( 0x1F83D9ABFB41BD6B );
-   casti_m256i( sc->H, 7 ) = m256_const1_64( 0x5BE0CD19137E2179 );
+   casti_m256i( sc->H, 0 ) = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
+   casti_m256i( sc->H, 1 ) = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
+   casti_m256i( sc->H, 2 ) = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
+   casti_m256i( sc->H, 3 ) = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
+   casti_m256i( sc->H, 4 ) = _mm256_set1_epi64x( 0x510E527FADE682D1 );
+   casti_m256i( sc->H, 5 ) = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
+   casti_m256i( sc->H, 6 ) = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
+   casti_m256i( sc->H, 7 ) = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

   sc->T0 = sc->T1 = 0;
   sc->ptr = 0;
@@ -1592,7 +1548,7 @@ void blake512_4way_full( blake_4way_big_context *sc, void * dst,
   uint64_t th, tl;

   bit_len = sc->ptr << 3;
-   sc->buf[ptr64] = m256_const1_64( 0x80 );
+   sc->buf[ptr64] = _mm256_set1_epi64x( 0x80 );
   tl = sc->T0 + bit_len;
   th = sc->T1;
   if ( sc->ptr == 0 )
@@ -1609,9 +1565,9 @@ void blake512_4way_full( blake_4way_big_context *sc, void * dst,
        sc->T0 -= 1024 - bit_len;

   memset_zero_256( sc->buf + ptr64 + 1, 13 - ptr64 );
-   sc->buf[13] = m256_const1_64( 0x0100000000000000ULL );
-   sc->buf[14] = m256_const1_64( bswap_64( th ) );
-   sc->buf[15] = m256_const1_64( bswap_64( tl ) );
+   sc->buf[13] = _mm256_set1_epi64x( 0x0100000000000000ULL );
+   sc->buf[14] = _mm256_set1_epi64x( bswap_64( th ) );
+   sc->buf[15] = _mm256_set1_epi64x( bswap_64( tl ) );

   if ( ( sc->T0 = sc->T0 + 1024 ) < 1024 )
       sc->T1 = sc->T1 + 1;
@@ -1633,8 +1589,4 @@ blake512_4way_close(void *cc, void *dst)
   blake64_4way_close( cc, dst );
 }

-#ifdef __cplusplus
-}
-#endif
-
 #endif
--- a/algo/blake/blakecoin-4way.c
+++ b/algo/blake/blakecoin-4way.c
@@ -4,7 +4,149 @@
 #include <stdint.h>
 #include <memory.h>

-#if defined (BLAKECOIN_4WAY)
+#define rounds 8
+
+#if defined (BLAKECOIN_16WAY)
+
+int scanhash_blakecoin_16way( struct work *work, uint32_t max_nonce,
+                         uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t hash32[8*16] __attribute__ ((aligned (64)));
+   uint32_t midstate_vars[16*16] __attribute__ ((aligned (64)));
+   __m512i block0_hash[8] __attribute__ ((aligned (64)));
+   __m512i block_buf[16] __attribute__ ((aligned (64)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 =  (uint32_t*)&( ((__m512i*)hash32)[7] );
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t targ32_d7 = ptarget[7];
+   uint32_t phash[8] __attribute__ ((aligned (64))) =
+   {
+      0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
+      0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
+   };
+   uint32_t n = pdata[19];
+   const uint32_t first_nonce = (const uint32_t) n;
+   const uint32_t last_nonce = max_nonce - 16;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+   const __m512i sixteen = _mm512_set1_epi32( 16 );
+
+   // Prehash first block
+   blake256_transform_le( phash, pdata, 512, 0, rounds );
+
+   block0_hash[0] = _mm512_set1_epi32( phash[0] );
+   block0_hash[1] = _mm512_set1_epi32( phash[1] );
+   block0_hash[2] = _mm512_set1_epi32( phash[2] );
+   block0_hash[3] = _mm512_set1_epi32( phash[3] );
+   block0_hash[4] = _mm512_set1_epi32( phash[4] );
+   block0_hash[5] = _mm512_set1_epi32( phash[5] );
+   block0_hash[6] = _mm512_set1_epi32( phash[6] );
+   block0_hash[7] = _mm512_set1_epi32( phash[7] );
+
+   // Build vectored second block, interleave last 16 bytes of data using
+   // unique nonces.
+   block_buf[0] = _mm512_set1_epi32( pdata[16] );
+   block_buf[1] = _mm512_set1_epi32( pdata[17] );
+   block_buf[2] = _mm512_set1_epi32( pdata[18] );
+   block_buf[3] =
+             _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 );
+
+   // Partialy prehash second block without touching nonces in block_buf[3].
+   blake256_16way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
+
+   do {
+      blake256_16way_final_rounds_le( hash32, midstate_vars, block0_hash,
+                                      block_buf, rounds );
+      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] = n + lane;
+            submit_solution( work, lane_hash, mythr );
+         }
+      }
+      block_buf[3] = _mm512_add_epi32( block_buf[3], sixteen );
+      n += 16;
+   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart) );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+
+#elif defined (BLAKECOIN_8WAY)
+
+int scanhash_blakecoin_8way( struct work *work, uint32_t max_nonce,
+                         uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t hash32[8*8] __attribute__ ((aligned (64)));
+   uint32_t midstate_vars[16*8] __attribute__ ((aligned (32)));
+   __m256i block0_hash[8] __attribute__ ((aligned (32)));
+   __m256i block_buf[16] __attribute__ ((aligned (32)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 =  (uint32_t*)&( ((__m256i*)hash32)[7] );
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t targ32_d7 = ptarget[7];
+   uint32_t phash[8] __attribute__ ((aligned (32))) =
+   {
+      0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
+      0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
+   };
+   uint32_t n = pdata[19];
+   const uint32_t first_nonce = (const uint32_t) n;
+   const uint32_t last_nonce = max_nonce - 8;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+   const __m256i eight = _mm256_set1_epi32( 8 );
+
+   // Prehash first block
+   blake256_transform_le( phash, pdata, 512, 0, rounds );
+
+   block0_hash[0] = _mm256_set1_epi32( phash[0] );
+   block0_hash[1] = _mm256_set1_epi32( phash[1] );
+   block0_hash[2] = _mm256_set1_epi32( phash[2] );
+   block0_hash[3] = _mm256_set1_epi32( phash[3] );
+   block0_hash[4] = _mm256_set1_epi32( phash[4] );
+   block0_hash[5] = _mm256_set1_epi32( phash[5] );
+   block0_hash[6] = _mm256_set1_epi32( phash[6] );
+   block0_hash[7] = _mm256_set1_epi32( phash[7] );
+
+   // Build vectored second block, interleave last 16 bytes of data using
+   // unique nonces.
+   block_buf[0] = _mm256_set1_epi32( pdata[16] );
+   block_buf[1] = _mm256_set1_epi32( pdata[17] );
+   block_buf[2] = _mm256_set1_epi32( pdata[18] );
+   block_buf[3] = _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n );
+
+   // Partialy prehash second block without touching nonces in block_buf[3].
+   blake256_8way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
+
+   do {
+      blake256_8way_final_rounds_le( hash32, midstate_vars, block0_hash,
+                                     block_buf, rounds );
+      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] = n + lane;
+            submit_solution( work, lane_hash, mythr );
+         }
+      }
+      block_buf[3] = _mm256_add_epi32( block_buf[3], eight );
+      n += 8;
+   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart) );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+   
+#elif defined (BLAKECOIN_4WAY)

 blake256r8_4way_context blakecoin_4w_ctx;

@@ -61,7 +203,8 @@ int scanhash_blakecoin_4way( struct work *work, uint32_t max_nonce,

 #endif

-#if defined(BLAKECOIN_8WAY)
+#if 0
+//#if defined(BLAKECOIN_8WAY)

 blake256r8_8way_context blakecoin_8w_ctx;

@@ -78,11 +221,84 @@ void blakecoin_8way_hash( void *state, const void *input )
                   state+160, state+192, state+224, vhash, 256 );
 }

+/*
+int scanhash_blakecoin_8way( struct work *work, uint32_t max_nonce,
+                             uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t hash32[8*8] __attribute__ ((aligned (64)));
+   uint32_t midstate_vars[16*8] __attribute__ ((aligned (64)));
+   __m256i block0_hash[8] __attribute__ ((aligned (64)));
+   __m256i block_buf[16] __attribute__ ((aligned (64)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 =  (uint32_t*)&( hash32[7] );
+   uint32_t phash[8] __attribute__ ((aligned (32))) =
+   {
+      0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
+      0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
+   };
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = (uint32_t*)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;
+   const int thr_id = mythr->id;
+   const bool bench = opt_benchmark;
+   const __m256i eight = _mm256_set1_epi32( 8 );
+
+   // Prehash first block
+   blake256_transform_le( phash, pdata, 512, 0, 8 );
+
+   block0_hash[0] = _mm256_set1_epi32( phash[0] );
+   block0_hash[1] = _mm256_set1_epi32( phash[1] );
+   block0_hash[2] = _mm256_set1_epi32( phash[2] );
+   block0_hash[3] = _mm256_set1_epi32( phash[3] );
+   block0_hash[4] = _mm256_set1_epi32( phash[4] );
+   block0_hash[5] = _mm256_set1_epi32( phash[5] );
+   block0_hash[6] = _mm256_set1_epi32( phash[6] );
+   block0_hash[7] = _mm256_set1_epi32( phash[7] );
+
+   // Build vectored second block, interleave last 16 bytes of data using
+   // unique nonces.
+   block_buf[0] = _mm256_set1_epi32( pdata[16] );
+   block_buf[1] = _mm256_set1_epi32( pdata[17] );
+   block_buf[2] = _mm256_set1_epi32( pdata[18] );
+   block_buf[3] = _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n );
+
+   // Partialy prehash second block without touching nonces
+   blake256_8way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
+
+   do {
+      blake256_8way_final_rounds_le( hash32, midstate_vars, block0_hash, 
+                                     block_buf );
+
+      for ( int lane = 0; lane < 8; lane++ )
+      if ( hash32_d7[ lane ] <= targ32_d7 )
+      {
+         extr_lane_8x32( lane_hash, hash32, lane, 256 );
+         if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
+         {
+            pdata[19] = n + lane;
+            submit_solution( work, lane_hash, mythr );
+         }
+      }
+      block_buf[3] = _mm256_add_epi32( block_buf[3], eight );
+      n += 8;
+   } while ( (n < last_nonce) && !work_restart[thr_id].restart );
+   pdata[19] = n;
+   *hashes_done = n - first_nonce;
+   return 0;
+}
+*/
+
 int scanhash_blakecoin_8way( struct work *work, 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)));
+   blake256r8_8way_context ctx __attribute__ ((aligned (32)));
+   uint32_t *hash32_d7 =  (uint32_t*)&( ((__m256i*)hash32)[7] );
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
   const uint32_t first_nonce = pdata[19];
@@ -101,15 +317,22 @@ int scanhash_blakecoin_8way( struct work *work, uint32_t max_nonce,
      *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;
-      blakecoin_8way_hash( hash, vdata );

-      for ( int i = 0; i < 8; i++ )
-      if (  (hash+(i<<3))[7] <= HTarget && fulltest( hash+(i<<3), ptarget )
-          && !opt_benchmark )
+      memcpy( &ctx, &blakecoin_8w_ctx, sizeof ctx );
+      blake256r8_8way_update( &ctx, (const void*)vdata + (64<<3), 16 );
+      blake256r8_8way_close( &ctx, hash32 );
+
+      for ( int lane = 0; lane < 8; lane++ )
+      if ( hash32_d7[ lane ] <= HTarget )
      {
-          pdata[19] = n+i;
-          submit_solution( work, hash+(i<<3), mythr );
+         extr_lane_8x32( lane_hash, hash32, lane, 256 );
+         if ( likely( valid_hash( lane_hash, ptarget ) && !opt_benchmark ) )
+         {
+            pdata[19] = n + lane;
+            submit_solution( work, lane_hash, mythr );
         }
+      }
+     
      n += 8;
   } while ( (n < max_nonce) && !work_restart[thr_id].restart );

--- a/algo/blake/blakecoin-gate.c
+++ b/algo/blake/blakecoin-gate.c
@@ -4,10 +4,10 @@
 // vanilla uses default gen merkle root, otherwise identical to blakecoin
 bool register_vanilla_algo( algo_gate_t* gate )
 {
-#if defined(BLAKECOIN_8WAY)
+#if defined(BLAKECOIN_16WAY)
+  gate->scanhash  = (void*)&scanhash_blakecoin_16way;
+#elif defined(BLAKECOIN_8WAY)
  gate->scanhash  = (void*)&scanhash_blakecoin_8way;
-  gate->hash      = (void*)&blakecoin_8way_hash;
-
 #elif defined(BLAKECOIN_4WAY)
  gate->scanhash  = (void*)&scanhash_blakecoin_4way;
  gate->hash      = (void*)&blakecoin_4way_hash;
@@ -15,14 +15,14 @@ bool register_vanilla_algo( algo_gate_t* gate )
  gate->scanhash = (void*)&scanhash_blakecoin;
  gate->hash     = (void*)&blakecoinhash;
 #endif
-  gate->optimizations = SSE42_OPT | AVX2_OPT;
+  gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
  return true;
 }

 bool register_blakecoin_algo( algo_gate_t* gate )
 {
  register_vanilla_algo( gate );
-  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
+  gate->gen_merkle_root = (void*)&sha256_gen_merkle_root;
  return true;
 }

--- a/algo/blake/blakecoin-gate.h
+++ b/algo/blake/blakecoin-gate.h
@@ -1,30 +1,36 @@
-#ifndef __BLAKECOIN_GATE_H__
-#define __BLAKECOIN_GATE_H__ 1
+#ifndef BLAKECOIN_GATE_H__
+#define BLAKECOIN_GATE_H__ 1

 #include "algo-gate-api.h"
 #include <stdint.h>

-#if defined(__SSE4_2__)
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+  #define BLAKECOIN_16WAY
+#elif defined(__AVX2__)
+  #define BLAKECOIN_8WAY
+#elif defined(__SSE2__)  // always true
  #define BLAKECOIN_4WAY
 #endif
-#if defined(__AVX2__)
-  #define BLAKECOIN_8WAY
-#endif

-#if defined (BLAKECOIN_8WAY)
-void blakecoin_8way_hash(void *state, const void *input);
+#if defined (BLAKECOIN_16WAY)
+int scanhash_blakecoin_16way( struct work *work, uint32_t max_nonce,
+                         uint64_t *hashes_done, struct thr_info *mythr );
+
+#elif defined (BLAKECOIN_8WAY)
+//void blakecoin_8way_hash(void *state, const void *input);
 int scanhash_blakecoin_8way( struct work *work, uint32_t max_nonce,
                         uint64_t *hashes_done, struct thr_info *mythr );
-#endif

-#if defined (BLAKECOIN_4WAY)
+#elif defined (BLAKECOIN_4WAY)
 void blakecoin_4way_hash(void *state, const void *input);
 int scanhash_blakecoin_4way( struct work *work, uint32_t max_nonce,
                         uint64_t *hashes_done, struct thr_info *mythr );
-#endif
+#else  // never used

 void blakecoinhash( void *state, const void *input );
 int scanhash_blakecoin( struct work *work, uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr );

 #endif
+
+#endif
--- a/algo/blake/blakecoin.c
+++ b/algo/blake/blakecoin.c
@@ -1,6 +1,6 @@
 #include "blakecoin-gate.h"

-#if !defined(BLAKECOIN_8WAY) && !defined(BLAKECOIN_4WAY)
+#if !defined(BLAKECOIN_16WAY) && !defined(BLAKECOIN_8WAY) && !defined(BLAKECOIN_4WAY)

 #define BLAKE32_ROUNDS 8
 #include "sph_blake.h"
@@ -12,7 +12,6 @@ void blakecoin_close(void *cc, void *dst);
 #include <string.h>
 #include <stdint.h>
 #include <memory.h>
-#include <openssl/sha.h>

 // context management is staged for efficiency.
 // 1. global initial ctx cached on startup
--- a/algo/blake/decred-4way.c
+++ b/algo/blake/decred-4way.c
@@ -1,74 +0,0 @@
-#include "decred-gate.h"
-#include "blake-hash-4way.h"
-#include <string.h>
-#include <stdint.h>
-#include <memory.h>
-#include <unistd.h>
-
-#if defined (DECRED_4WAY)
-
-static __thread blake256_4way_context blake_mid;
-
-void decred_hash_4way( void *state, const void *input )
-{
-     uint32_t vhash[8*4] __attribute__ ((aligned (64)));
-//     uint32_t hash0[8] __attribute__ ((aligned (32)));
-//     uint32_t hash1[8] __attribute__ ((aligned (32)));
-//     uint32_t hash2[8] __attribute__ ((aligned (32)));
-//     uint32_t hash3[8] __attribute__ ((aligned (32)));
-     const void *tail = input + ( DECRED_MIDSTATE_LEN << 2 );
-     int tail_len = 180 - DECRED_MIDSTATE_LEN; 
-     blake256_4way_context ctx __attribute__ ((aligned (64)));
-
-     memcpy( &ctx, &blake_mid, sizeof(blake_mid) );
-     blake256_4way_update( &ctx, tail, tail_len );
-     blake256_4way_close( &ctx, vhash );
-     dintrlv_4x32( state, state+32, state+64, state+96, vhash, 256 );
-}
-
-int scanhash_decred_4way( struct work *work, uint32_t max_nonce,
-                          uint64_t *hashes_done, struct thr_info *mythr )
-{
-   uint32_t vdata[48*4] __attribute__ ((aligned (64)));
-   uint32_t hash[8*4] __attribute__ ((aligned (32)));
-   uint32_t _ALIGN(64) edata[48];
-   uint32_t *pdata = work->data;
-   uint32_t *ptarget = work->target;
-   const uint32_t first_nonce = pdata[DECRED_NONCE_INDEX];
-   uint32_t n = first_nonce;
-   const uint32_t HTarget = opt_benchmark ? 0x7f : ptarget[7];
-   int thr_id = mythr->id;  // thr_id arg is deprecated
-
-   // copy to buffer guaranteed to be aligned.
-   memcpy( edata, pdata, 180 );
-
-   // use the old way until  new way updated for size.
-   mm128_intrlv_4x32x( vdata, edata, edata, edata, edata, 180*8 );
-
-   blake256_4way_init( &blake_mid );
-   blake256_4way_update( &blake_mid, vdata, DECRED_MIDSTATE_LEN );
-
-   uint32_t *noncep = vdata + DECRED_NONCE_INDEX * 4;
-   do {
-      * noncep    = n;
-      *(noncep+1) = n+1;
-      *(noncep+2) = n+2;
-      *(noncep+3) = n+3;
-
-      decred_hash_4way( hash, vdata );
-
-      for ( int i = 0; i < 4; i++ )
-      if (  (hash+(i<<3))[7] <= HTarget )
-      if ( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
-      {
-          pdata[DECRED_NONCE_INDEX] = n+i;
-          submit_solution( work, hash+(i<<3), mythr );
-      }
-      n += 4;
-  } while ( (n < max_nonce) && !work_restart[thr_id].restart );
-
-  *hashes_done = n - first_nonce + 1;
-  return 0;
-}
-
-#endif
--- a/algo/blake/decred-gate.c
+++ b/algo/blake/decred-gate.c
@@ -1,171 +0,0 @@
-#include "decred-gate.h"
-#include <unistd.h>
-#include <memory.h>
-#include <string.h>
-
-uint32_t *decred_get_nonceptr( uint32_t *work_data )
-{
-   return &work_data[ DECRED_NONCE_INDEX ];
-}
-
-long double decred_calc_network_diff( struct work* work )
-{
-   // sample for diff 43.281 : 1c05ea29
-   // todo: endian reversed on longpoll could be zr5 specific...
-   uint32_t nbits = work->data[ DECRED_NBITS_INDEX ];
-   uint32_t bits = ( nbits & 0xffffff );
-   int16_t shift = ( swab32(nbits) & 0xff ); // 0x1c = 28
-   int m;
-   long double d = (long double)0x0000ffff / (long double)bits;
-
-   for ( m = shift; m < 29; m++ )
-       d *= 256.0;
-   for ( m = 29; m < shift; m++ )
-       d /= 256.0;
-   if ( shift == 28 )
-       d *= 256.0; // testnet
-   if ( opt_debug_diff )
-       applog( LOG_DEBUG, "net diff: %f -> shift %u, bits %08x", (double)d,
-                           shift, bits );
-   return net_diff;
-}
-
-void decred_decode_extradata( struct work* work, uint64_t* net_blocks )
-{
-   // some random extradata to make the work unique
-   work->data[ DECRED_XNONCE_INDEX ] = (rand()*4);
-   work->height = work->data[32];
-   if (!have_longpoll && work->height > *net_blocks + 1)
-   {
-      char netinfo[64] = { 0 };
-      if ( net_diff > 0. )
-      {
-         if (net_diff != work->targetdiff)
-            sprintf(netinfo, ", diff %.3f, target %.1f", net_diff,
-                   work->targetdiff);
-         else
-             sprintf(netinfo, ", diff %.3f", net_diff);
-       }
-       applog(LOG_BLUE, "%s block %d%s", algo_names[opt_algo], work->height,
-                       netinfo);
-       *net_blocks = work->height - 1;
-   }
-}
-
-void decred_be_build_stratum_request( char *req, struct work *work,
-                                      struct stratum_ctx *sctx )
-{
-   unsigned char *xnonce2str;
-   uint32_t ntime, nonce;
-   char ntimestr[9], noncestr[9];
-
-   be32enc( &ntime, work->data[ DECRED_NTIME_INDEX ] );
-   be32enc( &nonce, work->data[ DECRED_NONCE_INDEX ] );
-   bin2hex( ntimestr, (char*)(&ntime), sizeof(uint32_t) );
-   bin2hex( noncestr, (char*)(&nonce), sizeof(uint32_t) );
-   xnonce2str = abin2hex( (char*)( &work->data[ DECRED_XNONCE_INDEX ] ),
-                                     sctx->xnonce1_size );
-   snprintf( req, JSON_BUF_LEN,
-        "{\"method\": \"mining.submit\", \"params\": [\"%s\", \"%s\", \"%s\", \"%s\", \"%s\"], \"id\":4}",
-         rpc_user, work->job_id, xnonce2str, ntimestr, noncestr );
-   free(xnonce2str);
-}
-
-#if !defined(min)
-#define min(a,b) (a>b ? (b) :(a))
-#endif
-
-void decred_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
-{
-   uchar merkle_root[64] = { 0 };
-   uint32_t extraheader[32] = { 0 };
-   int headersize = 0;
-   uint32_t* extradata = (uint32_t*) sctx->xnonce1;
-   int i;
-
-   // getwork over stratum, getwork merkle + header passed in coinb1
-   memcpy(merkle_root, sctx->job.coinbase, 32);
-   headersize = min((int)sctx->job.coinbase_size - 32,
-                  sizeof(extraheader) );
-   memcpy( extraheader, &sctx->job.coinbase[32], headersize );
-
-   // Assemble block header 
-   memset( g_work->data, 0, sizeof(g_work->data) );
-   g_work->data[0] = le32dec( sctx->job.version );
-   for ( i = 0; i < 8; i++ )
-      g_work->data[1 + i] = swab32(
-                              le32dec( (uint32_t *) sctx->job.prevhash + i ) );
-   for ( i = 0; i < 8; i++ )
-      g_work->data[9 + i] = swab32( be32dec( (uint32_t *) merkle_root + i ) );
-
-//   for ( i = 0; i < 8; i++ ) // prevhash
-//      g_work->data[1 + i] = swab32( g_work->data[1 + i] );
-//   for ( i = 0; i < 8; i++ ) // merkle
-//      g_work->data[9 + i] = swab32( g_work->data[9 + i] );
-
-   for ( i = 0; i < headersize/4; i++ ) // header
-      g_work->data[17 + i] = extraheader[i];
-   // extradata
-
-   for ( i = 0; i < sctx->xnonce1_size/4; i++ )
-      g_work->data[ DECRED_XNONCE_INDEX + i ] = extradata[i];
-   for ( i = DECRED_XNONCE_INDEX + sctx->xnonce1_size/4; i < 45; i++ )
-      g_work->data[i] = 0;
-   g_work->data[37] = (rand()*4) << 8;
-   // block header suffix from coinb2 (stake version)
-   memcpy( &g_work->data[44],
-           &sctx->job.coinbase[ sctx->job.coinbase_size-4 ], 4 );
-   sctx->block_height = g_work->data[32];
-   //applog_hex(work->data, 180);
-   //applog_hex(&work->data[36], 36);
-}
-
-#undef min
-
-bool decred_ready_to_mine( struct work* work, struct stratum_ctx* stratum,
-                           int thr_id )
-{
-   if ( have_stratum && strcmp(stratum->job.job_id, work->job_id)  )
-      // need to regen g_work..
-      return false;
-   if ( have_stratum && !work->data[0] && !opt_benchmark )
-   {
-      sleep(1);
-      return false;
-   }
-   // extradata: prevent duplicates
-   work->data[ DECRED_XNONCE_INDEX     ] += 1;
-   work->data[ DECRED_XNONCE_INDEX + 1 ] |= thr_id;
-   return true;
-}
-
-int decred_get_work_data_size() { return DECRED_DATA_SIZE; }
-
-bool register_decred_algo( algo_gate_t* gate )
-{
-#if defined(DECRED_4WAY)
-  four_way_not_tested();
-  gate->scanhash  = (void*)&scanhash_decred_4way;
-  gate->hash      = (void*)&decred_hash_4way;
-#else
-  gate->scanhash  = (void*)&scanhash_decred;
-  gate->hash      = (void*)&decred_hash;
-#endif
-  gate->optimizations = AVX2_OPT;
-//  gate->get_nonceptr          = (void*)&decred_get_nonceptr;
-  gate->decode_extra_data     = (void*)&decred_decode_extradata;
-  gate->build_stratum_request = (void*)&decred_be_build_stratum_request;
-  gate->work_decode           = (void*)&std_be_work_decode;
-  gate->submit_getwork_result = (void*)&std_be_submit_getwork_result;
-  gate->build_extraheader     = (void*)&decred_build_extraheader;
-  gate->ready_to_mine         = (void*)&decred_ready_to_mine;
-  gate->nbits_index           = DECRED_NBITS_INDEX;
-  gate->ntime_index           = DECRED_NTIME_INDEX;
-  gate->nonce_index           = DECRED_NONCE_INDEX;
-  gate->get_work_data_size    = (void*)&decred_get_work_data_size;
-  gate->work_cmp_size         = DECRED_WORK_COMPARE_SIZE;
-  allow_mininginfo            = false;
-  have_gbt                    = false;
-  return true;
-}
-
--- a/algo/blake/decred-gate.h
+++ b/algo/blake/decred-gate.h
@@ -1,36 +0,0 @@
-#ifndef __DECRED_GATE_H__
-#define __DECRED_GATE_H__
-
-#include "algo-gate-api.h"
-#include <stdint.h>
-
-#define DECRED_NBITS_INDEX 29
-#define DECRED_NTIME_INDEX 34
-#define DECRED_NONCE_INDEX 35
-#define DECRED_XNONCE_INDEX 36
-#define DECRED_DATA_SIZE 192
-#define DECRED_WORK_COMPARE_SIZE 140
-#define DECRED_MIDSTATE_LEN 128
-
-#if defined (__AVX2__) 
-//void blakehash_84way(void *state, const void *input);
-//int scanhash_blake_8way( struct work *work, uint32_t max_nonce,
-//                         uint64_t *hashes_done );
-#endif
-
-#if defined(__SSE4_2__)
-  #define DECRED_4WAY
-#endif
-
-#if defined (DECRED_4WAY)
-void decred_hash_4way(void *state, const void *input);
-int scanhash_decred_4way( struct work *work, uint32_t max_nonce,
-                          uint64_t *hashes_done, struct thr_info *mythr );
-#endif
-
-void decred_hash( void *state, const void *input );
-int scanhash_decred( struct work *work, uint32_t max_nonce,
-                     uint64_t *hashes_done, struct thr_info *mythr );
-
-#endif
-
--- a/algo/blake/decred.c
+++ b/algo/blake/decred.c
@@ -1,282 +0,0 @@
-#include "decred-gate.h"
-
-#if !defined(DECRED_8WAY) && !defined(DECRED_4WAY)
-
-#include "sph_blake.h"
-
-#include <string.h>
-#include <stdint.h>
-#include <memory.h>
-#include <unistd.h>
-
-/*
-#ifndef min
-#define min(a,b) (a>b ? b : a)
-#endif
-#ifndef max 
-#define max(a,b) (a<b ? b : a)
-#endif
-*/
-/*
-#define DECRED_NBITS_INDEX 29
-#define DECRED_NTIME_INDEX 34
-#define DECRED_NONCE_INDEX 35
-#define DECRED_XNONCE_INDEX 36
-#define DECRED_DATA_SIZE 192
-#define DECRED_WORK_COMPARE_SIZE 140
-*/
-static __thread sph_blake256_context blake_mid;
-static __thread bool ctx_midstate_done = false;
-
-void decred_hash(void *state, const void *input)
-{
-//        #define MIDSTATE_LEN 128
-        sph_blake256_context ctx __attribute__ ((aligned (64)));
-
-        uint8_t *ending = (uint8_t*) input;
-        ending += DECRED_MIDSTATE_LEN;
-
-        if (!ctx_midstate_done) {
-                sph_blake256_init(&blake_mid);
-                sph_blake256(&blake_mid, input, DECRED_MIDSTATE_LEN);
-                ctx_midstate_done = true;
-        }
-        memcpy(&ctx, &blake_mid, sizeof(blake_mid));
-
-        sph_blake256(&ctx, ending, (180 - DECRED_MIDSTATE_LEN));
-        sph_blake256_close(&ctx, state);
-}
-
-void decred_hash_simple(void *state, const void *input)
-{
-        sph_blake256_context ctx;
-        sph_blake256_init(&ctx);
-        sph_blake256(&ctx, input, 180);
-        sph_blake256_close(&ctx, state);
-}
-
-int scanhash_decred( struct work *work, uint32_t max_nonce,
-               uint64_t *hashes_done, struct thr_info *mythr )
-{
-        uint32_t _ALIGN(64) endiandata[48];
-        uint32_t _ALIGN(64) hash32[8];
-        uint32_t *pdata = work->data;
-        uint32_t *ptarget = work->target;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
-
-//        #define DCR_NONCE_OFT32 35
-
-        const uint32_t first_nonce = pdata[DECRED_NONCE_INDEX];
-        const uint32_t HTarget = opt_benchmark ? 0x7f : ptarget[7];
-
-        uint32_t n = first_nonce;
-
-        ctx_midstate_done = false;
-
-#if 1
-        memcpy(endiandata, pdata, 180);
-#else
-        for (int k=0; k < (180/4); k++)
-                be32enc(&endiandata[k], pdata[k]);
-#endif
-
-        do {
-                //be32enc(&endiandata[DCR_NONCE_OFT32], n);
-                endiandata[DECRED_NONCE_INDEX] = n;
-                decred_hash(hash32, endiandata);
-
-                if (hash32[7] <= HTarget && fulltest(hash32, ptarget))
-                {
-                   pdata[DECRED_NONCE_INDEX] = n;
-                   submit_solution( work, hash32, mythr );
-                }
-
-                n++;
-
-        } while (n < max_nonce && !work_restart[thr_id].restart);
-
-        *hashes_done = n - first_nonce + 1;
-        pdata[DECRED_NONCE_INDEX] = n;
-        return 0;
-}
-
-/*
-uint32_t *decred_get_nonceptr( uint32_t *work_data )
-{
-   return &work_data[ DECRED_NONCE_INDEX ];
-}
-
-double decred_calc_network_diff( struct work* work )
-{
-   // sample for diff 43.281 : 1c05ea29
-   // todo: endian reversed on longpoll could be zr5 specific...
-   uint32_t nbits = work->data[ DECRED_NBITS_INDEX ];
-   uint32_t bits = ( nbits & 0xffffff );
-   int16_t shift = ( swab32(nbits) & 0xff ); // 0x1c = 28
-   int m;
-   double d = (double)0x0000ffff / (double)bits;
-
-   for ( m = shift; m < 29; m++ )
-       d *= 256.0;
-   for ( m = 29; m < shift; m++ )
-       d /= 256.0;
-   if ( shift == 28 )
-       d *= 256.0; // testnet
-   if ( opt_debug_diff )
-       applog( LOG_DEBUG, "net diff: %f -> shift %u, bits %08x", d,
-                           shift, bits );
-   return net_diff;
-}
-
-void decred_decode_extradata( struct work* work, uint64_t* net_blocks )
-{
-   // some random extradata to make the work unique
-   work->data[ DECRED_XNONCE_INDEX ] = (rand()*4);
-   work->height = work->data[32];
-   if (!have_longpoll && work->height > *net_blocks + 1)
-   {
-      char netinfo[64] = { 0 };
-      if (net_diff > 0.)
-      {
-         if (net_diff != work->targetdiff)
-	    sprintf(netinfo, ", diff %.3f, target %.1f", net_diff,
-                   work->targetdiff);
-	 else
-	     sprintf(netinfo, ", diff %.3f", net_diff);
-       }
-       applog(LOG_BLUE, "%s block %d%s", algo_names[opt_algo], work->height,
-                       netinfo);
-       *net_blocks = work->height - 1;
-   }
-}
-
-void decred_be_build_stratum_request( char *req, struct work *work,
-                                      struct stratum_ctx *sctx )
-{
-   unsigned char *xnonce2str;
-   uint32_t ntime, nonce;
-   char ntimestr[9], noncestr[9];
-
-   be32enc( &ntime, work->data[ DECRED_NTIME_INDEX ] );
-   be32enc( &nonce, work->data[ DECRED_NONCE_INDEX ] );
-   bin2hex( ntimestr, (char*)(&ntime), sizeof(uint32_t) );
-   bin2hex( noncestr, (char*)(&nonce), sizeof(uint32_t) );
-   xnonce2str = abin2hex( (char*)( &work->data[ DECRED_XNONCE_INDEX ] ),
-                                     sctx->xnonce1_size );
-   snprintf( req, JSON_BUF_LEN,
-        "{\"method\": \"mining.submit\", \"params\": [\"%s\", \"%s\", \"%s\", \"%s\", \"%s\"], \"id\":4}",
-         rpc_user, work->job_id, xnonce2str, ntimestr, noncestr );
-   free(xnonce2str);
-}
-*/
-/*
-// data shared between gen_merkle_root and build_extraheader.
-__thread uint32_t decred_extraheader[32] = { 0 };
-__thread int decred_headersize = 0;
-
-void decred_gen_merkle_root( char* merkle_root, struct stratum_ctx* sctx )
-{
-   // getwork over stratum, getwork merkle + header passed in coinb1
-   memcpy(merkle_root, sctx->job.coinbase, 32);
-   decred_headersize = min((int)sctx->job.coinbase_size - 32, 
-                  sizeof(decred_extraheader) );
-   memcpy( decred_extraheader, &sctx->job.coinbase[32], decred_headersize);
-}
-*/
-
-/*
-#define min(a,b) (a>b ? (b) :(a))
-
-void decred_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
-{
-   uchar merkle_root[64] = { 0 };
-   uint32_t extraheader[32] = { 0 };
-   int headersize = 0;
-   uint32_t* extradata = (uint32_t*) sctx->xnonce1;
-   size_t t;
-   int i;
-
-   // getwork over stratum, getwork merkle + header passed in coinb1
-   memcpy(merkle_root, sctx->job.coinbase, 32);
-   headersize = min((int)sctx->job.coinbase_size - 32,
-                  sizeof(extraheader) );
-   memcpy( extraheader, &sctx->job.coinbase[32], headersize );
-
-   // Increment extranonce2 
-   for ( t = 0; t < sctx->xnonce2_size && !( ++sctx->job.xnonce2[t] ); t++ );
-
-   // Assemble block header 
-   memset( g_work->data, 0, sizeof(g_work->data) );
-   g_work->data[0] = le32dec( sctx->job.version );
-   for ( i = 0; i < 8; i++ )
-      g_work->data[1 + i] = swab32(
-                              le32dec( (uint32_t *) sctx->job.prevhash + i ) );
-   for ( i = 0; i < 8; i++ )
-      g_work->data[9 + i] = swab32( be32dec( (uint32_t *) merkle_root + i ) );
-
-//   for ( i = 0; i < 8; i++ ) // prevhash
-//      g_work->data[1 + i] = swab32( g_work->data[1 + i] );
-//   for ( i = 0; i < 8; i++ ) // merkle
-//      g_work->data[9 + i] = swab32( g_work->data[9 + i] );
-
-   for ( i = 0; i < headersize/4; i++ ) // header
-      g_work->data[17 + i] = extraheader[i];
-   // extradata
-
-   for ( i = 0; i < sctx->xnonce1_size/4; i++ )
-      g_work->data[ DECRED_XNONCE_INDEX + i ] = extradata[i];
-   for ( i = DECRED_XNONCE_INDEX + sctx->xnonce1_size/4; i < 45; i++ )
-      g_work->data[i] = 0;
-   g_work->data[37] = (rand()*4) << 8;
-   // block header suffix from coinb2 (stake version)
-   memcpy( &g_work->data[44],
-           &sctx->job.coinbase[ sctx->job.coinbase_size-4 ], 4 );
-   sctx->bloc_height = g_work->data[32];
-   //applog_hex(work->data, 180);
-   //applog_hex(&work->data[36], 36);
-}
-
-#undef min
-
-bool decred_ready_to_mine( struct work* work, struct stratum_ctx* stratum,
-                           int thr_id )
-{
-   if ( have_stratum && strcmp(stratum->job.job_id, work->job_id)  )
-      // need to regen g_work..
-      return false;
-   if ( have_stratum && !work->data[0] && !opt_benchmark )
-   {
-      sleep(1);
-      return false;
-   }      
-   // extradata: prevent duplicates
-   work->data[ DECRED_XNONCE_INDEX     ] += 1;
-   work->data[ DECRED_XNONCE_INDEX + 1 ] |= thr_id;
-   return true;
-}
-
-
-bool register_decred_algo( algo_gate_t* gate )
-{
-  gate->optimizations         = SSE2_OPT;
-  gate->scanhash              = (void*)&scanhash_decred;
-  gate->hash                  = (void*)&decred_hash;
-  gate->get_nonceptr          = (void*)&decred_get_nonceptr;
-  gate->decode_extra_data     = (void*)&decred_decode_extradata;
-  gate->build_stratum_request = (void*)&decred_be_build_stratum_request;
-  gate->work_decode           = (void*)&std_be_work_decode;
-  gate->submit_getwork_result = (void*)&std_be_submit_getwork_result;
-  gate->build_extraheader     = (void*)&decred_build_extraheader;
-  gate->ready_to_mine         = (void*)&decred_ready_to_mine;
-  gate->nbits_index           = DECRED_NBITS_INDEX;
-  gate->ntime_index           = DECRED_NTIME_INDEX;
-  gate->nonce_index           = DECRED_NONCE_INDEX;
-  gate->work_data_size        = DECRED_DATA_SIZE;
-  gate->work_cmp_size         = DECRED_WORK_COMPARE_SIZE; 
-  allow_mininginfo            = false;
-  have_gbt                    = false;
-  return true;
-}
-*/
-
-#endif
--- a/algo/blake/pentablake-4way.c
+++ b/algo/blake/pentablake-4way.c
@@ -1,6 +1,6 @@
 #include "pentablake-gate.h"

-#if defined (__AVX2__)
+#if defined(PENTABLAKE_4WAY)

 #include <stdlib.h>
 #include <stdint.h>
--- a/algo/blake/pentablake-gate.h
+++ b/algo/blake/pentablake-gate.h
@@ -4,9 +4,10 @@
 #include "algo-gate-api.h"
 #include <stdint.h>

-#if defined(__AVX2__)
-  #define PENTABLAKE_4WAY
-#endif
+// 4way is broken
+//#if defined(__AVX2__)
+//  #define PENTABLAKE_4WAY
+//#endif

 #if defined(PENTABLAKE_4WAY)
 void pentablakehash_4way( void *state, const void *input );
--- a/algo/blake/sph-blake2s.c
+++ b/algo/blake/sph-blake2s.c
@@ -15,7 +15,7 @@
 #include <string.h>
 #include <stdio.h>

-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"
 #include "sph-blake2s.h"

 static const uint32_t blake2s_IV[8] =
--- a/algo/blake/sph_blake.h
+++ b/algo/blake/sph_blake.h
@@ -42,7 +42,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for BLAKE-224.
--- a/algo/blake/sph_blake2b.c
+++ b/algo/blake/sph_blake2b.c
@@ -31,11 +31,10 @@
 #include <stdint.h>
 #include <string.h>
 #include "simd-utils.h"
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"
 #include "sph_blake2b.h"

 // Little-endian byte access.
-
 #define B2B_GET64(p)                            \
 	(((uint64_t) ((uint8_t *) (p))[0]) ^        \
 	(((uint64_t) ((uint8_t *) (p))[1]) << 8) ^  \
@@ -46,69 +45,95 @@
 	(((uint64_t) ((uint8_t *) (p))[6]) << 48) ^ \
 	(((uint64_t) ((uint8_t *) (p))[7]) << 56))

-// G Mixing function.
-
 #if defined(__AVX2__)

-#define BLAKE2B_G( R, Sa, Sb, Sc, Sd, Na, Nb ) \
+#define BLAKE2B_G( Sa, Sb, Sc, Sd, Se, Sf, Sg, Sh ) \
 { \
  V[0] = _mm256_add_epi64( V[0], _mm256_add_epi64( V[1], \
-              _mm256_set_epi64x( m[ sigma[R][Sd] ], m[ sigma[R][Sc] ], \
-                                 m[ sigma[R][Sb] ], m[ sigma[R][Sa] ] ) ) ); \
-  V[3] = mm256_ror_64( _mm256_xor_si256( V[3], V[0] ), Na ); \
+              _mm256_set_epi64x( m[ sigmaR[ Sg ] ], m[ sigmaR[ Se ] ], \
+                                 m[ sigmaR[ Sc ] ], m[ sigmaR[ Sa ] ] ) ) ); \
+  V[3] = mm256_swap64_32( _mm256_xor_si256( V[3], V[0] ) ); \
  V[2] = _mm256_add_epi64( V[2], V[3] ); \
-  V[1] = mm256_ror_64( _mm256_xor_si256( V[1], V[2] ), Nb ); \
+  V[1] = mm256_shuflr64_24( _mm256_xor_si256( V[1], V[2] ) ); \
+\
+  V[0] = _mm256_add_epi64( V[0], _mm256_add_epi64( V[1], \
+              _mm256_set_epi64x( m[ sigmaR[ Sh ] ], m[ sigmaR[ Sf ] ], \
+                                 m[ sigmaR[ Sd ] ], m[ sigmaR[ Sb ] ] ) ) ); \
+  V[3] = mm256_shuflr64_16( _mm256_xor_si256( V[3], V[0] ) ); \
+  V[2] = _mm256_add_epi64( V[2], V[3] ); \
+  V[1] = mm256_ror_64( _mm256_xor_si256( V[1], V[2] ), 63 ); \
 }

+// Pivot about V[1] instead of V[0] reduces latency.
 #define BLAKE2B_ROUND( R ) \
 { \
  __m256i *V = (__m256i*)v; \
-  BLAKE2B_G( R,  0,  2,  4,  6, 32, 24 ); \
-  BLAKE2B_G( R,  1,  3,  5,  7, 16, 63 ); \
+  const uint8_t *sigmaR = sigma[R]; \
+  BLAKE2B_G(  0,  1,  2,  3,  4,  5,  6,  7 ); \
+  V[0] = mm256_shufll_64( V[0] ); \
+  V[3] = mm256_swap_128( V[3] ); \
+  V[2] = mm256_shuflr_64( V[2] ); \
+  BLAKE2B_G( 14, 15,  8,  9, 10, 11, 12, 13 ); \
+  V[0] = mm256_shuflr_64( V[0] ); \
+  V[3] = mm256_swap_128( V[3] ); \
+  V[2] = mm256_shufll_64( V[2] ); \
+}
+
+/*
+#define BLAKE2B_ROUND( R ) \
+{ \
+  __m256i *V = (__m256i*)v; \
+  const uint8_t *sigmaR = sigma[R]; \
+  BLAKE2B_G(  0,  1,  2,  3,  4,  5,  6,  7 ); \
  V[3] = mm256_shufll_64( V[3] ); \
  V[2] = mm256_swap_128( V[2] ); \
  V[1] = mm256_shuflr_64( V[1] ); \
-  BLAKE2B_G( R,  8, 10, 12, 14, 32, 24 ); \
-  BLAKE2B_G( R,  9, 11, 13, 15, 16, 63 ); \
+  BLAKE2B_G(  8,  9, 10, 11, 12, 13, 14, 15 ); \
  V[3] = mm256_shuflr_64( V[3] ); \
  V[2] = mm256_swap_128( V[2] ); \
  V[1] = mm256_shufll_64( V[1] ); \
 }
+*/

-#elif defined(__SSSE3__)
+#elif defined(__SSE2__)
+// always true

-#define BLAKE2B_G( R, Va, Vb, Vc, Vd, Sa, Sb, Na, Nb ) \
+#define BLAKE2B_G( Va, Vb, Vc, Vd, Sa, Sb, Sc, Sd ) \
 { \
   Va = _mm_add_epi64( Va, _mm_add_epi64( Vb, \
-                 _mm_set_epi64x( m[ sigma[R][Sb] ], m[ sigma[R][Sa] ] ) ) ); \
-   Vd = mm128_ror_64( _mm_xor_si128( Vd, Va ), Na ); \
+                 _mm_set_epi64x( m[ sigmaR[ Sc ] ], m[ sigmaR[ Sa ] ] ) ) ); \
+   Vd = mm128_swap64_32( _mm_xor_si128( Vd, Va ) ); \
   Vc = _mm_add_epi64( Vc, Vd ); \
-   Vb = mm128_ror_64( _mm_xor_si128( Vb, Vc ), Nb ); \
+   Vb = mm128_shuflr64_24( _mm_xor_si128( Vb, Vc ) ); \
+\
+   Va = _mm_add_epi64( Va, _mm_add_epi64( Vb, \
+                 _mm_set_epi64x( m[ sigmaR[ Sd ] ], m[ sigmaR[ Sb ] ] ) ) ); \
+   Vd = mm128_shuflr64_16( _mm_xor_si128( Vd, Va ) ); \
+   Vc = _mm_add_epi64( Vc, Vd ); \
+   Vb = mm128_ror_64( _mm_xor_si128( Vb, Vc ), 63 ); \
 }

 #define BLAKE2B_ROUND( R ) \
 { \
   __m128i *V = (__m128i*)v; \
   __m128i V2, V3, V6, V7; \
-   BLAKE2B_G( R, V[0], V[2], V[4], V[6], 0, 2, 32, 24 ); \
-   BLAKE2B_G( R, V[0], V[2], V[4], V[6], 1, 3, 16, 63 ); \
-   BLAKE2B_G( R, V[1], V[3], V[5], V[7], 4, 6, 32, 24 ); \
-   BLAKE2B_G( R, V[1], V[3], V[5], V[7], 5, 7, 16, 63 ); \
-   V2 = mm128_shufl2r_64( V[2], V[3] ); \
-   V3 = mm128_shufl2r_64( V[3], V[2] ); \
-   V6 = mm128_shufl2l_64( V[6], V[7] ); \
-   V7 = mm128_shufl2l_64( V[7], V[6] ); \
-   BLAKE2B_G( R, V[0], V2, V[5], V6,  8, 10, 32, 24 ); \
-   BLAKE2B_G( R, V[0], V2, V[5], V6,  9, 11, 16, 63 ); \
-   BLAKE2B_G( R, V[1], V3, V[4], V7, 12, 14, 32, 24 ); \
-   BLAKE2B_G( R, V[1], V3, V[4], V7, 13, 15, 16, 63 ); \
-   V[2] = mm128_shufl2l_64( V2, V3 ); \
-   V[3] = mm128_shufl2l_64( V3, V2 ); \
-   V[6] = mm128_shufl2r_64( V6, V7 ); \
-   V[7] = mm128_shufl2r_64( V7, V6 ); \
+   const uint8_t *sigmaR = sigma[R]; \
+   BLAKE2B_G( V[0], V[2], V[4], V[6], 0, 1, 2, 3 ); \
+   BLAKE2B_G( V[1], V[3], V[5], V[7], 4, 5, 6, 7 ); \
+   V2 = mm128_alignr_64( V[3], V[2], 1 ); \
+   V3 = mm128_alignr_64( V[2], V[3], 1 ); \
+   V6 = mm128_alignr_64( V[6], V[7], 1 ); \
+   V7 = mm128_alignr_64( V[7], V[6], 1 ); \
+   BLAKE2B_G( V[0], V2, V[5], V6,  8,  9, 10, 11 ); \
+   BLAKE2B_G( V[1], V3, V[4], V7, 12, 13, 14, 15 ); \
+   V[2] = mm128_alignr_64( V2, V3, 1 ); \
+   V[3] = mm128_alignr_64( V3, V2, 1 ); \
+   V[6] = mm128_alignr_64( V7, V6, 1 ); \
+   V[7] = mm128_alignr_64( V6, V7, 1 ); \
 }

 #else
+// never used, SSE2 is always available

 #ifndef ROTR64
 #define ROTR64(x, y)  (((x) >> (y)) ^ ((x) << (64 - (y))))
@@ -120,6 +145,7 @@
   Vd = ROTR64( Vd ^ Va, 32 ); \
   Vc = Vc + Vd; \
   Vb = ROTR64( Vb ^ Vc, 24 ); \
+\
   Va = Va + Vb + m[ sigma[R][Sb] ]; \
   Vd = ROTR64( Vd ^ Va, 16 ); \
   Vc = Vc + Vd; \
--- a/algo/bmw/bmw-hash-4way.h
+++ b/algo/bmw/bmw-hash-4way.h
@@ -41,8 +41,6 @@ extern "C"{
 #endif

 #include <stddef.h>
-
-#include "algo/sha/sph_types.h"
 #include "simd-utils.h"

 #define SPH_SIZE_bmw256   256
@@ -57,7 +55,7 @@ typedef struct {
   __m128i buf[64];
   __m128i H[16];
   size_t ptr;
-   sph_u32 bit_count;  // assume bit_count fits in 32 bits
+   uint32_t bit_count;  // assume bit_count fits in 32 bits
 } bmw_4way_small_context;

 typedef bmw_4way_small_context bmw256_4way_context;
@@ -144,7 +142,7 @@ typedef struct {
   __m256i buf[16];
   __m256i H[16];
   size_t ptr;
-   sph_u64 bit_count;
+   uint64_t bit_count;
 } bmw_4way_big_context __attribute__((aligned(128)));

 typedef bmw_4way_big_context bmw512_4way_context;
--- a/algo/bmw/bmw256-hash-4way.c
+++ b/algo/bmw/bmw256-hash-4way.c
@@ -109,7 +109,7 @@ static const uint32_t IV256[] = {
             _mm_sub_epi32( _mm_add_epi32( rol_off_32( M, j, 0 ), \
                                           rol_off_32( M, j, 3 ) ), \
                            rol_off_32( M, j, 10 ) ), \
-       _mm_set1_epi32( ( (j)+16 ) * SPH_C32(0x05555555UL) ) ), \
+       _mm_set1_epi32( ( (j)+16 ) * 0x05555555UL ) ), \
   H[ ( (j)+7 ) & 0xF ] )


@@ -451,22 +451,22 @@ static const __m128i final_s[16] =
 */
 void bmw256_4way_init( bmw256_4way_context *ctx )
 {
-   ctx->H[ 0] = m128_const1_64( 0x4041424340414243 );
-   ctx->H[ 1] = m128_const1_64( 0x4445464744454647 );
-   ctx->H[ 2] = m128_const1_64( 0x48494A4B48494A4B );
-   ctx->H[ 3] = m128_const1_64( 0x4C4D4E4F4C4D4E4F );
-   ctx->H[ 4] = m128_const1_64( 0x5051525350515253 );
-   ctx->H[ 5] = m128_const1_64( 0x5455565754555657 );
-   ctx->H[ 6] = m128_const1_64( 0x58595A5B58595A5B );
-   ctx->H[ 7] = m128_const1_64( 0x5C5D5E5F5C5D5E5F );
-   ctx->H[ 8] = m128_const1_64( 0x6061626360616263 );
-   ctx->H[ 9] = m128_const1_64( 0x6465666764656667 );
-   ctx->H[10] = m128_const1_64( 0x68696A6B68696A6B );
-   ctx->H[11] = m128_const1_64( 0x6C6D6E6F6C6D6E6F );
-   ctx->H[12] = m128_const1_64( 0x7071727370717273 );
-   ctx->H[13] = m128_const1_64( 0x7475767774757677 );
-   ctx->H[14] = m128_const1_64( 0x78797A7B78797A7B );
-   ctx->H[15] = m128_const1_64( 0x7C7D7E7F7C7D7E7F );
+   ctx->H[ 0] = _mm_set1_epi64x( 0x4041424340414243 );
+   ctx->H[ 1] = _mm_set1_epi64x( 0x4445464744454647 );
+   ctx->H[ 2] = _mm_set1_epi64x( 0x48494A4B48494A4B );
+   ctx->H[ 3] = _mm_set1_epi64x( 0x4C4D4E4F4C4D4E4F );
+   ctx->H[ 4] = _mm_set1_epi64x( 0x5051525350515253 );
+   ctx->H[ 5] = _mm_set1_epi64x( 0x5455565754555657 );
+   ctx->H[ 6] = _mm_set1_epi64x( 0x58595A5B58595A5B );
+   ctx->H[ 7] = _mm_set1_epi64x( 0x5C5D5E5F5C5D5E5F );
+   ctx->H[ 8] = _mm_set1_epi64x( 0x6061626360616263 );
+   ctx->H[ 9] = _mm_set1_epi64x( 0x6465666764656667 );
+   ctx->H[10] = _mm_set1_epi64x( 0x68696A6B68696A6B );
+   ctx->H[11] = _mm_set1_epi64x( 0x6C6D6E6F6C6D6E6F );
+   ctx->H[12] = _mm_set1_epi64x( 0x7071727370717273 );
+   ctx->H[13] = _mm_set1_epi64x( 0x7475767774757677 );
+   ctx->H[14] = _mm_set1_epi64x( 0x78797A7B78797A7B );
+   ctx->H[15] = _mm_set1_epi64x( 0x7C7D7E7F7C7D7E7F );


 //   for ( int i = 0; i < 16; i++ )
@@ -485,7 +485,7 @@ bmw32_4way(bmw_4way_small_context *sc, const void *data, size_t len)
   size_t ptr;
   const int buf_size = 64;  // bytes of one lane, compatible with len

-   sc->bit_count += (sph_u32)len << 3;
+   sc->bit_count += (uint32_t)len << 3;
   buf = sc->buf;
   ptr = sc->ptr;
   h1 = sc->H;
@@ -529,7 +529,7 @@ bmw32_4way_close(bmw_4way_small_context *sc, unsigned ub, unsigned n,

   buf = sc->buf;
   ptr = sc->ptr;
-   buf[ ptr>>2 ] = m128_const1_64( 0x0000008000000080 );
+   buf[ ptr>>2 ] = _mm_set1_epi64x( 0x0000008000000080 );
   ptr += 4;
   h = sc->H;

@@ -959,22 +959,22 @@ static const __m256i final_s8[16] =

 void bmw256_8way_init( bmw256_8way_context *ctx )
 {
-   ctx->H[ 0] = m256_const1_64( 0x4041424340414243 );
-   ctx->H[ 1] = m256_const1_64( 0x4445464744454647 );
-   ctx->H[ 2] = m256_const1_64( 0x48494A4B48494A4B );
-   ctx->H[ 3] = m256_const1_64( 0x4C4D4E4F4C4D4E4F );
-   ctx->H[ 4] = m256_const1_64( 0x5051525350515253 );
-   ctx->H[ 5] = m256_const1_64( 0x5455565754555657 );
-   ctx->H[ 6] = m256_const1_64( 0x58595A5B58595A5B );
-   ctx->H[ 7] = m256_const1_64( 0x5C5D5E5F5C5D5E5F );
-   ctx->H[ 8] = m256_const1_64( 0x6061626360616263 );
-   ctx->H[ 9] = m256_const1_64( 0x6465666764656667 );
-   ctx->H[10] = m256_const1_64( 0x68696A6B68696A6B );
-   ctx->H[11] = m256_const1_64( 0x6C6D6E6F6C6D6E6F );
-   ctx->H[12] = m256_const1_64( 0x7071727370717273 );
-   ctx->H[13] = m256_const1_64( 0x7475767774757677 );
-   ctx->H[14] = m256_const1_64( 0x78797A7B78797A7B );
-   ctx->H[15] = m256_const1_64( 0x7C7D7E7F7C7D7E7F );
+   ctx->H[ 0] = _mm256_set1_epi64x( 0x4041424340414243 );
+   ctx->H[ 1] = _mm256_set1_epi64x( 0x4445464744454647 );
+   ctx->H[ 2] = _mm256_set1_epi64x( 0x48494A4B48494A4B );
+   ctx->H[ 3] = _mm256_set1_epi64x( 0x4C4D4E4F4C4D4E4F );
+   ctx->H[ 4] = _mm256_set1_epi64x( 0x5051525350515253 );
+   ctx->H[ 5] = _mm256_set1_epi64x( 0x5455565754555657 );
+   ctx->H[ 6] = _mm256_set1_epi64x( 0x58595A5B58595A5B );
+   ctx->H[ 7] = _mm256_set1_epi64x( 0x5C5D5E5F5C5D5E5F );
+   ctx->H[ 8] = _mm256_set1_epi64x( 0x6061626360616263 );
+   ctx->H[ 9] = _mm256_set1_epi64x( 0x6465666764656667 );
+   ctx->H[10] = _mm256_set1_epi64x( 0x68696A6B68696A6B );
+   ctx->H[11] = _mm256_set1_epi64x( 0x6C6D6E6F6C6D6E6F );
+   ctx->H[12] = _mm256_set1_epi64x( 0x7071727370717273 );
+   ctx->H[13] = _mm256_set1_epi64x( 0x7475767774757677 );
+   ctx->H[14] = _mm256_set1_epi64x( 0x78797A7B78797A7B );
+   ctx->H[15] = _mm256_set1_epi64x( 0x7C7D7E7F7C7D7E7F );
   ctx->ptr       = 0;
   ctx->bit_count = 0;
 }
@@ -1030,7 +1030,7 @@ void bmw256_8way_close( bmw256_8way_context *ctx, void *dst )

   buf = ctx->buf;
   ptr = ctx->ptr;
-   buf[ ptr>>2 ] = m256_const1_64( 0x0000008000000080 );
+   buf[ ptr>>2 ] = _mm256_set1_epi64x( 0x0000008000000080 );
   ptr += 4;
   h = ctx->H;

@@ -1460,22 +1460,22 @@ static const __m512i final_s16[16] =

 void bmw256_16way_init( bmw256_16way_context *ctx )
 {
-   ctx->H[ 0] = m512_const1_64( 0x4041424340414243 );
-   ctx->H[ 1] = m512_const1_64( 0x4445464744454647 );
-   ctx->H[ 2] = m512_const1_64( 0x48494A4B48494A4B );
-   ctx->H[ 3] = m512_const1_64( 0x4C4D4E4F4C4D4E4F );
-   ctx->H[ 4] = m512_const1_64( 0x5051525350515253 );
-   ctx->H[ 5] = m512_const1_64( 0x5455565754555657 );
-   ctx->H[ 6] = m512_const1_64( 0x58595A5B58595A5B );
-   ctx->H[ 7] = m512_const1_64( 0x5C5D5E5F5C5D5E5F );
-   ctx->H[ 8] = m512_const1_64( 0x6061626360616263 );
-   ctx->H[ 9] = m512_const1_64( 0x6465666764656667 );
-   ctx->H[10] = m512_const1_64( 0x68696A6B68696A6B );
-   ctx->H[11] = m512_const1_64( 0x6C6D6E6F6C6D6E6F );
-   ctx->H[12] = m512_const1_64( 0x7071727370717273 );
-   ctx->H[13] = m512_const1_64( 0x7475767774757677 );
-   ctx->H[14] = m512_const1_64( 0x78797A7B78797A7B );
-   ctx->H[15] = m512_const1_64( 0x7C7D7E7F7C7D7E7F );
+   ctx->H[ 0] = _mm512_set1_epi64( 0x4041424340414243 );
+   ctx->H[ 1] = _mm512_set1_epi64( 0x4445464744454647 );
+   ctx->H[ 2] = _mm512_set1_epi64( 0x48494A4B48494A4B );
+   ctx->H[ 3] = _mm512_set1_epi64( 0x4C4D4E4F4C4D4E4F );
+   ctx->H[ 4] = _mm512_set1_epi64( 0x5051525350515253 );
+   ctx->H[ 5] = _mm512_set1_epi64( 0x5455565754555657 );
+   ctx->H[ 6] = _mm512_set1_epi64( 0x58595A5B58595A5B );
+   ctx->H[ 7] = _mm512_set1_epi64( 0x5C5D5E5F5C5D5E5F );
+   ctx->H[ 8] = _mm512_set1_epi64( 0x6061626360616263 );
+   ctx->H[ 9] = _mm512_set1_epi64( 0x6465666764656667 );
+   ctx->H[10] = _mm512_set1_epi64( 0x68696A6B68696A6B );
+   ctx->H[11] = _mm512_set1_epi64( 0x6C6D6E6F6C6D6E6F );
+   ctx->H[12] = _mm512_set1_epi64( 0x7071727370717273 );
+   ctx->H[13] = _mm512_set1_epi64( 0x7475767774757677 );
+   ctx->H[14] = _mm512_set1_epi64( 0x78797A7B78797A7B );
+   ctx->H[15] = _mm512_set1_epi64( 0x7C7D7E7F7C7D7E7F );
   ctx->ptr       = 0;
   ctx->bit_count = 0;
 }
@@ -1531,7 +1531,7 @@ void bmw256_16way_close( bmw256_16way_context *ctx, void *dst )

   buf = ctx->buf;
   ptr = ctx->ptr;
-   buf[ ptr>>2 ] = m512_const1_64( 0x0000008000000080 );
+   buf[ ptr>>2 ] = _mm512_set1_epi64( 0x0000008000000080 );
   ptr += 4;
   h = ctx->H;

--- a/algo/bmw/bmw512-hash-4way.c
+++ b/algo/bmw/bmw512-hash-4way.c
@@ -45,15 +45,15 @@ extern "C"{

 #define LPAR   (

-static const sph_u64 IV512[] = {
-        SPH_C64(0x8081828384858687), SPH_C64(0x88898A8B8C8D8E8F),
-        SPH_C64(0x9091929394959697), SPH_C64(0x98999A9B9C9D9E9F),
-        SPH_C64(0xA0A1A2A3A4A5A6A7), SPH_C64(0xA8A9AAABACADAEAF),
-        SPH_C64(0xB0B1B2B3B4B5B6B7), SPH_C64(0xB8B9BABBBCBDBEBF),
-        SPH_C64(0xC0C1C2C3C4C5C6C7), SPH_C64(0xC8C9CACBCCCDCECF),
-        SPH_C64(0xD0D1D2D3D4D5D6D7), SPH_C64(0xD8D9DADBDCDDDEDF),
-        SPH_C64(0xE0E1E2E3E4E5E6E7), SPH_C64(0xE8E9EAEBECEDEEEF),
-        SPH_C64(0xF0F1F2F3F4F5F6F7), SPH_C64(0xF8F9FAFBFCFDFEFF)
+static const uint64_t IV512[] = {
+        0x8081828384858687, 0x88898A8B8C8D8E8F,
+        0x9091929394959697, 0x98999A9B9C9D9E9F,
+        0xA0A1A2A3A4A5A6A7, 0xA8A9AAABACADAEAF,
+        0xB0B1B2B3B4B5B6B7, 0xB8B9BABBBCBDBEBF,
+        0xC0C1C2C3C4C5C6C7, 0xC8C9CACBCCCDCECF,
+        0xD0D1D2D3D4D5D6D7, 0xD8D9DADBDCDDDEDF,
+        0xE0E1E2E3E4E5E6E7, 0xE8E9EAEBECEDEEEF,
+        0xF0F1F2F3F4F5F6F7, 0xF8F9FAFBFCFDFEFF
 };

 #if defined(__SSE2__)
@@ -747,38 +747,40 @@ void compress_big( const __m256i *M, const __m256i H[16], __m256i dH[16] )
   mj[14] = mm256_rol_64( M[14], 15 );
   mj[15] = mm256_rol_64( M[15], 16 );

-   qt[16] = add_elt_b( mj[ 0], mj[ 3], mj[10], H[ 7],
-              (const __m256i)_mm256_set1_epi64x( 16 * 0x0555555555555555ULL ) );
-   qt[17] = add_elt_b( mj[ 1], mj[ 4], mj[11], H[ 8],
-              (const __m256i)_mm256_set1_epi64x( 17 * 0x0555555555555555ULL ) );
-   qt[18] = add_elt_b( mj[ 2], mj[ 5], mj[12], H[ 9],
-              (const __m256i)_mm256_set1_epi64x( 18 * 0x0555555555555555ULL ) );
-   qt[19] = add_elt_b( mj[ 3], mj[ 6], mj[13], H[10],
-              (const __m256i)_mm256_set1_epi64x( 19 * 0x0555555555555555ULL ) );
-   qt[20] = add_elt_b( mj[ 4], mj[ 7], mj[14], H[11],
-              (const __m256i)_mm256_set1_epi64x( 20 * 0x0555555555555555ULL ) );
-   qt[21] = add_elt_b( mj[ 5], mj[ 8], mj[15], H[12],
-              (const __m256i)_mm256_set1_epi64x( 21 * 0x0555555555555555ULL ) );
-   qt[22] = add_elt_b( mj[ 6], mj[ 9], mj[ 0], H[13],
-              (const __m256i)_mm256_set1_epi64x( 22 * 0x0555555555555555ULL ) );
-   qt[23] = add_elt_b( mj[ 7], mj[10], mj[ 1], H[14],
-              (const __m256i)_mm256_set1_epi64x( 23 * 0x0555555555555555ULL ) );
-   qt[24] = add_elt_b( mj[ 8], mj[11], mj[ 2], H[15],
-              (const __m256i)_mm256_set1_epi64x( 24 * 0x0555555555555555ULL ) );
-   qt[25] = add_elt_b( mj[ 9], mj[12], mj[ 3], H[ 0],
-              (const __m256i)_mm256_set1_epi64x( 25 * 0x0555555555555555ULL ) );
-   qt[26] = add_elt_b( mj[10], mj[13], mj[ 4], H[ 1],
-              (const __m256i)_mm256_set1_epi64x( 26 * 0x0555555555555555ULL ) );
-   qt[27] = add_elt_b( mj[11], mj[14], mj[ 5], H[ 2],
-              (const __m256i)_mm256_set1_epi64x( 27 * 0x0555555555555555ULL ) );
-   qt[28] = add_elt_b( mj[12], mj[15], mj[ 6], H[ 3],
-              (const __m256i)_mm256_set1_epi64x( 28 * 0x0555555555555555ULL ) );
-   qt[29] = add_elt_b( mj[13], mj[ 0], mj[ 7], H[ 4],
-              (const __m256i)_mm256_set1_epi64x( 29 * 0x0555555555555555ULL ) );
-   qt[30] = add_elt_b( mj[14], mj[ 1], mj[ 8], H[ 5],
-              (const __m256i)_mm256_set1_epi64x( 30 * 0x0555555555555555ULL ) );
-   qt[31] = add_elt_b( mj[15], mj[ 2], mj[ 9], H[ 6],
-              (const __m256i)_mm256_set1_epi64x( 31 * 0x0555555555555555ULL ) );
+   __m256i K = _mm256_set1_epi64x( 16 * 0x0555555555555555ULL );
+   const __m256i Kincr = _mm256_set1_epi64x( 0x0555555555555555ULL );
+
+   qt[16] = add_elt_b( mj[ 0], mj[ 3], mj[10], H[ 7], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[17] = add_elt_b( mj[ 1], mj[ 4], mj[11], H[ 8], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[18] = add_elt_b( mj[ 2], mj[ 5], mj[12], H[ 9], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[19] = add_elt_b( mj[ 3], mj[ 6], mj[13], H[10], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[20] = add_elt_b( mj[ 4], mj[ 7], mj[14], H[11], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[21] = add_elt_b( mj[ 5], mj[ 8], mj[15], H[12], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[22] = add_elt_b( mj[ 6], mj[ 9], mj[ 0], H[13], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[23] = add_elt_b( mj[ 7], mj[10], mj[ 1], H[14], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[24] = add_elt_b( mj[ 8], mj[11], mj[ 2], H[15], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[25] = add_elt_b( mj[ 9], mj[12], mj[ 3], H[ 0], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[26] = add_elt_b( mj[10], mj[13], mj[ 4], H[ 1], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[27] = add_elt_b( mj[11], mj[14], mj[ 5], H[ 2], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[28] = add_elt_b( mj[12], mj[15], mj[ 6], H[ 3], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[29] = add_elt_b( mj[13], mj[ 0], mj[ 7], H[ 4], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[30] = add_elt_b( mj[14], mj[ 1], mj[ 8], H[ 5], K );
+   K = _mm256_add_epi64( K, Kincr );
+   qt[31] = add_elt_b( mj[15], mj[ 2], mj[ 9], H[ 6], K );

   qt[16] = _mm256_add_epi64( qt[16], expand1_b( qt, 16 ) );
   qt[17] = _mm256_add_epi64( qt[17], expand1_b( qt, 17 ) );
@@ -892,24 +894,24 @@ static const __m256i final_b[16] =
 };

 static void
-bmw64_4way_init( bmw_4way_big_context *sc, const sph_u64 *iv )
+bmw64_4way_init( bmw_4way_big_context *sc, const uint64_t *iv )
 {
-   sc->H[ 0] = m256_const1_64( 0x8081828384858687 );
-   sc->H[ 1] = m256_const1_64( 0x88898A8B8C8D8E8F );
-   sc->H[ 2] = m256_const1_64( 0x9091929394959697 );
-   sc->H[ 3] = m256_const1_64( 0x98999A9B9C9D9E9F );
-   sc->H[ 4] = m256_const1_64( 0xA0A1A2A3A4A5A6A7 );
-   sc->H[ 5] = m256_const1_64( 0xA8A9AAABACADAEAF );
-   sc->H[ 6] = m256_const1_64( 0xB0B1B2B3B4B5B6B7 );
-   sc->H[ 7] = m256_const1_64( 0xB8B9BABBBCBDBEBF );
-   sc->H[ 8] = m256_const1_64( 0xC0C1C2C3C4C5C6C7 );
-   sc->H[ 9] = m256_const1_64( 0xC8C9CACBCCCDCECF );
-   sc->H[10] = m256_const1_64( 0xD0D1D2D3D4D5D6D7 );
-   sc->H[11] = m256_const1_64( 0xD8D9DADBDCDDDEDF );
-   sc->H[12] = m256_const1_64( 0xE0E1E2E3E4E5E6E7 );
-   sc->H[13] = m256_const1_64( 0xE8E9EAEBECEDEEEF );
-   sc->H[14] = m256_const1_64( 0xF0F1F2F3F4F5F6F7 );
-   sc->H[15] = m256_const1_64( 0xF8F9FAFBFCFDFEFF );
+   sc->H[ 0] = _mm256_set1_epi64x( 0x8081828384858687 );
+   sc->H[ 1] = _mm256_set1_epi64x( 0x88898A8B8C8D8E8F );
+   sc->H[ 2] = _mm256_set1_epi64x( 0x9091929394959697 );
+   sc->H[ 3] = _mm256_set1_epi64x( 0x98999A9B9C9D9E9F );
+   sc->H[ 4] = _mm256_set1_epi64x( 0xA0A1A2A3A4A5A6A7 );
+   sc->H[ 5] = _mm256_set1_epi64x( 0xA8A9AAABACADAEAF );
+   sc->H[ 6] = _mm256_set1_epi64x( 0xB0B1B2B3B4B5B6B7 );
+   sc->H[ 7] = _mm256_set1_epi64x( 0xB8B9BABBBCBDBEBF );
+   sc->H[ 8] = _mm256_set1_epi64x( 0xC0C1C2C3C4C5C6C7 );
+   sc->H[ 9] = _mm256_set1_epi64x( 0xC8C9CACBCCCDCECF );
+   sc->H[10] = _mm256_set1_epi64x( 0xD0D1D2D3D4D5D6D7 );
+   sc->H[11] = _mm256_set1_epi64x( 0xD8D9DADBDCDDDEDF );
+   sc->H[12] = _mm256_set1_epi64x( 0xE0E1E2E3E4E5E6E7 );
+   sc->H[13] = _mm256_set1_epi64x( 0xE8E9EAEBECEDEEEF );
+   sc->H[14] = _mm256_set1_epi64x( 0xF0F1F2F3F4F5F6F7 );
+   sc->H[15] = _mm256_set1_epi64x( 0xF8F9FAFBFCFDFEFF );
   sc->ptr = 0;
   sc->bit_count = 0;
 }
@@ -924,7 +926,7 @@ bmw64_4way( bmw_4way_big_context *sc, const void *data, size_t len )
   size_t ptr;
   const int buf_size = 128;  // bytes of one lane, compatible with len

-   sc->bit_count += (sph_u64)len << 3;
+   sc->bit_count += (uint64_t)len << 3;
   buf = sc->buf;
   ptr = sc->ptr;
   h1 = sc->H;
@@ -965,7 +967,7 @@ bmw64_4way_close(bmw_4way_big_context *sc, unsigned ub, unsigned n,

   buf = sc->buf;
   ptr = sc->ptr;
-   buf[ ptr>>3 ] = m256_const1_64( 0x80 );
+   buf[ ptr>>3 ] = _mm256_set1_epi64x( 0x80 );
   ptr += 8;
   h = sc->H;

@@ -1180,7 +1182,6 @@ void compress_big_8way( const __m512i *M, const __m512i H[16],
   qt[15] = _mm512_add_epi64( s8b0( W8b15), H[ 0] );

   __m512i mj[16];
-   uint64_t K = 16 * 0x0555555555555555ULL;
 
   mj[ 0] = mm512_rol_64( M[ 0],  1 );
   mj[ 1] = mm512_rol_64( M[ 1],  2 );
@@ -1199,54 +1200,40 @@ void compress_big_8way( const __m512i *M, const __m512i H[16],
   mj[14] = mm512_rol_64( M[14], 15 );
   mj[15] = mm512_rol_64( M[15], 16 );

-   qt[16] = add_elt_b8( mj[ 0], mj[ 3], mj[10], H[ 7],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[17] = add_elt_b8( mj[ 1], mj[ 4], mj[11], H[ 8],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[18] = add_elt_b8( mj[ 2], mj[ 5], mj[12], H[ 9],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[19] = add_elt_b8( mj[ 3], mj[ 6], mj[13], H[10],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[20] = add_elt_b8( mj[ 4], mj[ 7], mj[14], H[11],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[21] = add_elt_b8( mj[ 5], mj[ 8], mj[15], H[12],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[22] = add_elt_b8( mj[ 6], mj[ 9], mj[ 0], H[13],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[23] = add_elt_b8( mj[ 7], mj[10], mj[ 1], H[14],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[24] = add_elt_b8( mj[ 8], mj[11], mj[ 2], H[15],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[25] = add_elt_b8( mj[ 9], mj[12], mj[ 3], H[ 0],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[26] = add_elt_b8( mj[10], mj[13], mj[ 4], H[ 1],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[27] = add_elt_b8( mj[11], mj[14], mj[ 5], H[ 2],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[28] = add_elt_b8( mj[12], mj[15], mj[ 6], H[ 3],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[29] = add_elt_b8( mj[13], mj[ 0], mj[ 7], H[ 4],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[30] = add_elt_b8( mj[14], mj[ 1], mj[ 8], H[ 5],
-                        (const __m512i)_mm512_set1_epi64( K ) );
-   K += 0x0555555555555555ULL;
-   qt[31] = add_elt_b8( mj[15], mj[ 2], mj[ 9], H[ 6],
-                        (const __m512i)_mm512_set1_epi64( K ) );
+   __m512i K = _mm512_set1_epi64( 16 * 0x0555555555555555ULL );
+   const __m512i Kincr = _mm512_set1_epi64( 0x0555555555555555ULL );

+   qt[16] = add_elt_b8( mj[ 0], mj[ 3], mj[10], H[ 7], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[17] = add_elt_b8( mj[ 1], mj[ 4], mj[11], H[ 8], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[18] = add_elt_b8( mj[ 2], mj[ 5], mj[12], H[ 9], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[19] = add_elt_b8( mj[ 3], mj[ 6], mj[13], H[10], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[20] = add_elt_b8( mj[ 4], mj[ 7], mj[14], H[11], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[21] = add_elt_b8( mj[ 5], mj[ 8], mj[15], H[12], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[22] = add_elt_b8( mj[ 6], mj[ 9], mj[ 0], H[13], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[23] = add_elt_b8( mj[ 7], mj[10], mj[ 1], H[14], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[24] = add_elt_b8( mj[ 8], mj[11], mj[ 2], H[15], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[25] = add_elt_b8( mj[ 9], mj[12], mj[ 3], H[ 0], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[26] = add_elt_b8( mj[10], mj[13], mj[ 4], H[ 1], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[27] = add_elt_b8( mj[11], mj[14], mj[ 5], H[ 2], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[28] = add_elt_b8( mj[12], mj[15], mj[ 6], H[ 3], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[29] = add_elt_b8( mj[13], mj[ 0], mj[ 7], H[ 4], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[30] = add_elt_b8( mj[14], mj[ 1], mj[ 8], H[ 5], K );
+   K = _mm512_add_epi64( K, Kincr );
+   qt[31] = add_elt_b8( mj[15], mj[ 2], mj[ 9], H[ 6], K );

   qt[16] = _mm512_add_epi64( qt[16], expand1_b8( qt, 16 ) );
   qt[17] = _mm512_add_epi64( qt[17], expand1_b8( qt, 17 ) );
@@ -1390,24 +1377,24 @@ static const __m512i final_b8[16] =


 void bmw512_8way_init( bmw512_8way_context *ctx )
-//bmw64_4way_init( bmw_4way_big_context *sc, const sph_u64 *iv )
+//bmw64_4way_init( bmw_4way_big_context *sc, const uint64_t *iv )
 {
-   ctx->H[ 0] = m512_const1_64( 0x8081828384858687 );
-   ctx->H[ 1] = m512_const1_64( 0x88898A8B8C8D8E8F );
-   ctx->H[ 2] = m512_const1_64( 0x9091929394959697 );
-   ctx->H[ 3] = m512_const1_64( 0x98999A9B9C9D9E9F );
-   ctx->H[ 4] = m512_const1_64( 0xA0A1A2A3A4A5A6A7 );
-   ctx->H[ 5] = m512_const1_64( 0xA8A9AAABACADAEAF );
-   ctx->H[ 6] = m512_const1_64( 0xB0B1B2B3B4B5B6B7 );
-   ctx->H[ 7] = m512_const1_64( 0xB8B9BABBBCBDBEBF );
-   ctx->H[ 8] = m512_const1_64( 0xC0C1C2C3C4C5C6C7 );
-   ctx->H[ 9] = m512_const1_64( 0xC8C9CACBCCCDCECF );
-   ctx->H[10] = m512_const1_64( 0xD0D1D2D3D4D5D6D7 );
-   ctx->H[11] = m512_const1_64( 0xD8D9DADBDCDDDEDF );
-   ctx->H[12] = m512_const1_64( 0xE0E1E2E3E4E5E6E7 );
-   ctx->H[13] = m512_const1_64( 0xE8E9EAEBECEDEEEF );
-   ctx->H[14] = m512_const1_64( 0xF0F1F2F3F4F5F6F7 );
-   ctx->H[15] = m512_const1_64( 0xF8F9FAFBFCFDFEFF );
+   ctx->H[ 0] = _mm512_set1_epi64( 0x8081828384858687 );
+   ctx->H[ 1] = _mm512_set1_epi64( 0x88898A8B8C8D8E8F );
+   ctx->H[ 2] = _mm512_set1_epi64( 0x9091929394959697 );
+   ctx->H[ 3] = _mm512_set1_epi64( 0x98999A9B9C9D9E9F );
+   ctx->H[ 4] = _mm512_set1_epi64( 0xA0A1A2A3A4A5A6A7 );
+   ctx->H[ 5] = _mm512_set1_epi64( 0xA8A9AAABACADAEAF );
+   ctx->H[ 6] = _mm512_set1_epi64( 0xB0B1B2B3B4B5B6B7 );
+   ctx->H[ 7] = _mm512_set1_epi64( 0xB8B9BABBBCBDBEBF );
+   ctx->H[ 8] = _mm512_set1_epi64( 0xC0C1C2C3C4C5C6C7 );
+   ctx->H[ 9] = _mm512_set1_epi64( 0xC8C9CACBCCCDCECF );
+   ctx->H[10] = _mm512_set1_epi64( 0xD0D1D2D3D4D5D6D7 );
+   ctx->H[11] = _mm512_set1_epi64( 0xD8D9DADBDCDDDEDF );
+   ctx->H[12] = _mm512_set1_epi64( 0xE0E1E2E3E4E5E6E7 );
+   ctx->H[13] = _mm512_set1_epi64( 0xE8E9EAEBECEDEEEF );
+   ctx->H[14] = _mm512_set1_epi64( 0xF0F1F2F3F4F5F6F7 );
+   ctx->H[15] = _mm512_set1_epi64( 0xF8F9FAFBFCFDFEFF );
   ctx->ptr = 0;
   ctx->bit_count = 0;
 }
@@ -1461,7 +1448,7 @@ void bmw512_8way_close( bmw512_8way_context *ctx, void *dst )

   buf = ctx->buf;
   ptr = ctx->ptr;
-   buf[ ptr>>3 ] = m512_const1_64( 0x80 );
+   buf[ ptr>>3 ] = _mm512_set1_epi64( 0x80 );
   ptr += 8;
   h = ctx->H;

@@ -1496,22 +1483,22 @@ void bmw512_8way_full( bmw512_8way_context *ctx, void *out, const void *data,

 // Init

-   H[ 0] = m512_const1_64( 0x8081828384858687 );
-   H[ 1] = m512_const1_64( 0x88898A8B8C8D8E8F );
-   H[ 2] = m512_const1_64( 0x9091929394959697 );
-   H[ 3] = m512_const1_64( 0x98999A9B9C9D9E9F );
-   H[ 4] = m512_const1_64( 0xA0A1A2A3A4A5A6A7 );
-   H[ 5] = m512_const1_64( 0xA8A9AAABACADAEAF );
-   H[ 6] = m512_const1_64( 0xB0B1B2B3B4B5B6B7 );
-   H[ 7] = m512_const1_64( 0xB8B9BABBBCBDBEBF );
-   H[ 8] = m512_const1_64( 0xC0C1C2C3C4C5C6C7 );
-   H[ 9] = m512_const1_64( 0xC8C9CACBCCCDCECF );
-   H[10] = m512_const1_64( 0xD0D1D2D3D4D5D6D7 );
-   H[11] = m512_const1_64( 0xD8D9DADBDCDDDEDF );
-   H[12] = m512_const1_64( 0xE0E1E2E3E4E5E6E7 );
-   H[13] = m512_const1_64( 0xE8E9EAEBECEDEEEF );
-   H[14] = m512_const1_64( 0xF0F1F2F3F4F5F6F7 );
-   H[15] = m512_const1_64( 0xF8F9FAFBFCFDFEFF );
+   H[ 0] = _mm512_set1_epi64( 0x8081828384858687 );
+   H[ 1] = _mm512_set1_epi64( 0x88898A8B8C8D8E8F );
+   H[ 2] = _mm512_set1_epi64( 0x9091929394959697 );
+   H[ 3] = _mm512_set1_epi64( 0x98999A9B9C9D9E9F );
+   H[ 4] = _mm512_set1_epi64( 0xA0A1A2A3A4A5A6A7 );
+   H[ 5] = _mm512_set1_epi64( 0xA8A9AAABACADAEAF );
+   H[ 6] = _mm512_set1_epi64( 0xB0B1B2B3B4B5B6B7 );
+   H[ 7] = _mm512_set1_epi64( 0xB8B9BABBBCBDBEBF );
+   H[ 8] = _mm512_set1_epi64( 0xC0C1C2C3C4C5C6C7 );
+   H[ 9] = _mm512_set1_epi64( 0xC8C9CACBCCCDCECF );
+   H[10] = _mm512_set1_epi64( 0xD0D1D2D3D4D5D6D7 );
+   H[11] = _mm512_set1_epi64( 0xD8D9DADBDCDDDEDF );
+   H[12] = _mm512_set1_epi64( 0xE0E1E2E3E4E5E6E7 );
+   H[13] = _mm512_set1_epi64( 0xE8E9EAEBECEDEEEF );
+   H[14] = _mm512_set1_epi64( 0xF0F1F2F3F4F5F6F7 );
+   H[15] = _mm512_set1_epi64( 0xF8F9FAFBFCFDFEFF );

 // Update

@@ -1543,7 +1530,7 @@ void bmw512_8way_full( bmw512_8way_context *ctx, void *out, const void *data,
   __m512i h1[16], h2[16];
   size_t u, v;

-   buf[ ptr>>3 ] = m512_const1_64( 0x80 );
+   buf[ ptr>>3 ] = _mm512_set1_epi64( 0x80 );
   ptr += 8;

   if (  ptr > (buf_size - 8) )
--- a/algo/bmw/sph_bmw.h
+++ b/algo/bmw/sph_bmw.h
@@ -41,7 +41,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for BMW-224.
--- a/algo/cubehash/cube-hash-2way.c
+++ b/algo/cubehash/cube-hash-2way.c
@@ -221,14 +221,14 @@ int cube_4way_init( cube_4way_context *sp, int hashbitlen, int rounds,
    sp->rounds    = rounds;
    sp->pos       = 0;

-    h[ 0] = m512_const1_128( iv[0] );
-    h[ 1] = m512_const1_128( iv[1] );
-    h[ 2] = m512_const1_128( iv[2] );
-    h[ 3] = m512_const1_128( iv[3] );
-    h[ 4] = m512_const1_128( iv[4] );
-    h[ 5] = m512_const1_128( iv[5] );
-    h[ 6] = m512_const1_128( iv[6] );
-    h[ 7] = m512_const1_128( iv[7] );
+    h[ 0] = mm512_bcast_m128( iv[0] );
+    h[ 1] = mm512_bcast_m128( iv[1] );
+    h[ 2] = mm512_bcast_m128( iv[2] );
+    h[ 3] = mm512_bcast_m128( iv[3] );
+    h[ 4] = mm512_bcast_m128( iv[4] );
+    h[ 5] = mm512_bcast_m128( iv[5] );
+    h[ 6] = mm512_bcast_m128( iv[6] );
+    h[ 7] = mm512_bcast_m128( iv[7] );

    return 0;
 }
@@ -259,11 +259,11 @@ int cube_4way_close( cube_4way_context *sp, void *output )

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm512_xor_si512( sp->h[ sp->pos ],
-                                 m512_const2_64( 0, 0x0000000000000080 ) );
+                         mm512_bcast128lo_64( 0x0000000000000080 ) );
    transform_4way( sp );

    sp->h[7] = _mm512_xor_si512( sp->h[7],
-                                 m512_const2_64( 0x0000000100000000, 0 ) );
+                         mm512_bcast128hi_64( 0x0000000100000000 ) );

    for ( i = 0; i < 10; ++i ) 
       transform_4way( sp );
@@ -283,14 +283,14 @@ int cube_4way_full( cube_4way_context *sp, void *output,  int hashbitlen,
    sp->rounds    = 16;
    sp->pos       = 0;

-    h[ 0] = m512_const1_128( iv[0] );
-    h[ 1] = m512_const1_128( iv[1] );
-    h[ 2] = m512_const1_128( iv[2] );
-    h[ 3] = m512_const1_128( iv[3] );
-    h[ 4] = m512_const1_128( iv[4] );
-    h[ 5] = m512_const1_128( iv[5] );
-    h[ 6] = m512_const1_128( iv[6] );
-    h[ 7] = m512_const1_128( iv[7] );
+    h[ 0] = mm512_bcast_m128( iv[0] );
+    h[ 1] = mm512_bcast_m128( iv[1] );
+    h[ 2] = mm512_bcast_m128( iv[2] );
+    h[ 3] = mm512_bcast_m128( iv[3] );
+    h[ 4] = mm512_bcast_m128( iv[4] );
+    h[ 5] = mm512_bcast_m128( iv[5] );
+    h[ 6] = mm512_bcast_m128( iv[6] );
+    h[ 7] = mm512_bcast_m128( iv[7] );

    const int len = size >> 4;
    const __m512i *in = (__m512i*)data;
@@ -310,11 +310,11 @@ int cube_4way_full( cube_4way_context *sp, void *output,  int hashbitlen,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm512_xor_si512( sp->h[ sp->pos ],
-                                    m512_const2_64( 0, 0x0000000000000080 ) );
+                         mm512_bcast128lo_64( 0x0000000000000080 ) );
    transform_4way( sp );

    sp->h[7] = _mm512_xor_si512( sp->h[7],
-                                    m512_const2_64( 0x0000000100000000, 0 ) );
+                         mm512_bcast128hi_64( 0x0000000100000000 ) );

    for ( i = 0; i < 10; ++i )
       transform_4way( sp );
@@ -336,14 +336,14 @@ int cube_4way_2buf_full( cube_4way_2buf_context *sp,
    sp->rounds    = 16;
    sp->pos       = 0;

-    h1[0] = h0[0] = m512_const1_128( iv[0] );
-    h1[1] = h0[1] = m512_const1_128( iv[1] );
-    h1[2] = h0[2] = m512_const1_128( iv[2] );
-    h1[3] = h0[3] = m512_const1_128( iv[3] );
-    h1[4] = h0[4] = m512_const1_128( iv[4] );
-    h1[5] = h0[5] = m512_const1_128( iv[5] );
-    h1[6] = h0[6] = m512_const1_128( iv[6] );
-    h1[7] = h0[7] = m512_const1_128( iv[7] );
+    h1[0] = h0[0] = mm512_bcast_m128( iv[0] );
+    h1[1] = h0[1] = mm512_bcast_m128( iv[1] );
+    h1[2] = h0[2] = mm512_bcast_m128( iv[2] );
+    h1[3] = h0[3] = mm512_bcast_m128( iv[3] );
+    h1[4] = h0[4] = mm512_bcast_m128( iv[4] );
+    h1[5] = h0[5] = mm512_bcast_m128( iv[5] );
+    h1[6] = h0[6] = mm512_bcast_m128( iv[6] );
+    h1[7] = h0[7] = mm512_bcast_m128( iv[7] );

    const int len = size >> 4;
    const __m512i *in0 = (__m512i*)data0;
@@ -365,13 +365,13 @@ int cube_4way_2buf_full( cube_4way_2buf_context *sp,
    }

    // pos is zero for 64 byte data, 1 for 80 byte data.
-    __m512i tmp = m512_const2_64( 0, 0x0000000000000080 );
+    __m512i tmp = mm512_bcast128lo_64( 0x0000000000000080 );
    sp->h0[ sp->pos ] = _mm512_xor_si512( sp->h0[ sp->pos ], tmp );
    sp->h1[ sp->pos ] = _mm512_xor_si512( sp->h1[ sp->pos ], tmp );

    transform_4way_2buf( sp );

-    tmp = m512_const2_64( 0x0000000100000000, 0 );
+    tmp = mm512_bcast128hi_64( 0x0000000100000000 );
    sp->h0[7] = _mm512_xor_si512( sp->h0[7], tmp );
    sp->h1[7] = _mm512_xor_si512( sp->h1[7], tmp );

@@ -384,7 +384,6 @@ int cube_4way_2buf_full( cube_4way_2buf_context *sp,
    return 0;
 }

-
 int cube_4way_update_close( cube_4way_context *sp, void *output,
                               const void *data, size_t size )
 {
@@ -406,11 +405,11 @@ int cube_4way_update_close( cube_4way_context *sp, void *output,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm512_xor_si512( sp->h[ sp->pos ],
-                                    m512_const2_64( 0, 0x0000000000000080 ) );
+                          mm512_bcast128lo_64( 0x0000000000000080 ) );
    transform_4way( sp );

    sp->h[7] = _mm512_xor_si512( sp->h[7],
-                                    m512_const2_64( 0x0000000100000000, 0 ) );
+                          mm512_bcast128hi_64( 0x0000000100000000 ) );

    for ( i = 0; i < 10; ++i )
       transform_4way( sp );
@@ -424,21 +423,6 @@ int cube_4way_update_close( cube_4way_context *sp, void *output,

 // 2 way 128 

-// This isn't expected to be used with AVX512 so HW rotate intruction
-// is assumed not avaiable.
-// Use double buffering to optimize serial bit rotations. Full double
-// buffering isn't practical because it needs twice as many registers
-// with AVX2 having only half as many as AVX512.
-#define ROL2( out0, out1, in0, in1, c ) \
-{ \
- __m256i t0 = _mm256_slli_epi32( in0, c ); \
- __m256i t1 = _mm256_slli_epi32( in1, c ); \
- out0 = _mm256_srli_epi32( in0, 32-(c) ); \
- out1 = _mm256_srli_epi32( in1, 32-(c) ); \
- out0 = _mm256_or_si256( out0, t0 ); \
- out1 = _mm256_or_si256( out1, t1 ); \
-}
-
 static void transform_2way( cube_2way_context *sp )
 {
    int r;
@@ -461,8 +445,10 @@ static void transform_2way( cube_2way_context *sp )
        x5 = _mm256_add_epi32( x1, x5 );
        x6 = _mm256_add_epi32( x2, x6 );
        x7 = _mm256_add_epi32( x3, x7 );
-        ROL2( y0, y1, x2, x3, 7 );
-        ROL2( x2, x3, x0, x1, 7 );
+        y0 = mm256_rol_32( x2, 7 );
+        y1 = mm256_rol_32( x3, 7 );
+        x2 = mm256_rol_32( x0, 7 );
+        x3 = mm256_rol_32( x1, 7 );
        x0 = _mm256_xor_si256( y0, x4 );
        x1 = _mm256_xor_si256( y1, x5 );
        x2 = _mm256_xor_si256( x2, x6 );
@@ -475,8 +461,10 @@ static void transform_2way( cube_2way_context *sp )
        x5 = _mm256_add_epi32( x1, x5 );
        x6 = _mm256_add_epi32( x2, x6 );
        x7 = _mm256_add_epi32( x3, x7 );
-        ROL2( y0, x1, x1, x0, 11 );
-        ROL2( y1, x3, x3, x2, 11 );
+        y0 = mm256_rol_32( x1, 11 );
+        x1 = mm256_rol_32( x0, 11 );
+        y1 = mm256_rol_32( x3, 11 );
+        x3 = mm256_rol_32( x2, 11 );
        x0 = _mm256_xor_si256( y0, x4 );
        x1 = _mm256_xor_si256( x1, x5 );
        x2 = _mm256_xor_si256( y1, x6 );
@@ -508,14 +496,14 @@ int cube_2way_init( cube_2way_context *sp, int hashbitlen, int rounds,
    sp->rounds    = rounds;
    sp->pos       = 0;

-    h[ 0] = m256_const1_128( iv[0] );
-    h[ 1] = m256_const1_128( iv[1] );
-    h[ 2] = m256_const1_128( iv[2] );
-    h[ 3] = m256_const1_128( iv[3] );
-    h[ 4] = m256_const1_128( iv[4] );
-    h[ 5] = m256_const1_128( iv[5] );
-    h[ 6] = m256_const1_128( iv[6] );
-    h[ 7] = m256_const1_128( iv[7] );
+    h[ 0] = mm256_bcast_m128( iv[0] );
+    h[ 1] = mm256_bcast_m128( iv[1] );
+    h[ 2] = mm256_bcast_m128( iv[2] );
+    h[ 3] = mm256_bcast_m128( iv[3] );
+    h[ 4] = mm256_bcast_m128( iv[4] );
+    h[ 5] = mm256_bcast_m128( iv[5] );
+    h[ 6] = mm256_bcast_m128( iv[6] );
+    h[ 7] = mm256_bcast_m128( iv[7] );
    
    return 0;
 }
@@ -546,13 +534,14 @@ int cube_2way_close( cube_2way_context *sp, void *output )

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ],
-                                   m256_const2_64( 0, 0x0000000000000080 ) );
+                                   mm256_bcast128lo_64( 0x0000000000000080 ) );
    transform_2way( sp );

    sp->h[7] = _mm256_xor_si256( sp->h[7],
-                                   m256_const2_64( 0x0000000100000000, 0 ) );
+                                   mm256_bcast128hi_64( 0x0000000100000000 ) );

-    for ( i = 0; i < 10; ++i )           transform_2way( sp );
+    for ( i = 0; i < 10; ++i )  
+       transform_2way( sp );

    memcpy( hash, sp->h, sp->hashlen<<5 );
    return 0;
@@ -579,13 +568,14 @@ int cube_2way_update_close( cube_2way_context *sp, void *output,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ],
-                                    m256_const2_64( 0, 0x0000000000000080 ) );
+                                    mm256_bcast128lo_64( 0x0000000000000080 ) );
    transform_2way( sp );

    sp->h[7] = _mm256_xor_si256( sp->h[7],
-                                    m256_const2_64( 0x0000000100000000, 0 ) );
+                                    mm256_bcast128hi_64( 0x0000000100000000 ) );

-    for ( i = 0; i < 10; ++i )    transform_2way( sp );
+    for ( i = 0; i < 10; ++i )
+       transform_2way( sp );

    memcpy( hash, sp->h, sp->hashlen<<5 );
    return 0;
@@ -602,14 +592,14 @@ int cube_2way_full( cube_2way_context *sp, void *output, int hashbitlen,
    sp->rounds    = 16;
    sp->pos       = 0;

-    h[ 0] = m256_const1_128( iv[0] );
-    h[ 1] = m256_const1_128( iv[1] );
-    h[ 2] = m256_const1_128( iv[2] );
-    h[ 3] = m256_const1_128( iv[3] );
-    h[ 4] = m256_const1_128( iv[4] );
-    h[ 5] = m256_const1_128( iv[5] );
-    h[ 6] = m256_const1_128( iv[6] );
-    h[ 7] = m256_const1_128( iv[7] );
+    h[ 0] = mm256_bcast_m128( iv[0] );
+    h[ 1] = mm256_bcast_m128( iv[1] );
+    h[ 2] = mm256_bcast_m128( iv[2] );
+    h[ 3] = mm256_bcast_m128( iv[3] );
+    h[ 4] = mm256_bcast_m128( iv[4] );
+    h[ 5] = mm256_bcast_m128( iv[5] );
+    h[ 6] = mm256_bcast_m128( iv[6] );
+    h[ 7] = mm256_bcast_m128( iv[7] );

    const int len = size >> 4;
    const __m256i *in = (__m256i*)data;
@@ -629,13 +619,14 @@ int cube_2way_full( cube_2way_context *sp, void *output, int hashbitlen,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->h[ sp->pos ] = _mm256_xor_si256( sp->h[ sp->pos ],
-                                    m256_const2_64( 0, 0x0000000000000080 ) );
+                                    mm256_bcast128lo_64( 0x0000000000000080 ) );
    transform_2way( sp );

    sp->h[7] = _mm256_xor_si256( sp->h[7],
-                                    m256_const2_64( 0x0000000100000000, 0 ) );
+                                    mm256_bcast128hi_64( 0x0000000100000000 ) );

-    for ( i = 0; i < 10; ++i )    transform_2way( sp );
+    for ( i = 0; i < 10; ++i )
+       transform_2way( sp );

    memcpy( hash, sp->h, sp->hashlen<<5 );
    return 0;
--- a/algo/cubehash/cubehash_sse2.c
+++ b/algo/cubehash/cubehash_sse2.c
@@ -9,7 +9,6 @@
 #include <immintrin.h>
 #endif
 #include "cubehash_sse2.h"
-#include "algo/sha/sha3-defs.h"
 #include <stdbool.h>
 #include <unistd.h>
 #include <memory.h>
@@ -58,19 +57,18 @@ static void transform( cubehashParam *sp )
    { 
        x2 = _mm256_add_epi32( x0, x2 );
        x3 = _mm256_add_epi32( x1, x3 );
-        y0 = x0;
-        x0 = mm256_rol_32( x1, 7 );
-        x1 = mm256_rol_32( y0, 7 );
-        x0 = _mm256_xor_si256( x0, x2 );
-        x1 = _mm256_xor_si256( x1, x3 );
+        y0 = mm256_rol_32( x1, 7 );
+        y1 = mm256_rol_32( x0, 7 );
+        x0 = _mm256_xor_si256( y0, x2 );
+        x1 = _mm256_xor_si256( y1, x3 );
        x2 = mm256_swap128_64( x2 );
        x3 = mm256_swap128_64( x3 );
        x2 = _mm256_add_epi32( x0, x2 );
        x3 = _mm256_add_epi32( x1, x3 );
-        y0 = mm256_swap_128( x0 );
-        y1 = mm256_swap_128( x1 );
-        x0 = mm256_rol_32( y0, 11 );
-        x1 = mm256_rol_32( y1, 11 );
+        x0 = mm256_swap_128( x0 );
+        x1 = mm256_swap_128( x1 );
+        x0 = mm256_rol_32( x0, 11 );
+        x1 = mm256_rol_32( x1, 11 );
        x0 = _mm256_xor_si256( x0, x2 );
        x1 = _mm256_xor_si256( x1, x3 );
        x2 = mm256_swap64_32( x2 );
@@ -94,47 +92,48 @@ static void transform( cubehashParam *sp )
    x6 = _mm_load_si128( (__m128i*)sp->x + 6 );
    x7 = _mm_load_si128( (__m128i*)sp->x + 7 );

-    for (r = 0; r < rounds; ++r) {
-	x4 = _mm_add_epi32(x0, x4);
-	x5 = _mm_add_epi32(x1, x5);
-	x6 = _mm_add_epi32(x2, x6);
-	x7 = _mm_add_epi32(x3, x7);
+    for ( r = 0; r < rounds; ++r )
+    {
+       x4 = _mm_add_epi32( x0, x4 );
+       x5 = _mm_add_epi32( x1, x5 );
+       x6 = _mm_add_epi32( x2, x6 );
+       x7 = _mm_add_epi32( x3, x7 );
       y0 = x2;
       y1 = x3;
       y2 = x0;
       y3 = x1;
-	x0 = _mm_xor_si128(_mm_slli_epi32(y0, 7), _mm_srli_epi32(y0, 25));
-	x1 = _mm_xor_si128(_mm_slli_epi32(y1, 7), _mm_srli_epi32(y1, 25));
-	x2 = _mm_xor_si128(_mm_slli_epi32(y2, 7), _mm_srli_epi32(y2, 25));
-	x3 = _mm_xor_si128(_mm_slli_epi32(y3, 7), _mm_srli_epi32(y3, 25));
-	x0 = _mm_xor_si128(x0, x4);
-	x1 = _mm_xor_si128(x1, x5);
-	x2 = _mm_xor_si128(x2, x6);
-	x3 = _mm_xor_si128(x3, x7);
-	x4 = _mm_shuffle_epi32(x4, 0x4e);
-	x5 = _mm_shuffle_epi32(x5, 0x4e);
-	x6 = _mm_shuffle_epi32(x6, 0x4e);
-	x7 = _mm_shuffle_epi32(x7, 0x4e);
-	x4 = _mm_add_epi32(x0, x4);
-	x5 = _mm_add_epi32(x1, x5);
-	x6 = _mm_add_epi32(x2, x6);
-	x7 = _mm_add_epi32(x3, x7);
+       x0 = mm128_rol_32( y0, 7 );
+       x1 = mm128_rol_32( y1, 7 );
+       x2 = mm128_rol_32( y2, 7 );
+       x3 = mm128_rol_32( y3, 7 );
+       x0 = _mm_xor_si128( x0, x4 );
+       x1 = _mm_xor_si128( x1, x5 );
+       x2 = _mm_xor_si128( x2, x6 );
+       x3 = _mm_xor_si128( x3, x7 );
+       x4 = _mm_shuffle_epi32( x4, 0x4e );
+       x5 = _mm_shuffle_epi32( x5, 0x4e );
+       x6 = _mm_shuffle_epi32( x6, 0x4e );
+       x7 = _mm_shuffle_epi32( x7, 0x4e );
+       x4 = _mm_add_epi32( x0, x4 );
+       x5 = _mm_add_epi32( x1, x5 );
+       x6 = _mm_add_epi32( x2, x6 );
+       x7 = _mm_add_epi32( x3, x7 );
       y0 = x1;
       y1 = x0;
       y2 = x3;
       y3 = x2;
-	x0 = _mm_xor_si128(_mm_slli_epi32(y0, 11), _mm_srli_epi32(y0, 21));
-	x1 = _mm_xor_si128(_mm_slli_epi32(y1, 11), _mm_srli_epi32(y1, 21));
-	x2 = _mm_xor_si128(_mm_slli_epi32(y2, 11), _mm_srli_epi32(y2, 21));
-	x3 = _mm_xor_si128(_mm_slli_epi32(y3, 11), _mm_srli_epi32(y3, 21));
-	x0 = _mm_xor_si128(x0, x4);
-	x1 = _mm_xor_si128(x1, x5);
-	x2 = _mm_xor_si128(x2, x6);
-	x3 = _mm_xor_si128(x3, x7);
-	x4 = _mm_shuffle_epi32(x4, 0xb1);
-	x5 = _mm_shuffle_epi32(x5, 0xb1);
-	x6 = _mm_shuffle_epi32(x6, 0xb1);
-	x7 = _mm_shuffle_epi32(x7, 0xb1);
+       x0 = mm128_rol_32( y0, 11 );
+       x1 = mm128_rol_32( y1, 11 );
+       x2 = mm128_rol_32( y2, 11 );
+       x3 = mm128_rol_32( y3, 11 );
+	    x0 = _mm_xor_si128( x0, x4 );
+	    x1 = _mm_xor_si128( x1, x5 );
+	    x2 = _mm_xor_si128( x2, x6 );
+	    x3 = _mm_xor_si128( x3, x7 );
+	    x4 = _mm_shuffle_epi32( x4, 0xb1 );
+	    x5 = _mm_shuffle_epi32( x5, 0xb1 );
+	    x6 = _mm_shuffle_epi32( x6, 0xb1 );
+	    x7 = _mm_shuffle_epi32( x7, 0xb1 );
    }

    _mm_store_si128( (__m128i*)sp->x,     x0 );
@@ -180,25 +179,25 @@ int cubehashInit(cubehashParam *sp, int hashbitlen, int rounds, int blockbytes)
    if ( hashbitlen == 512 )
    {

-       x[0] = m128_const_64( 0x4167D83E2D538B8B, 0x50F494D42AEA2A61 );
-       x[1] = m128_const_64( 0x50AC5695CC39968E, 0xC701CF8C3FEE2313 );
-       x[2] = m128_const_64( 0x825B453797CF0BEF, 0xA647A8B34D42C787 );
-       x[3] = m128_const_64( 0xA23911AED0E5CD33, 0xF22090C4EEF864D2 );
-       x[4] = m128_const_64( 0xB64445321B017BEF, 0x148FE485FCD398D9 );
-       x[5] = m128_const_64( 0x0DBADEA991FA7934, 0x2FF5781C6A536159 );
-       x[6] = m128_const_64( 0xBC796576B1C62456, 0xA5A70E75D65C8A2B );
-       x[7] = m128_const_64( 0xD43E3B447795D246, 0xE7989AF11921C8F7 );
+       x[0] = _mm_set_epi64x( 0x4167D83E2D538B8B, 0x50F494D42AEA2A61 );
+       x[1] = _mm_set_epi64x( 0x50AC5695CC39968E, 0xC701CF8C3FEE2313 );
+       x[2] = _mm_set_epi64x( 0x825B453797CF0BEF, 0xA647A8B34D42C787 );
+       x[3] = _mm_set_epi64x( 0xA23911AED0E5CD33, 0xF22090C4EEF864D2 );
+       x[4] = _mm_set_epi64x( 0xB64445321B017BEF, 0x148FE485FCD398D9 );
+       x[5] = _mm_set_epi64x( 0x0DBADEA991FA7934, 0x2FF5781C6A536159 );
+       x[6] = _mm_set_epi64x( 0xBC796576B1C62456, 0xA5A70E75D65C8A2B );
+       x[7] = _mm_set_epi64x( 0xD43E3B447795D246, 0xE7989AF11921C8F7 );
    }
    else
    {
-       x[0] = m128_const_64( 0x35481EAE63117E71, 0xCCD6F29FEA2BD4B4 );
-       x[1] = m128_const_64( 0xF4CC12BE7E624131, 0xE5D94E6322512D5B );
-       x[2] = m128_const_64( 0x3361DA8CD0720C35, 0x42AF2070C2D0B696 );
-       x[3] = m128_const_64( 0x40E5FBAB4680AC00, 0x8EF8AD8328CCECA4 );
-       x[4] = m128_const_64( 0xF0B266796C859D41, 0x6107FBD5D89041C3 );
-       x[5] = m128_const_64( 0x93CB628565C892FD, 0x5FA2560309392549 );
-       x[6] = m128_const_64( 0x85254725774ABFDD, 0x9E4B4E602AF2B5AE );
-       x[7] = m128_const_64( 0xD6032C0A9CDAF8AF, 0x4AB6AAD615815AEB );
+       x[0] = _mm_set_epi64x( 0x35481EAE63117E71, 0xCCD6F29FEA2BD4B4 );
+       x[1] = _mm_set_epi64x( 0xF4CC12BE7E624131, 0xE5D94E6322512D5B );
+       x[2] = _mm_set_epi64x( 0x3361DA8CD0720C35, 0x42AF2070C2D0B696 );
+       x[3] = _mm_set_epi64x( 0x40E5FBAB4680AC00, 0x8EF8AD8328CCECA4 );
+       x[4] = _mm_set_epi64x( 0xF0B266796C859D41, 0x6107FBD5D89041C3 );
+       x[5] = _mm_set_epi64x( 0x93CB628565C892FD, 0x5FA2560309392549 );
+       x[6] = _mm_set_epi64x( 0x85254725774ABFDD, 0x9E4B4E602AF2B5AE );
+       x[7] = _mm_set_epi64x( 0xD6032C0A9CDAF8AF, 0x4AB6AAD615815AEB );
    }   

    return SUCCESS;
@@ -234,10 +233,10 @@ int cubehashDigest( cubehashParam *sp, byte *digest )

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->x[ sp->pos ] = _mm_xor_si128( sp->x[ sp->pos ],
-                                      m128_const_64( 0, 0x80 ) );
+                                      _mm_set_epi64x( 0, 0x80 ) );
    transform( sp );

-    sp->x[7] = _mm_xor_si128( sp->x[7], m128_const_64( 0x100000000, 0 ) );
+    sp->x[7] = _mm_xor_si128( sp->x[7], _mm_set_epi64x( 0x100000000, 0 ) );
    transform( sp );
    transform( sp );
    transform( sp );
@@ -279,10 +278,10 @@ int cubehashUpdateDigest( cubehashParam *sp, byte *digest,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->x[ sp->pos ] = _mm_xor_si128( sp->x[ sp->pos ],
-                                      m128_const_64( 0, 0x80 ) );
+                                      _mm_set_epi64x( 0, 0x80 ) );
    transform( sp );

-    sp->x[7] = _mm_xor_si128( sp->x[7], m128_const_64( 0x100000000, 0 ) );
+    sp->x[7] = _mm_xor_si128( sp->x[7], _mm_set_epi64x( 0x100000000, 0 ) );

    transform( sp );
    transform( sp );
@@ -313,25 +312,25 @@ int cubehash_full( cubehashParam *sp, byte *digest, int hashbitlen,
    if ( hashbitlen == 512 )
    {

-       x[0] = m128_const_64( 0x4167D83E2D538B8B, 0x50F494D42AEA2A61 );
-       x[1] = m128_const_64( 0x50AC5695CC39968E, 0xC701CF8C3FEE2313 );
-       x[2] = m128_const_64( 0x825B453797CF0BEF, 0xA647A8B34D42C787 );
-       x[3] = m128_const_64( 0xA23911AED0E5CD33, 0xF22090C4EEF864D2 );
-       x[4] = m128_const_64( 0xB64445321B017BEF, 0x148FE485FCD398D9 );
-       x[5] = m128_const_64( 0x0DBADEA991FA7934, 0x2FF5781C6A536159 );
-       x[6] = m128_const_64( 0xBC796576B1C62456, 0xA5A70E75D65C8A2B );
-       x[7] = m128_const_64( 0xD43E3B447795D246, 0xE7989AF11921C8F7 );
+       x[0] = _mm_set_epi64x( 0x4167D83E2D538B8B, 0x50F494D42AEA2A61 );
+       x[1] = _mm_set_epi64x( 0x50AC5695CC39968E, 0xC701CF8C3FEE2313 );
+       x[2] = _mm_set_epi64x( 0x825B453797CF0BEF, 0xA647A8B34D42C787 );
+       x[3] = _mm_set_epi64x( 0xA23911AED0E5CD33, 0xF22090C4EEF864D2 );
+       x[4] = _mm_set_epi64x( 0xB64445321B017BEF, 0x148FE485FCD398D9 );
+       x[5] = _mm_set_epi64x( 0x0DBADEA991FA7934, 0x2FF5781C6A536159 );
+       x[6] = _mm_set_epi64x( 0xBC796576B1C62456, 0xA5A70E75D65C8A2B );
+       x[7] = _mm_set_epi64x( 0xD43E3B447795D246, 0xE7989AF11921C8F7 );
    }
    else
    {
-       x[0] = m128_const_64( 0x35481EAE63117E71, 0xCCD6F29FEA2BD4B4 );
-       x[1] = m128_const_64( 0xF4CC12BE7E624131, 0xE5D94E6322512D5B );
-       x[2] = m128_const_64( 0x3361DA8CD0720C35, 0x42AF2070C2D0B696 );
-       x[3] = m128_const_64( 0x40E5FBAB4680AC00, 0x8EF8AD8328CCECA4 );
-       x[4] = m128_const_64( 0xF0B266796C859D41, 0x6107FBD5D89041C3 );
-       x[5] = m128_const_64( 0x93CB628565C892FD, 0x5FA2560309392549 );
-       x[6] = m128_const_64( 0x85254725774ABFDD, 0x9E4B4E602AF2B5AE );
-       x[7] = m128_const_64( 0xD6032C0A9CDAF8AF, 0x4AB6AAD615815AEB );
+       x[0] = _mm_set_epi64x( 0x35481EAE63117E71, 0xCCD6F29FEA2BD4B4 );
+       x[1] = _mm_set_epi64x( 0xF4CC12BE7E624131, 0xE5D94E6322512D5B );
+       x[2] = _mm_set_epi64x( 0x3361DA8CD0720C35, 0x42AF2070C2D0B696 );
+       x[3] = _mm_set_epi64x( 0x40E5FBAB4680AC00, 0x8EF8AD8328CCECA4 );
+       x[4] = _mm_set_epi64x( 0xF0B266796C859D41, 0x6107FBD5D89041C3 );
+       x[5] = _mm_set_epi64x( 0x93CB628565C892FD, 0x5FA2560309392549 );
+       x[6] = _mm_set_epi64x( 0x85254725774ABFDD, 0x9E4B4E602AF2B5AE );
+       x[7] = _mm_set_epi64x( 0xD6032C0A9CDAF8AF, 0x4AB6AAD615815AEB );
    }


@@ -358,10 +357,10 @@ int cubehash_full( cubehashParam *sp, byte *digest, int hashbitlen,

    // pos is zero for 64 byte data, 1 for 80 byte data.
    sp->x[ sp->pos ] = _mm_xor_si128( sp->x[ sp->pos ],
-                                      m128_const_64( 0, 0x80 ) );
+                                      _mm_set_epi64x( 0, 0x80 ) );
    transform( sp );

-    sp->x[7] = _mm_xor_si128( sp->x[7], m128_const_64( 0x100000000, 0 ) );
+    sp->x[7] = _mm_xor_si128( sp->x[7], _mm_set_epi64x( 0x100000000, 0 ) );

    transform( sp );
    transform( sp );
--- a/algo/cubehash/cubehash_sse2.h
+++ b/algo/cubehash/cubehash_sse2.h
@@ -3,7 +3,7 @@

 #include "compat.h"
 #include <stdint.h>
-#include "algo/sha/sha3-defs.h"
+#include "compat/sha3-defs.h"

 #define	OPTIMIZE_SSE2

--- a/algo/cubehash/sph_cubehash.h
+++ b/algo/cubehash/sph_cubehash.h
@@ -42,7 +42,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for CubeHash-224.
--- a/algo/echo/aes_ni/hash.c
+++ b/algo/echo/aes_ni/hash.c
@@ -566,16 +566,16 @@ HashReturn echo_full( hashState_echo *state, BitSequence *hashval,
         state->uHashSize = 256;
         state->uBlockLength = 192;
         state->uRounds = 8;
-         state->hashsize = m128_const_64( 0, 0x100 );
-         state->const1536 = m128_const_64( 0, 0x600 );
+         state->hashsize = _mm_set_epi64x( 0, 0x100 );
+         state->const1536 = _mm_set_epi64x( 0, 0x600 );
         break;

      case 512:
         state->uHashSize = 512;
         state->uBlockLength = 128;
         state->uRounds = 10;
-         state->hashsize = m128_const_64( 0, 0x200 );
-         state->const1536 = m128_const_64( 0, 0x400 );
+         state->hashsize = _mm_set_epi64x( 0, 0x200 );
+         state->const1536 = _mm_set_epi64x( 0, 0x400 );
         break;

      default:
--- a/algo/echo/aes_ni/hash_api.h
+++ b/algo/echo/aes_ni/hash_api.h
@@ -22,7 +22,7 @@
 #endif


-#include "algo/sha/sha3_common.h"
+#include "compat/sha3_common.h"

 #include <emmintrin.h>

--- a/algo/echo/echo-hash-4way.c
+++ b/algo/echo/echo-hash-4way.c
@@ -162,9 +162,9 @@ 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 ); 
+  const __m512i one = mm512_bcast128lo_64( 1 ); 
+  const __m512i mul2mask = mm512_bcast128lo_64( 0x00001b00 );
+  const __m512i lsbmask  = _mm512_set1_epi32( 0x01010101 ); 

  _state[ 0 ][ 0 ] = ctx->state[ 0 ][ 0 ];
  _state[ 0 ][ 1 ] = ctx->state[ 0 ][ 1 ];
@@ -264,16 +264,16 @@ int echo_4way_init( echo_4way_context *ctx, int nHashSize )
 		ctx->uHashSize = 256;
 		ctx->uBlockLength = 192;
 		ctx->uRounds = 8;
-		ctx->hashsize = m512_const2_64( 0, 0x100 );
-		ctx->const1536 = m512_const2_64( 0, 0x600 );
+      ctx->hashsize = mm512_bcast128lo_64( 0x100 );
+      ctx->const1536 = mm512_bcast128lo_64( 0x600 );
 		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);
+      ctx->hashsize = mm512_bcast128lo_64( 0x200 );
+      ctx->const1536 = mm512_bcast128lo_64( 0x400);
 		break;

 	default:
@@ -305,7 +305,7 @@ int echo_4way_update_close( echo_4way_context *state, void *hashval,
   {
      echo_4way_compress( state, data, 1 );
      state->processed_bits = 1024;
-      remainingbits = m512_const2_64( 0, -1024 );
+      remainingbits = mm512_bcast128lo_64( -1024 );
      vlen = 0;
   }
   else
@@ -313,13 +313,15 @@ 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 = m512_const2_64( 0, (uint64_t)databitlen );
+      remainingbits = mm512_bcast128lo_64( (uint64_t)databitlen );
   }

-   state->buffer[ vlen ] = m512_const2_64( 0, 0x80 );
+   state->buffer[ vlen ] = mm512_bcast128lo_64( 0x80 );
   memset_zero_512( state->buffer + vlen + 1, vblen - vlen - 2 );
-   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->buffer[ vblen-2 ] =
+           mm512_bcast128hi_64( (uint64_t)state->uHashSize << 48 );
+   state->buffer[ vblen-1 ] =
+           mm512_bcast128lo_64( state->processed_bits );

   state->k = _mm512_add_epi64( state->k, remainingbits );
   state->k = _mm512_sub_epi64( state->k, state->const1536 );
@@ -352,16 +354,16 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
         ctx->uHashSize = 256;
         ctx->uBlockLength = 192;
         ctx->uRounds = 8;
-         ctx->hashsize = m512_const2_64( 0, 0x100 );
-         ctx->const1536 = m512_const2_64( 0, 0x600 );
+         ctx->hashsize = mm512_bcast128lo_64( 0x100 );
+         ctx->const1536 = mm512_bcast128lo_64( 0x600 );
         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 );
+         ctx->hashsize = mm512_bcast128lo_64( 0x200 );
+         ctx->const1536 = mm512_bcast128lo_64( 0x400 );
         break;

      default:
@@ -388,7 +390,7 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
   {
      echo_4way_compress( ctx, data, 1 );
      ctx->processed_bits = 1024;
-      remainingbits = m512_const2_64( 0, -1024 );
+      remainingbits = mm512_bcast128lo_64( -1024 );
      vlen = 0;
   }
   else
@@ -396,14 +398,14 @@ 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 = m512_const2_64( 0, databitlen );
+      remainingbits = mm512_bcast128lo_64( databitlen );
   }

-   ctx->buffer[ vlen ] = m512_const2_64( 0, 0x80 );
+   ctx->buffer[ vlen ] = mm512_bcast128lo_64( 0x80 );
   memset_zero_512( ctx->buffer + vlen + 1, vblen - vlen - 2 );
   ctx->buffer[ vblen-2 ] =
-                     m512_const2_64( (uint64_t)ctx->uHashSize << 48, 0 );
-   ctx->buffer[ vblen-1 ] = m512_const2_64( 0, ctx->processed_bits);
+               mm512_bcast128hi_64( (uint64_t)ctx->uHashSize << 48 );
+   ctx->buffer[ vblen-1 ] = mm512_bcast128lo_64( ctx->processed_bits);

   ctx->k = _mm512_add_epi64( ctx->k, remainingbits );
   ctx->k = _mm512_sub_epi64( ctx->k, ctx->const1536 );
@@ -425,9 +427,9 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,

 // AVX2 + VAES

-#define mul2mask_2way   m256_const2_64( 0, 0x0000000000001b00 ) 
+#define mul2mask_2way   mm256_bcast128lo_64( 0x0000000000001b00 ) 

-#define lsbmask_2way    m256_const1_32( 0x01010101 ) 
+#define lsbmask_2way    _mm256_set1_epi32( 0x01010101 ) 

 #define ECHO_SUBBYTES4_2WAY( state, j ) \
   state[0][j] = _mm256_aesenc_epi128( state[0][j], k1 ); \
@@ -467,8 +469,7 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
   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[ 0 ][ j ] = mm256_xor3( state2[ 0 ][ j ], 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 ] ); \
@@ -478,8 +479,7 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
   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[ 1 ][ j ] = mm256_xor3( state2[ 1 ][ j ], 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 ] ); \
@@ -489,8 +489,7 @@ int echo_4way_full( echo_4way_context *ctx, void *hashval, int nHashSize,
   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[ 2 ][ j ] = mm256_xor3( state2[ 2 ][ j ], s2, state1[ 3 ][ j3] ); \
   state2[ 3 ][ j ] = _mm256_xor_si256( state2[ 3 ][ j ], s2 ); \
 } while(0)

@@ -679,16 +678,16 @@ int echo_2way_init( echo_2way_context *ctx, int nHashSize )
                        ctx->uHashSize = 256;
                        ctx->uBlockLength = 192;
                        ctx->uRounds = 8;
-                        ctx->hashsize = m256_const2_64( 0, 0x100 );
-                        ctx->const1536 = m256_const2_64( 0, 0x600 );
+                        ctx->hashsize = mm256_bcast128lo_64( 0x100 );
+                        ctx->const1536 = mm256_bcast128lo_64( 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 );
+                        ctx->hashsize = mm256_bcast128lo_64( 0x200 );
+                        ctx->const1536 = mm256_bcast128lo_64( 0x400 );
                        break;

                default:
@@ -720,20 +719,20 @@ int echo_2way_update_close( echo_2way_context *state, void *hashval,
   {
      echo_2way_compress( state, data, 1 );
      state->processed_bits = 1024;
-      remainingbits = m256_const2_64( 0, -1024 );
+      remainingbits = mm256_bcast128lo_64( -1024 );
      vlen = 0;
   }
   else
   {
      memcpy_256( state->buffer, data, vlen );
      state->processed_bits += (unsigned int)( databitlen );
-      remainingbits = m256_const2_64( 0, databitlen );
+      remainingbits = mm256_bcast128lo_64( databitlen );
   }

-   state->buffer[ vlen ] = m256_const2_64( 0, 0x80 );
+   state->buffer[ vlen ] = mm256_bcast128lo_64( 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->buffer[ vblen-2 ] = mm256_bcast128hi_64( (uint64_t)state->uHashSize << 48 );
+   state->buffer[ vblen-1 ] = mm256_bcast128lo_64( state->processed_bits );

   state->k = _mm256_add_epi64( state->k, remainingbits );
   state->k = _mm256_sub_epi64( state->k, state->const1536 );
@@ -766,16 +765,16 @@ int echo_2way_full( echo_2way_context *ctx, void *hashval, int nHashSize,
         ctx->uHashSize = 256;
         ctx->uBlockLength = 192;
         ctx->uRounds = 8;
-         ctx->hashsize = m256_const2_64( 0, 0x100 );
-         ctx->const1536 = m256_const2_64( 0, 0x600 );
+         ctx->hashsize = mm256_bcast128lo_64( 0x100 );
+         ctx->const1536 = mm256_bcast128lo_64( 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 );
+         ctx->hashsize = mm256_bcast128lo_64( 0x200 );
+         ctx->const1536 = mm256_bcast128lo_64( 0x400 );
         break;

      default:
@@ -798,7 +797,7 @@ int echo_2way_full( echo_2way_context *ctx, void *hashval, int nHashSize,
   {
      echo_2way_compress( ctx, data, 1 );
      ctx->processed_bits = 1024;
-      remainingbits = m256_const2_64( 0, -1024 );
+      remainingbits = mm256_bcast128lo_64( -1024 );
      vlen = 0;
   }
   else
@@ -806,13 +805,13 @@ int echo_2way_full( echo_2way_context *ctx, void *hashval, int nHashSize,
      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 );
+      remainingbits = mm256_bcast128lo_64( databitlen );
   }

-   ctx->buffer[ vlen ] = m256_const2_64( 0, 0x80 );
+   ctx->buffer[ vlen ] = mm256_bcast128lo_64( 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->buffer[ vblen-2 ] = mm256_bcast128hi_64( (uint64_t)ctx->uHashSize << 48 );
+   ctx->buffer[ vblen-1 ] = mm256_bcast128lo_64( ctx->processed_bits );

   ctx->k = _mm256_add_epi64( ctx->k, remainingbits );
   ctx->k = _mm256_sub_epi64( ctx->k, ctx->const1536 );
--- a/algo/echo/sph_echo.c
+++ b/algo/echo/sph_echo.c
@@ -73,7 +73,7 @@ extern "C"{
 #endif

 #define AES_BIG_ENDIAN   0
-#include "algo/sha/aes_helper.c"
+#include "compat/aes_helper.c"

 #if SPH_ECHO_64

--- a/algo/echo/sph_echo.h
+++ b/algo/echo/sph_echo.h
@@ -43,7 +43,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for ECHO-224.
--- a/algo/fugue/fugue-aesni.c
+++ b/algo/fugue/fugue-aesni.c
@@ -33,11 +33,11 @@ MYALIGN const unsigned long long _supermix4b[]	= {0x07020d08080e0d0d, 0x07070908
 MYALIGN const unsigned long long _supermix4c[]	= {0x0706050403020000, 0x0302000007060504};
 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 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 long long _k_n[] = {0x4E4E4E4E4E4E4E4E, 0x1B1B1B1B0E0E0E0E};
+//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 _mul2mask[] = {0x1b1b0000, 0x00000000, 0x00000000, 0x00000000};
 MYALIGN const unsigned int _mul4mask[] = {0x2d361b00, 0x00000000, 0x00000000, 0x00000000};
--- a/algo/fugue/fugue-aesni.h
+++ b/algo/fugue/fugue-aesni.h
@@ -20,7 +20,7 @@
 #error "Unsupported configuration, AES needs SSE4.1. Compile without AES."
 #endif

-#include "algo/sha/sha3_common.h"
+#include "compat/sha3_common.h"
 #include "simd-utils.h"


--- a/algo/fugue/sph_fugue.h
+++ b/algo/fugue/sph_fugue.h
@@ -2,7 +2,7 @@
 #define SPH_FUGUE_H__

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 #ifdef __cplusplus
 extern "C"{
--- a/algo/gost/sph_gost.h
+++ b/algo/gost/sph_gost.h
@@ -41,7 +41,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for GOST-256.
--- a/algo/groestl/aes_ni/groestl-intr-aes.h
+++ b/algo/groestl/aes_ni/groestl-intr-aes.h
@@ -139,7 +139,7 @@ static const __m128i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003 };
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m128_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2(a0, b0, b1);\
  a0 = _mm_xor_si128(a0, TEMP0);\
  MUL2(a1, b0, b1);\
@@ -237,7 +237,7 @@ static const __m128i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003 };
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m128_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2(a0, b0, b1);\
  a0 = _mm_xor_si128(a0, TEMP0);\
  MUL2(a1, b0, b1);\
--- a/algo/groestl/aes_ni/groestl256-intr-aes.h
+++ b/algo/groestl/aes_ni/groestl256-intr-aes.h
@@ -128,7 +128,7 @@ static const __m128i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e };
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m128_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2(a0, b0, b1);\
  a0 = _mm_xor_si128(a0, TEMP0);\
  MUL2(a1, b0, b1);\
@@ -226,7 +226,7 @@ static const __m128i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e };
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m128_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2(a0, b0, b1);\
  a0 = _mm_xor_si128(a0, TEMP0);\
  MUL2(a1, b0, b1);\
@@ -275,7 +275,7 @@ static const __m128i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e };
 */
 #define ROUND(i, a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
  /* AddRoundConstant */\
-  b1 = m128_const_64( 0xffffffffffffffff, 0 ); \
+  b1 = _mm_set_epi64x( 0xffffffffffffffff, 0 ); \
  a0 = _mm_xor_si128( a0, casti_m128i( round_const_l0, i ) ); \
  a1 = _mm_xor_si128( a1, b1 ); \
  a2 = _mm_xor_si128( a2, b1 ); \
--- a/algo/groestl/aes_ni/hash-groestl.c
+++ b/algo/groestl/aes_ni/hash-groestl.c
@@ -24,9 +24,6 @@ HashReturn_gr init_groestl( hashState_groestl* ctx, int hashlen )

  ctx->hashlen = hashlen;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return FAIL_GR;
-
  for ( i = 0; i < SIZE512; i++ )
  {
     ctx->chaining[i] = _mm_setzero_si128();
@@ -34,7 +31,7 @@ HashReturn_gr init_groestl( hashState_groestl* ctx, int hashlen )
  }

  // The only non-zero in the IV is len. It can be hard coded.
-  ctx->chaining[ 6 ] = m128_const_64( 0x0200000000000000, 0 );
+  ctx->chaining[ 6 ] = _mm_set_epi64x( 0x0200000000000000, 0 );

  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;
@@ -46,15 +43,12 @@ HashReturn_gr reinit_groestl( hashState_groestl* ctx )
 {
  int i;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return FAIL_GR;
-
  for ( i = 0; i < SIZE512; i++ )
  {
     ctx->chaining[i] = _mm_setzero_si128();
     ctx->buffer[i]   = _mm_setzero_si128();
  }
-  ctx->chaining[ 6 ] = m128_const_64( 0x0200000000000000, 0 );
+  ctx->chaining[ 6 ] = _mm_set_epi64x( 0x0200000000000000, 0 );
  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;

@@ -122,7 +116,7 @@ HashReturn_gr final_groestl( hashState_groestl* ctx, void* output )
   else
   {
       // add first padding
-       ctx->buffer[rem_ptr] = m128_const_64( 0, 0x80 );
+       ctx->buffer[rem_ptr] = _mm_set_epi64x( 0, 0x80 );
       // add zero padding
       for ( i = rem_ptr + 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = _mm_setzero_si128();
@@ -154,7 +148,7 @@ int groestl512_full( hashState_groestl* ctx, void* output,
      ctx->chaining[i] = _mm_setzero_si128();
      ctx->buffer[i]   = _mm_setzero_si128();
   }
-   ctx->chaining[ 6 ] = m128_const_64( 0x0200000000000000, 0 );
+   ctx->chaining[ 6 ] = _mm_set_epi64x( 0x0200000000000000, 0 );
   ctx->buf_ptr = 0;

   // --- update ---
@@ -188,7 +182,7 @@ int groestl512_full( hashState_groestl* ctx, void* output,
   else
   {
       // add first padding
-       ctx->buffer[i] = m128_const_64( 0, 0x80 );
+       ctx->buffer[i] = _mm_set_epi64x( 0, 0x80 );
       // add zero padding
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = _mm_setzero_si128();
@@ -245,7 +239,7 @@ HashReturn_gr update_and_final_groestl( hashState_groestl* ctx, void* output,
   else
   {
       // add first padding
-       ctx->buffer[i] = m128_const_64( 0, 0x80 );
+       ctx->buffer[i] = _mm_set_epi64x( 0, 0x80 );
       // add zero padding
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = _mm_setzero_si128();
--- a/algo/groestl/aes_ni/hash-groestl.h
+++ b/algo/groestl/aes_ni/hash-groestl.h
@@ -20,8 +20,8 @@
 #define LENGTH (512)

 #include "brg_endian.h"
-#define NEED_UINT_64T
-#include "algo/sha/brg_types.h"
+//#define NEED_UINT_64T
+#include "compat/brg_types.h"

 /* some sizes (number of bytes) */
 #define ROWS (8)
--- a/algo/groestl/aes_ni/hash-groestl256.c
+++ b/algo/groestl/aes_ni/hash-groestl256.c
@@ -22,9 +22,6 @@ HashReturn_gr init_groestl256( hashState_groestl256* ctx, int hashlen )

  ctx->hashlen = hashlen;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return FAIL_GR;
-
  for ( i = 0; i < SIZE256; i++ )
  {
     ctx->chaining[i] = _mm_setzero_si128();
@@ -43,19 +40,14 @@ HashReturn_gr reinit_groestl256(hashState_groestl256* ctx)
 {
  int i;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return FAIL_GR;
-
  for ( i = 0; i < SIZE256; i++ )
  {
     ctx->chaining[i] = _mm_setzero_si128();
     ctx->buffer[i]   = _mm_setzero_si128();
  }

-  ctx->chaining[ 3 ] = m128_const_64( 0, 0x0100000000000000 );
+  ctx->chaining[ 3 ] = _mm_set_epi64x( 0, 0x0100000000000000 );

-//  ((u64*)ctx->chaining)[COLS-1] = U64BIG((u64)LENGTH);
-//  INIT256(ctx->chaining);
  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;

--- a/algo/groestl/aes_ni/hash-groestl256.h
+++ b/algo/groestl/aes_ni/hash-groestl256.h
@@ -34,8 +34,7 @@ typedef crypto_uint64 u64;
 //#define LENGTH (512)

 #include "brg_endian.h"
-#define NEED_UINT_64T
-#include "algo/sha/brg_types.h"
+#include "compat/brg_types.h"

 #ifdef IACA_TRACE
  #include IACA_MARKS
--- a/algo/groestl/groestl-gate.c
+++ b/algo/groestl/groestl-gate.c
@@ -17,7 +17,7 @@ bool register_dmd_gr_algo( algo_gate_t *gate )
 bool register_groestl_algo( algo_gate_t* gate )
 {
    register_dmd_gr_algo( gate );
-    gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
+    gate->gen_merkle_root = (void*)&sha256_gen_merkle_root;
    return true;
 };

--- a/algo/groestl/groestl256-hash-4way.c
+++ b/algo/groestl/groestl256-hash-4way.c
@@ -26,9 +26,6 @@ int groestl256_4way_init( groestl256_4way_context* ctx, uint64_t hashlen )

  ctx->hashlen = hashlen;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return 1;
-
  for ( i = 0; i < SIZE256; i++ )
  {
     ctx->chaining[i] = m512_zero;
@@ -36,8 +33,7 @@ int groestl256_4way_init( groestl256_4way_context* ctx, uint64_t hashlen )
  }

  // The only non-zero in the IV is len. It can be hard coded.
-  ctx->chaining[ 3 ] = m512_const2_64( 0, 0x0100000000000000 );
-
+  ctx->chaining[ 3 ] = mm512_bcast128lo_64( 0x0100000000000000 );
  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;

@@ -54,9 +50,6 @@ int groestl256_4way_full( groestl256_4way_context* ctx, void* output,
   __m512i* in = (__m512i*)input;
   int i;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return 1;
-
  for ( i = 0; i < SIZE256; i++ )
  {
     ctx->chaining[i] = m512_zero;
@@ -64,7 +57,7 @@ int groestl256_4way_full( groestl256_4way_context* ctx, void* output,
  }

  // The only non-zero in the IV is len. It can be hard coded.
-  ctx->chaining[ 3 ] = m512_const2_64( 0, 0x0100000000000000 );
+  ctx->chaining[ 3 ] = mm512_bcast128lo_64( 0x0100000000000000 );
  ctx->buf_ptr = 0;
   
   // --- update ---
@@ -86,18 +79,18 @@ 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( blocks << 56, 0x80 ); 
+       ctx->buffer[i] = mm512_set2_64( blocks << 56, 0x80 ); 
   }   
   else
   {
       // add first padding
-       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm512_bcast128lo_64( 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( blocks << 56, 0 );
+       ctx->buffer[i] = mm512_bcast128hi_64( blocks << 56 );
   }

   // digest final padding block and do output transform
@@ -143,18 +136,18 @@ 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_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm512_set2_64( blocks << 56, 0x80 );
   }
   else
   {
       // add first padding
-       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm512_bcast128lo_64( 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( blocks << 56, 0 );
+       ctx->buffer[i] = mm512_bcast128hi_64( blocks << 56 );
   }

 // digest final padding block and do output transform
@@ -179,8 +172,8 @@ int groestl256_2way_init( groestl256_2way_context* ctx, uint64_t hashlen )

  ctx->hashlen = hashlen;

-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return 1;
+//  if (ctx->chaining == NULL || ctx->buffer == NULL)
+//    return 1;

  for ( i = 0; i < SIZE256; i++ )
  {
@@ -189,7 +182,7 @@ int groestl256_2way_init( groestl256_2way_context* ctx, uint64_t hashlen )
  }

  // The only non-zero in the IV is len. It can be hard coded.
-  ctx->chaining[ 3 ] = m256_const2_64( 0, 0x0100000000000000 );
+  ctx->chaining[ 3 ] = mm256_bcast128lo_64( 0x0100000000000000 );

  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;
@@ -207,9 +200,6 @@ int groestl256_2way_full( groestl256_2way_context* ctx, void* output,
   __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;
@@ -217,7 +207,7 @@ int groestl256_2way_full( groestl256_2way_context* ctx, void* output,
   }

   // The only non-zero in the IV is len. It can be hard coded.
-   ctx->chaining[ 3 ] = m256_const2_64( 0, 0x0100000000000000 );
+   ctx->chaining[ 3 ] = mm256_bcast128lo_64( 0x0100000000000000 );
   ctx->buf_ptr = 0;

   // --- update ---
@@ -239,18 +229,18 @@ int groestl256_2way_full( groestl256_2way_context* ctx, void* output,
   if ( i == SIZE256 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-      ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+      ctx->buffer[i] = mm256_set2_64( blocks << 56, 0x80 );
   }
   else
   {
       // add first padding
-       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm256_bcast128lo_64( 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 );
+       ctx->buffer[i] = mm256_bcast128hi_64( blocks << 56 );
   }

   // digest final padding block and do output transform
@@ -295,23 +285,22 @@ int groestl256_2way_update_close( groestl256_2way_context* ctx, void* output,
   if ( i == SIZE256 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm256_set2_64( blocks << 56, 0x80 );
   }
   else
   {
       // add first padding
-       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm256_bcast128lo_64( 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 );
+       ctx->buffer[i] = mm256_bcast128hi_64( blocks << 56 );
   }

 // digest final padding block and do output transform
   TF512_2way( ctx->chaining, ctx->buffer );
-
   OF512_2way( ctx->chaining );

   // store hash result in output 
--- a/algo/groestl/groestl256-hash-4way.h
+++ b/algo/groestl/groestl256-hash-4way.h
@@ -22,10 +22,6 @@

 #define LENGTH (256)

-//#include "brg_endian.h"
-//#define NEED_UINT_64T
-//#include "algo/sha/brg_types.h"
-
 /* some sizes (number of bytes) */
 #define ROWS (8)
 #define LENGTHFIELDLEN (ROWS)
--- a/algo/groestl/groestl256-intr-4way.h
+++ b/algo/groestl/groestl256-intr-4way.h
@@ -165,7 +165,7 @@ static const __m512i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e,
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m512_const1_64( 0x1b1b1b1b1b1b1b1b ); \
+  b1 = _mm512_set1_epi64( 0x1b1b1b1b1b1b1b1b ); \
  MUL2( a0, b0, b1 ); \
  a0 = _mm512_xor_si512( a0, TEMP0 ); \
  MUL2( a1, b0, b1 ); \
@@ -205,116 +205,18 @@ static const __m512i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e,
  b1 = _mm512_xor_si512( b1, a4 ); \
 }/*MixBytes*/

-
-#if 0
-#define MixBytes(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 = _mm512_xor_si512(a0, a1);\
-  b0 = a2;\
-  a1 = _mm512_xor_si512(a1, a2);\
-  b1 = a3;\
-  a2 = _mm512_xor_si512(a2, a3);\
-  b2 = a4;\
-  a3 = _mm512_xor_si512(a3, a4);\
-  b3 = a5;\
-  a4 = _mm512_xor_si512(a4, a5);\
-  b4 = a6;\
-  a5 = _mm512_xor_si512(a5, a6);\
-  b5 = a7;\
-  a6 = _mm512_xor_si512(a6, a7);\
-  a7 = _mm512_xor_si512(a7, b6);\
-  \
-  /* build y4 y5 y6 ... in regs xmm8, xmm9, xmm10 by adding t_i*/\
-  b0 = _mm512_xor_si512(b0, a4);\
-  b6 = _mm512_xor_si512(b6, a4);\
-  b1 = _mm512_xor_si512(b1, a5);\
-  b7 = _mm512_xor_si512(b7, a5);\
-  b2 = _mm512_xor_si512(b2, a6);\
-  b0 = _mm512_xor_si512(b0, a6);\
-  /* spill values y_4, y_5 to memory */\
-  TEMP0 = b0;\
-  b3 = _mm512_xor_si512(b3, a7);\
-  b1 = _mm512_xor_si512(b1, a7);\
-  TEMP1 = b1;\
-  b4 = _mm512_xor_si512(b4, a0);\
-  b2 = _mm512_xor_si512(b2, a0);\
-  /* save values t0, t1, t2 to xmm8, xmm9 and memory */\
-  b0 = a0;\
-  b5 = _mm512_xor_si512(b5, a1);\
-  b3 = _mm512_xor_si512(b3, a1);\
-  b1 = a1;\
-  b6 = _mm512_xor_si512(b6, a2);\
-  b4 = _mm512_xor_si512(b4, a2);\
-  TEMP2 = a2;\
-  b7 = _mm512_xor_si512(b7, a3);\
-  b5 = _mm512_xor_si512(b5, a3);\
-  \
-  /* compute x_i = t_i + t_{i+3} */\
-  a0 = _mm512_xor_si512(a0, a3);\
-  a1 = _mm512_xor_si512(a1, a4);\
-  a2 = _mm512_xor_si512(a2, a5);\
-  a3 = _mm512_xor_si512(a3, a6);\
-  a4 = _mm512_xor_si512(a4, a7);\
-  a5 = _mm512_xor_si512(a5, b0);\
-  a6 = _mm512_xor_si512(a6, b1);\
-  a7 = _mm512_xor_si512(a7, TEMP2);\
-  \
-  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
-  /* compute w_i : add y_{i+4} */\
-  b1 = m512_const1_64( 0x1b1b1b1b1b1b1b1b );\
-  MUL2(a0, b0, b1);\
-  a0 = _mm512_xor_si512(a0, TEMP0);\
-  MUL2(a1, b0, b1);\
-  a1 = _mm512_xor_si512(a1, TEMP1);\
-  MUL2(a2, b0, b1);\
-  a2 = _mm512_xor_si512(a2, b2);\
-  MUL2(a3, b0, b1);\
-  a3 = _mm512_xor_si512(a3, b3);\
-  MUL2(a4, b0, b1);\
-  a4 = _mm512_xor_si512(a4, b4);\
-  MUL2(a5, b0, b1);\
-  a5 = _mm512_xor_si512(a5, b5);\
-  MUL2(a6, b0, b1);\
-  a6 = _mm512_xor_si512(a6, b6);\
-  MUL2(a7, b0, b1);\
-  a7 = _mm512_xor_si512(a7, b7);\
-  \
-  /* compute v_i : double w_i      */\
-  /* add to y_4 y_5 .. v3, v4, ... */\
-  MUL2(a0, b0, b1);\
-  b5 = _mm512_xor_si512(b5, a0);\
-  MUL2(a1, b0, b1);\
-  b6 = _mm512_xor_si512(b6, a1);\
-  MUL2(a2, b0, b1);\
-  b7 = _mm512_xor_si512(b7, a2);\
-  MUL2(a5, b0, b1);\
-  b2 = _mm512_xor_si512(b2, a5);\
-  MUL2(a6, b0, b1);\
-  b3 = _mm512_xor_si512(b3, a6);\
-  MUL2(a7, b0, b1);\
-  b4 = _mm512_xor_si512(b4, a7);\
-  MUL2(a3, b0, b1);\
-  MUL2(a4, b0, b1);\
-  b0 = TEMP0;\
-  b1 = TEMP1;\
-  b0 = _mm512_xor_si512(b0, a3);\
-  b1 = _mm512_xor_si512(b1, a4);\
-}/*MixBytes*/
-#endif
+#define MASK_NOT( a )  _mm512_mask_ternarylogic_epi64( a, 0xaa, a, a, 1 )

 #define ROUND(i, a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
  /* AddRoundConstant */\
-  b1 = m512_const2_64( 0xffffffffffffffff, 0 ); \
-  a0 = _mm512_xor_si512( a0, m512_const1_128( round_const_l0[i] ) );\
-  a1 = _mm512_xor_si512( a1, b1 );\
-  a2 = _mm512_xor_si512( a2, b1 );\
-  a3 = _mm512_xor_si512( a3, b1 );\
-  a4 = _mm512_xor_si512( a4, b1 );\
-  a5 = _mm512_xor_si512( a5, b1 );\
-  a6 = _mm512_xor_si512( a6, b1 );\
-  a7 = _mm512_xor_si512( a7, m512_const1_128( round_const_l7[i] ) );\
+  a0 = _mm512_xor_si512( a0, mm512_bcast_m128( round_const_l0[i] ) );\
+  a1 = MASK_NOT( a1 ); \
+  a2 = MASK_NOT( a2 ); \
+  a3 = MASK_NOT( a3 ); \
+  a4 = MASK_NOT( a4 ); \
+  a5 = MASK_NOT( a5 ); \
+  a6 = MASK_NOT( a6 ); \
+  a7 = _mm512_xor_si512( a7, mm512_bcast_m128( round_const_l7[i] ) );\
  \
  /* ShiftBytes + SubBytes (interleaved) */\
  b0 = _mm512_xor_si512( b0, b0 );\
@@ -450,7 +352,7 @@ static const __m512i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e,
 * outputs: (i0-7) = (0|S)
 */
 #define Matrix_Transpose_O_B(i0, i1, i2, i3, i4, i5, i6, i7, t0){\
-  t0 = _mm512_xor_si512( t0, t0 );\
+  t0 = m512_zero;\
  i1 = i0;\
  i3 = i2;\
  i5 = i4;\
@@ -481,11 +383,11 @@ static const __m512i SUBSH_MASK7 = { 0x090c000306080b07, 0x02050f0a0d01040e,

 void TF512_4way( __m512i* chaining, __m512i* message )
 {
-  static __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
-  static __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
-  static __m512i TEMP0;
-  static __m512i TEMP1;
-  static __m512i TEMP2;
+  __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m512i TEMP0;
+  __m512i TEMP1;
+  __m512i TEMP2;

  /* load message into registers xmm12 - xmm15 */
  xmm12 = message[0];
@@ -547,11 +449,11 @@ void TF512_4way( __m512i* chaining, __m512i* message )

 void OF512_4way( __m512i* chaining )
 {
-  static __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
-  static __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
-  static __m512i TEMP0;
-  static __m512i TEMP1;
-  static __m512i TEMP2;
+  __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m512i TEMP0;
+  __m512i TEMP1;
+  __m512i TEMP2;

  /* load CV into registers xmm8, xmm10, xmm12, xmm14 */
  xmm8 = chaining[0];
@@ -637,7 +539,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  j = _mm256_cmpgt_epi8(j, i );\
  i = _mm256_add_epi8(i, i);\
  j = _mm256_and_si256(j, k);\
-  i = _mm256_xor_si256(i, j);\
+  i = mm256_xorand( i, j, k );\
 }

 #define MixBytes_2way(a0, a1, a2, a3, a4, a5, a6, a7, b0, b1, b2, b3, b4, b5, b6, b7){\
@@ -648,7 +550,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  b0 = a2;\
  a1 = _mm256_xor_si256(a1, a2);\
  b1 = a3;\
-  a2 = _mm256_xor_si256(a2, a3);\
+  TEMP2 = _mm256_xor_si256(a2, a3);\
  b2 = a4;\
  a3 = _mm256_xor_si256(a3, a4);\
  b3 = a5;\
@@ -660,34 +562,20 @@ static const __m256i SUBSH_MASK7_2WAY =
  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);\
-  \
+  TEMP0 = mm256_xor3( b0, a4, a6 ); \
+  TEMP1 = mm256_xor3( b1, a5, a7 ); \
+  b2 = mm256_xor3( b2, a6, a0 ); \
+  b0 = a0; \
+  b3 = mm256_xor3( b3, a7, a1 ); \
+  b1 = a1; \
+  b6 = mm256_xor3( b6, a4, TEMP2 ); \
+  b4 = mm256_xor3( b4, a0, TEMP2 ); \
+  b7 = mm256_xor3( b7, a5, a3 ); \
+  b5 = mm256_xor3( b5, a1, 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);\
+  a2 = _mm256_xor_si256( TEMP2, a5);\
  a3 = _mm256_xor_si256(a3, a6);\
  a4 = _mm256_xor_si256(a4, a7);\
  a5 = _mm256_xor_si256(a5, b0);\
@@ -696,7 +584,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m256_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm256_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2_2WAY(a0, b0, b1);\
  a0 = _mm256_xor_si256(a0, TEMP0);\
  MUL2_2WAY(a1, b0, b1);\
@@ -738,15 +626,15 @@ static const __m256i SUBSH_MASK7_2WAY =

 #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] ) );\
+  b1 = mm256_bcast_m128( mm128_mask_32( m128_neg1, 0x3 ) ); \
+  a0 = _mm256_xor_si256( a0, mm256_bcast_m128( 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] ) );\
+  a7 = _mm256_xor_si256( a7, mm256_bcast_m128( round_const_l7[i] ) );\
  \
  /* ShiftBytes + SubBytes (interleaved) */\
  b0 = _mm256_xor_si256( b0, b0 );\
@@ -769,7 +657,6 @@ static const __m256i SUBSH_MASK7_2WAY =
  \
  /* 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 */
@@ -850,7 +737,7 @@ static const __m256i SUBSH_MASK7_2WAY =
 }/**/

 #define Matrix_Transpose_O_B_2way(i0, i1, i2, i3, i4, i5, i6, i7, t0){\
-  t0 = _mm256_xor_si256( t0, t0 );\
+  t0 = m256_zero;\
  i1 = i0;\
  i3 = i2;\
  i5 = i4;\
@@ -874,11 +761,11 @@ static const __m256i SUBSH_MASK7_2WAY =

 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;
+  __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m256i TEMP0;
+  __m256i TEMP1;
+  __m256i TEMP2;

  /* load message into registers xmm12 - xmm15 */
  xmm12 = message[0];
@@ -940,11 +827,11 @@ void TF512_2way( __m256i* chaining, __m256i* message )
  
 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;
+  __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m256i TEMP0;
+  __m256i TEMP1;
+  __m256i TEMP2;

  /* load CV into registers xmm8, xmm10, xmm12, xmm14 */
  xmm8 = chaining[0];
--- a/algo/groestl/groestl512-hash-4way.c
+++ b/algo/groestl/groestl512-hash-4way.c
@@ -21,15 +21,11 @@

 int groestl512_4way_init( groestl512_4way_context* ctx, uint64_t hashlen )
 {
-  if (ctx->chaining == NULL || ctx->buffer == NULL)
-    return 1;
-
  memset_zero_512( ctx->chaining, SIZE512 );
  memset_zero_512( ctx->buffer, SIZE512 );

  // The only non-zero in the IV is len. It can be hard coded.
-  ctx->chaining[ 6 ] = m512_const2_64( 0x0200000000000000, 0 );
-
+  ctx->chaining[ 6 ] = mm512_bcast128hi_64( 0x0200000000000000 );
  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;

@@ -64,14 +60,14 @@ 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_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm512_set2_64( blocks << 56, 0x80 );
   }   
   else
   {
-       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm512_bcast128lo_64( 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m512_zero;
-       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
+       ctx->buffer[i] = mm512_bcast128hi_64( blocks << 56 );
   }

   TF1024_4way( ctx->chaining, ctx->buffer );
@@ -97,7 +93,7 @@ int groestl512_4way_full( groestl512_4way_context* ctx, void* output,

   memset_zero_512( ctx->chaining, SIZE512 );
   memset_zero_512( ctx->buffer, SIZE512 );
-   ctx->chaining[ 6 ] = m512_const2_64( 0x0200000000000000, 0 );
+   ctx->chaining[ 6 ] = mm512_bcast128hi_64( 0x0200000000000000 );
   ctx->buf_ptr = 0;

   // --- update ---
@@ -116,14 +112,14 @@ int groestl512_4way_full( groestl512_4way_context* ctx, void* output,
   if ( i == SIZE512 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m512_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm512_set2_64( blocks << 56, 0x80 );
   }
   else
   {
-       ctx->buffer[i] = m512_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm512_bcast128lo_64( 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m512_zero;
-       ctx->buffer[i] = m512_const2_64( blocks << 56, 0 );
+       ctx->buffer[i] = mm512_bcast128hi_64( blocks << 56 );
   }

   TF1024_4way( ctx->chaining, ctx->buffer );
@@ -142,14 +138,11 @@ int groestl512_4way_full( groestl512_4way_context* ctx, void* output,

 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->chaining[ 6 ] = mm256_bcast128hi_64( 0x0200000000000000 );

  ctx->buf_ptr = 0;
  ctx->rem_ptr = 0;
@@ -185,14 +178,14 @@ int groestl512_2way_update_close( groestl512_2way_context* ctx, void* output,
   if ( i == SIZE512 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm256_set2_64( blocks << 56, 0x80 );
   }
   else
   {
-       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm256_bcast128lo_64( 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m256_zero;
-       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+       ctx->buffer[i] = mm256_bcast128hi_64( blocks << 56 );
   }

   TF1024_2way( ctx->chaining, ctx->buffer );
@@ -218,7 +211,7 @@ int groestl512_2way_full( groestl512_2way_context* ctx, void* output,

   memset_zero_256( ctx->chaining, SIZE512 );
   memset_zero_256( ctx->buffer, SIZE512 );
-   ctx->chaining[ 6 ] = m256_const2_64( 0x0200000000000000, 0 );
+   ctx->chaining[ 6 ] = mm256_bcast128hi_64( 0x0200000000000000 );
   ctx->buf_ptr = 0;

   // --- update ---
@@ -237,14 +230,14 @@ int groestl512_2way_full( groestl512_2way_context* ctx, void* output,
   if ( i == SIZE512 - 1 )
   {
       // only 1 vector left in buffer, all padding at once
-       ctx->buffer[i] = m256_const2_64( blocks << 56, 0x80 );
+       ctx->buffer[i] = mm256_set2_64( blocks << 56, 0x80 );
   }
   else
   {
-       ctx->buffer[i] = m256_const2_64( 0, 0x80 );
+       ctx->buffer[i] = mm256_bcast128lo_64( 0x80 );
       for ( i += 1; i < SIZE512 - 1; i++ )
           ctx->buffer[i] = m256_zero;
-       ctx->buffer[i] = m256_const2_64( blocks << 56, 0 );
+       ctx->buffer[i] = mm256_bcast128hi_64( blocks << 56 );
   }

   TF1024_2way( ctx->chaining, ctx->buffer );
--- a/algo/groestl/groestl512-intr-4way.h
+++ b/algo/groestl/groestl512-intr-4way.h
@@ -174,7 +174,7 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m512_const1_64( 0x1b1b1b1b1b1b1b1b ); \
+  b1 = _mm512_set1_epi64( 0x1b1b1b1b1b1b1b1b ); \
  MUL2( a0, b0, b1 ); \
  a0 = _mm512_xor_si512( a0, TEMP0 ); \
  MUL2( a1, b0, b1 ); \
@@ -238,7 +238,7 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,
  for ( round_counter = 0; round_counter < 14; round_counter += 2 ) \
  { \
    /* AddRoundConstant P1024 */\
-    xmm8 = _mm512_xor_si512( xmm8, m512_const1_128( \
+    xmm8 = _mm512_xor_si512( xmm8, mm512_bcast_m128( \
             casti_m128i( round_const_p, round_counter ) ) ); \
    /* ShiftBytes P1024 + pre-AESENCLAST */\
    xmm8  = _mm512_shuffle_epi8( xmm8,  SUBSH_MASK0 ); \
@@ -253,7 +253,7 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,
    SUBMIX(xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15, xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7);\
    \
     /* AddRoundConstant P1024 */\
-    xmm0 = _mm512_xor_si512( xmm0, m512_const1_128( \
+    xmm0 = _mm512_xor_si512( xmm0, mm512_bcast_m128( \
             casti_m128i( round_const_p, round_counter+1 ) ) ); \
    /* ShiftBytes P1024 + pre-AESENCLAST */\
    xmm0 = _mm512_shuffle_epi8( xmm0, SUBSH_MASK0 );\
@@ -282,7 +282,7 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,
    xmm12 = _mm512_xor_si512( xmm12, xmm1 );\
    xmm13 = _mm512_xor_si512( xmm13, xmm1 );\
    xmm14 = _mm512_xor_si512( xmm14, xmm1 );\
-    xmm15 = _mm512_xor_si512( xmm15, m512_const1_128( \
+    xmm15 = _mm512_xor_si512( xmm15, mm512_bcast_m128( \
                 casti_m128i( round_const_q, round_counter ) ) ); \
    /* ShiftBytes Q1024 + pre-AESENCLAST */\
    xmm8  = _mm512_shuffle_epi8( xmm8,  SUBSH_MASK1 );\
@@ -305,7 +305,7 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,
    xmm4 = _mm512_xor_si512( xmm4, xmm9 );\
    xmm5 = _mm512_xor_si512( xmm5, xmm9 );\
    xmm6 = _mm512_xor_si512( xmm6, xmm9 );\
-    xmm7 = _mm512_xor_si512( xmm7, m512_const1_128( \
+    xmm7 = _mm512_xor_si512( xmm7, mm512_bcast_m128( \
             casti_m128i( round_const_q, round_counter+1 ) ) ); \
    /* ShiftBytes Q1024 + pre-AESENCLAST */\
    xmm0 = _mm512_shuffle_epi8( xmm0, SUBSH_MASK1 );\
@@ -471,8 +471,8 @@ static const __m512i SUBSH_MASK7 = { 0x06090c0f0205080b, 0x0e0104070a0d0003,

 void INIT_4way( __m512i* chaining )
 {
-  static __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
-  static __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;

  /* load IV into registers xmm8 - xmm15 */
  xmm8 = chaining[0];
@@ -500,12 +500,12 @@ void INIT_4way( __m512i* chaining )

 void TF1024_4way( __m512i* chaining, const __m512i* message )
 {
-  static __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
-  static __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
-  static __m512i QTEMP[8];
-  static __m512i TEMP0;
-  static __m512i TEMP1;
-  static __m512i TEMP2;
+  __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m512i QTEMP[8];
+  __m512i TEMP0;
+  __m512i TEMP1;
+  __m512i TEMP2;

  /* load message into registers xmm8 - xmm15 (Q = message) */
  xmm8 = message[0];
@@ -606,11 +606,11 @@ void TF1024_4way( __m512i* chaining, const __m512i* message )

 void OF1024_4way( __m512i* chaining )
 {
-  static __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
-  static __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
-  static __m512i TEMP0;
-  static __m512i TEMP1;
-  static __m512i TEMP2;
+  __m512i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m512i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m512i TEMP0;
+  __m512i TEMP1;
+  __m512i TEMP2;

  /* load CV into registers xmm8 - xmm15 */
  xmm8 = chaining[0];
@@ -710,7 +710,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  b0 = a2;\
  a1 = _mm256_xor_si256(a1, a2);\
  b1 = a3;\
-  a2 = _mm256_xor_si256(a2, a3);\
+  TEMP2 = _mm256_xor_si256(a2, a3);\
  b2 = a4;\
  a3 = _mm256_xor_si256(a3, a4);\
  b3 = a5;\
@@ -722,34 +722,23 @@ static const __m256i SUBSH_MASK7_2WAY =
  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);\
+  TEMP0 = mm256_xor3( b0, a4, 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);\
+  TEMP1 = mm256_xor3( b1, a5, a7 ); \
+  b2 = mm256_xor3( b2, a6, 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);\
+  b0 = a0; \
+  b3 = mm256_xor3( b3, a7, a1 ); \
+  b1 = a1; \
+  b6 = mm256_xor3( b6, a4, TEMP2 ); \
+  b4 = mm256_xor3( b4, a0, TEMP2 ); \
+  b7 = mm256_xor3( b7, a5, a3 ); \
+  b5 = mm256_xor3( b5, a1, 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);\
+  a2 = _mm256_xor_si256( TEMP2, a5);\
  a3 = _mm256_xor_si256(a3, a6);\
  a4 = _mm256_xor_si256(a4, a7);\
  a5 = _mm256_xor_si256(a5, b0);\
@@ -758,7 +747,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  \
  /* compute z_i : double x_i using temp xmm8 and 1B xmm9 */\
  /* compute w_i : add y_{i+4} */\
-  b1 = m256_const1_64( 0x1b1b1b1b1b1b1b1b );\
+  b1 = _mm256_set1_epi64x( 0x1b1b1b1b1b1b1b1b );\
  MUL2_2WAY(a0, b0, b1);\
  a0 = _mm256_xor_si256(a0, TEMP0);\
  MUL2_2WAY(a1, b0, b1);\
@@ -822,7 +811,7 @@ static const __m256i SUBSH_MASK7_2WAY =
  for ( round_counter = 0; round_counter < 14; round_counter += 2 ) \
  { \
    /* AddRoundConstant P1024 */\
-    xmm8 = _mm256_xor_si256( xmm8, m256_const1_128( \
+    xmm8 = _mm256_xor_si256( xmm8, mm256_bcast_m128( \
             casti_m128i( round_const_p, round_counter ) ) ); \
    /* ShiftBytes P1024 + pre-AESENCLAST */\
    xmm8  = _mm256_shuffle_epi8( xmm8,  SUBSH_MASK0_2WAY ); \
@@ -837,7 +826,7 @@ static const __m256i SUBSH_MASK7_2WAY =
    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( \
+    xmm0 = _mm256_xor_si256( xmm0, mm256_bcast_m128( \
             casti_m128i( round_const_p, round_counter+1 ) ) ); \
    /* ShiftBytes P1024 + pre-AESENCLAST */\
    xmm0 = _mm256_shuffle_epi8( xmm0, SUBSH_MASK0_2WAY );\
@@ -866,7 +855,7 @@ static const __m256i SUBSH_MASK7_2WAY =
    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( \
+    xmm15 = _mm256_xor_si256( xmm15, mm256_bcast_m128( \
                 casti_m128i( round_const_q, round_counter ) ) ); \
    /* ShiftBytes Q1024 + pre-AESENCLAST */\
    xmm8  = _mm256_shuffle_epi8( xmm8,  SUBSH_MASK1_2WAY );\
@@ -889,7 +878,7 @@ static const __m256i SUBSH_MASK7_2WAY =
    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( \
+    xmm7 = _mm256_xor_si256( xmm7, mm256_bcast_m128( \
             casti_m128i( round_const_q, round_counter+1 ) ) ); \
    /* ShiftBytes Q1024 + pre-AESENCLAST */\
    xmm0 = _mm256_shuffle_epi8( xmm0, SUBSH_MASK1_2WAY );\
@@ -1040,8 +1029,8 @@ static const __m256i SUBSH_MASK7_2WAY =

 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;
+  __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;

  /* load IV into registers xmm8 - xmm15 */
  xmm8 = chaining[0];
@@ -1069,12 +1058,12 @@ void INIT_2way( __m256i *chaining )

 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;
+  __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m256i QTEMP[8];
+  __m256i TEMP0;
+  __m256i TEMP1;
+  __m256i TEMP2;

  /* load message into registers xmm8 - xmm15 (Q = message) */
  xmm8 = message[0];
@@ -1175,11 +1164,11 @@ void TF1024_2way( __m256i *chaining, const __m256i *message )

 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;
+  __m256i xmm0, xmm1, xmm2, xmm3, xmm4, xmm5, xmm6, xmm7;
+  __m256i xmm8, xmm9, xmm10, xmm11, xmm12, xmm13, xmm14, xmm15;
+  __m256i TEMP0;
+  __m256i TEMP1;
+  __m256i TEMP2;

  /* load CV into registers xmm8 - xmm15 */
  xmm8 = chaining[0];
--- a/algo/groestl/myr-groestl.c
+++ b/algo/groestl/myr-groestl.c
@@ -73,11 +73,11 @@ int scanhash_myriad( struct work *work, uint32_t max_nonce,
      be32enc(&endiandata[19], nonce);
      myriad_hash(hash, endiandata);

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

--- a/algo/groestl/myrgr-4way.c
+++ b/algo/groestl/myrgr-4way.c
@@ -4,7 +4,7 @@
 #include <stdint.h>
 #include <string.h>
 #include "aes_ni/hash-groestl.h"
-#include "algo/sha/sha-hash-4way.h"
+#include "algo/sha/sha256-hash.h"
 #if defined(__VAES__)
  #include "groestl512-hash-4way.h"
 #endif
--- a/algo/groestl/sph_groestl.h
+++ b/algo/groestl/sph_groestl.h
@@ -40,7 +40,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 #if !defined(__AES__)   
 /**
--- a/algo/hamsi/hamsi-hash-4way.c
+++ b/algo/hamsi/hamsi-hash-4way.c
--- a/algo/hamsi/hamsi-hash-4way.h
+++ b/algo/hamsi/hamsi-hash-4way.h
@@ -36,44 +36,64 @@
 #define HAMSI_4WAY_H__

 #include <stddef.h>
-#include "algo/sha/sph_types.h"

 #if defined (__AVX2__)

 #include "simd-utils.h"

-#ifdef __cplusplus
-extern "C"{
-#endif
-
-#define SPH_SIZE_hamsi512   512
+// Hamsi-512 4x64

 // Partial is only scalar but needs pointer ref for hamsi-helper
 // deprecate partial_len
-typedef struct {
+typedef struct
+{
   __m256i h[8];
   __m256i buf[1];
   size_t partial_len;
-   sph_u32 count_high, count_low;
+   uint32_t count_high, count_low;
 } hamsi_4way_big_context;
-
 typedef hamsi_4way_big_context hamsi512_4way_context;

 void hamsi512_4way_init( hamsi512_4way_context *sc );
 void hamsi512_4way_update( hamsi512_4way_context *sc, const void *data,
      size_t len );
-//#define hamsi512_4way hamsi512_4way_update
 void hamsi512_4way_close( hamsi512_4way_context *sc, void *dst );

+#define hamsi512_4x64_context   hamsi512_4way_context
+#define hamsi512_4x64_init      hamsi512_4way_init
+#define hamsi512_4x64_update    hamsi512_4way_update
+#define hamsi512_4x64_close     hamsi512_4way_close
+
+// Hamsi-512 8x32
+
+typedef struct
+{
+   __m256i h[16];
+   __m256i buf[2];
+   size_t partial_len;
+   uint32_t count_high, count_low;
+} hamsi_8x32_big_context;
+typedef hamsi_8x32_big_context hamsi512_8x32_context;
+
+void hamsi512_8x32_init( hamsi512_8x32_context *sc );
+void hamsi512_8x32_update( hamsi512_8x32_context *sc, const void *data,
+      size_t len );
+void hamsi512_8x32_close( hamsi512_8x32_context *sc, void *dst );
+void hamsi512_8x32_full( hamsi512_8x32_context *sc, void *dst, const void *data,
+      size_t len );
+
+#endif
+
 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

+// Hamsi-512 8x64
+
 typedef struct {
   __m512i h[8];
   __m512i buf[1];
   size_t partial_len;
-   sph_u32 count_high, count_low;
+   uint32_t count_high, count_low;
 } hamsi_8way_big_context;
-
 typedef hamsi_8way_big_context hamsi512_8way_context;

 void hamsi512_8way_init( hamsi512_8way_context *sc );
@@ -81,15 +101,29 @@ void hamsi512_8way_update( hamsi512_8way_context *sc, const void *data,
                           size_t len );
 void hamsi512_8way_close( hamsi512_8way_context *sc, void *dst );

+#define hamsi512_8x64_context   hamsi512_8way_context
+#define hamsi512_8x64_init      hamsi512_8way_init
+#define hamsi512_8x64_update    hamsi512_8way_update
+#define hamsi512_8x64_close     hamsi512_8way_close

-
-#endif
-
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif
+// Hamsi-512 16x32
+
+typedef struct
+{
+   __m512i h[16];
+   __m512i buf[2];
+   size_t partial_len;
+   uint32_t count_high, count_low;
+} hamsi_16x32_big_context;
+typedef hamsi_16x32_big_context hamsi512_16x32_context;
+
+void hamsi512_16x32_init( hamsi512_16x32_context *sc );
+void hamsi512_16x32_update( hamsi512_16x32_context *sc, const void *data,
+                           size_t len );
+void hamsi512_16way_close( hamsi512_16x32_context *sc, void *dst );
+void hamsi512_16x32_full( hamsi512_16x32_context *sc, void *dst,
+                          const void *data, size_t len );
+
+#endif   // AVX512

 #endif
--- a/algo/hamsi/sph_hamsi.h
+++ b/algo/hamsi/sph_hamsi.h
@@ -36,7 +36,7 @@
 #define SPH_HAMSI_H__

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 #ifdef __cplusplus
 extern "C"{
--- a/algo/haval/haval-4way-helper.c
+++ b/algo/haval/haval-4way-helper.c
@@ -48,7 +48,7 @@ SPH_XCAT(SPH_XCAT(haval, PASSES), _4way_update)
   while ( len > 0 )
   {
      unsigned clen;
-      sph_u32 clow, clow2;
+      uint32_t clow, clow2;

      clen = 128U - current;
      if ( clen > len )
@@ -67,7 +67,7 @@ SPH_XCAT(SPH_XCAT(haval, PASSES), _4way_update)
         current = 0;
      }
      clow = sc->count_low;
-      clow2 = SPH_T32(clow + clen);
+      clow2 = clow + clen;
      sc->count_low = clow2;
      if ( clow2 < clow )
         sc->count_high ++;
--- a/algo/haval/haval-hash-4way.c
+++ b/algo/haval/haval-hash-4way.c
@@ -52,6 +52,56 @@ extern "C"{
 #define SPH_SMALL_FOOTPRINT_HAVAL   1
 //#endif

+#if defined(__AVX512VL__)
+
+// ( ~( a ^ b ) ) & c
+#define mm128_andnotxor( a, b, c ) \
+   _mm_ternarylogic_epi32( a, b, c, 0x82  )
+
+#else
+
+#define mm128_andnotxor( a, b, c ) \
+   _mm_andnot_si128( _mm_xor_si128( a, b ), c )
+
+#endif
+
+#define F1(x6, x5, x4, x3, x2, x1, x0) \
+ mm128_xor3( x0, mm128_andxor( x1, x0, x4 ), \
+                 _mm_xor_si128( _mm_and_si128( x2, x5 ), \
+                                _mm_and_si128( x3, x6 ) ) ) \
+
+#define F2(x6, x5, x4, x3, x2, x1, x0) \
+   mm128_xor3( mm128_andxor( x2, _mm_andnot_si128( x3, x1 ), \
+                       mm128_xor3( _mm_and_si128( x4, x5 ), x6, x0 )  ), \
+               mm128_andxor( x4, x1, x5 ), \
+               mm128_xorand( x0, x3, x5 ) ) \
+
+#define F3(x6, x5, x4, x3, x2, x1, x0) \
+  mm128_xor3( x0, \
+              _mm_and_si128( x3, \
+                         mm128_xor3( _mm_and_si128( x1, x2 ), x6, x0 ) ), \
+              _mm_xor_si128( _mm_and_si128( x1, x4 ), \
+                             _mm_and_si128( x2, x5 ) ) )
+
+#define F4(x6, x5, x4, x3, x2, x1, x0) \
+  mm128_xor3( \
+      mm128_andxor( x3, x5, \
+                    _mm_xor_si128( _mm_and_si128( x1, x2 ), \
+                                      _mm_or_si128( x4, x6 ) ) ), \
+      _mm_and_si128( x4, \
+                        mm128_xor3( x0, _mm_andnot_si128( x2, x5 ), \
+                                    _mm_xor_si128( x1, x6 ) ) ), \
+      mm128_xorand( x0, x2, x6 ) )
+
+#define F5(x6, x5, x4, x3, x2, x1, x0) \
+   _mm_xor_si128( \
+         mm128_andnotxor( mm128_and3( x1, x2, x3 ), x5, x0 ), \
+         mm128_xor3( _mm_and_si128( x1, x4 ), \
+                     _mm_and_si128( x2, x5 ), \
+                     _mm_and_si128( x3, x6 ) ) )
+  
+
+/*
 #define F1(x6, x5, x4, x3, x2, x1, x0) \
   _mm_xor_si128( x0, \
       _mm_xor_si128( _mm_and_si128(_mm_xor_si128( x0, x4 ), x1 ), \
@@ -96,6 +146,7 @@ extern "C"{
      _mm_xor_si128( _mm_xor_si128( _mm_and_si128( x1, x4 ), \
                                    _mm_and_si128( x2, x5 ) ), \
                                    _mm_and_si128( x3, x6 ) ) )
+*/

 /*
 * The macros below integrate the phi() permutations, depending on the
@@ -141,6 +192,13 @@ do { \
                       _mm_add_epi32( w, _mm_set1_epi32( c ) ) ); \
 } while (0)

+#define STEP1(n, p, x7, x6, x5, x4, x3, x2, x1, x0, w) \
+do { \
+   __m128i t = FP ## n ## _ ## p(x6, x5, x4, x3, x2, x1, x0); \
+   x7 = _mm_add_epi32( _mm_add_epi32( mm128_ror_32( t, 7 ), \
+                                      mm128_ror_32( x7, 11 ) ), w ); \
+} while (0)
+
 /*
 * PASSy(n, in) computes pass number "y", for a total of "n", using the
 * one-argument macro "in" to access input words. Current state is assumed
@@ -152,22 +210,22 @@ do { \
 #define PASS1(n, in)   do { \
 		unsigned pass_count; \
 		for (pass_count = 0; pass_count < 32; pass_count += 8) { \
-			STEP(n, 1, s7, s6, s5, s4, s3, s2, s1, s0, \
-				in(pass_count + 0), SPH_C32(0x00000000)); \
-			STEP(n, 1, s6, s5, s4, s3, s2, s1, s0, s7, \
-				in(pass_count + 1), SPH_C32(0x00000000)); \
-			STEP(n, 1, s5, s4, s3, s2, s1, s0, s7, s6, \
-				in(pass_count + 2), SPH_C32(0x00000000)); \
-			STEP(n, 1, s4, s3, s2, s1, s0, s7, s6, s5, \
-				in(pass_count + 3), SPH_C32(0x00000000)); \
-			STEP(n, 1, s3, s2, s1, s0, s7, s6, s5, s4, \
-				in(pass_count + 4), SPH_C32(0x00000000)); \
-			STEP(n, 1, s2, s1, s0, s7, s6, s5, s4, s3, \
-				in(pass_count + 5), SPH_C32(0x00000000)); \
-			STEP(n, 1, s1, s0, s7, s6, s5, s4, s3, s2, \
-				in(pass_count + 6), SPH_C32(0x00000000)); \
-			STEP(n, 1, s0, s7, s6, s5, s4, s3, s2, s1, \
-				in(pass_count + 7), SPH_C32(0x00000000)); \
+			STEP1(n, 1, s7, s6, s5, s4, s3, s2, s1, s0, \
+				in(pass_count + 0) ); \
+			STEP1(n, 1, s6, s5, s4, s3, s2, s1, s0, s7, \
+				in(pass_count + 1) ); \
+			STEP1(n, 1, s5, s4, s3, s2, s1, s0, s7, s6, \
+				in(pass_count + 2) ); \
+			STEP1(n, 1, s4, s3, s2, s1, s0, s7, s6, s5, \
+				in(pass_count + 3) ); \
+			STEP1(n, 1, s3, s2, s1, s0, s7, s6, s5, s4, \
+				in(pass_count + 4) ); \
+			STEP1(n, 1, s2, s1, s0, s7, s6, s5, s4, s3, \
+				in(pass_count + 5) ); \
+			STEP1(n, 1, s1, s0, s7, s6, s5, s4, s3, s2, \
+				in(pass_count + 6) ); \
+			STEP1(n, 1, s0, s7, s6, s5, s4, s3, s2, s1, \
+				in(pass_count + 7) ); \
   		} \
 	} while (0)

@@ -234,7 +292,9 @@ static const unsigned MP5[32] = {
 	 2, 23, 16, 22,  4,  1, 25, 15
 };

-static const sph_u32 RK2[32] = {
+#define SPH_C32(x) (x)
+
+static const uint32_t RK2[32] = {
 	SPH_C32(0x452821E6), SPH_C32(0x38D01377),
 	SPH_C32(0xBE5466CF), SPH_C32(0x34E90C6C),
 	SPH_C32(0xC0AC29B7), SPH_C32(0xC97C50DD),
@@ -253,7 +313,7 @@ static const sph_u32 RK2[32] = {
 	SPH_C32(0x7B54A41D), SPH_C32(0xC25A59B5)
 };

-static const sph_u32 RK3[32] = {
+static const uint32_t RK3[32] = {
 	SPH_C32(0x9C30D539), SPH_C32(0x2AF26013),
 	SPH_C32(0xC5D1B023), SPH_C32(0x286085F0),
 	SPH_C32(0xCA417918), SPH_C32(0xB8DB38EF),
@@ -272,7 +332,7 @@ static const sph_u32 RK3[32] = {
 	SPH_C32(0xAFD6BA33), SPH_C32(0x6C24CF5C)
 };

-static const sph_u32 RK4[32] = {
+static const uint32_t RK4[32] = {
 	SPH_C32(0x7A325381), SPH_C32(0x28958677),
 	SPH_C32(0x3B8F4898), SPH_C32(0x6B4BB9AF),
 	SPH_C32(0xC4BFE81B), SPH_C32(0x66282193),
@@ -291,7 +351,7 @@ static const sph_u32 RK4[32] = {
 	SPH_C32(0x6EEF0B6C), SPH_C32(0x137A3BE4)
 };

-static const sph_u32 RK5[32] = {
+static const uint32_t RK5[32] = {
 	SPH_C32(0xBA3BF050), SPH_C32(0x7EFB2A98),
 	SPH_C32(0xA1F1651D), SPH_C32(0x39AF0176),
 	SPH_C32(0x66CA593E), SPH_C32(0x82430E88),
@@ -605,25 +665,32 @@ do { \
                       _mm256_add_epi32( w, _mm256_set1_epi32( c ) ) ); \
 } while (0)

+#define STEP1_8W(n, p, x7, x6, x5, x4, x3, x2, x1, x0, w) \
+do { \
+   __m256i t = FP ## n ## _ ## p ## _8W(x6, x5, x4, x3, x2, x1, x0); \
+   x7 = _mm256_add_epi32( _mm256_add_epi32( mm256_ror_32( t, 7 ), \
+                                      mm256_ror_32( x7, 11 ) ), w ); \
+} while (0)
+   
 #define PASS1_8W(n, in)   do { \
      unsigned pass_count; \
      for (pass_count = 0; pass_count < 32; pass_count += 8) { \
-         STEP_8W(n, 1, s7, s6, s5, s4, s3, s2, s1, s0, \
-            in(pass_count + 0), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s6, s5, s4, s3, s2, s1, s0, s7, \
-            in(pass_count + 1), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s5, s4, s3, s2, s1, s0, s7, s6, \
-            in(pass_count + 2), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s4, s3, s2, s1, s0, s7, s6, s5, \
-            in(pass_count + 3), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s3, s2, s1, s0, s7, s6, s5, s4, \
-            in(pass_count + 4), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s2, s1, s0, s7, s6, s5, s4, s3, \
-            in(pass_count + 5), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s1, s0, s7, s6, s5, s4, s3, s2, \
-            in(pass_count + 6), SPH_C32(0x00000000)); \
-         STEP_8W(n, 1, s0, s7, s6, s5, s4, s3, s2, s1, \
-            in(pass_count + 7), SPH_C32(0x00000000)); \
+         STEP1_8W(n, 1, s7, s6, s5, s4, s3, s2, s1, s0, \
+            in(pass_count + 0) ); \
+         STEP1_8W(n, 1, s6, s5, s4, s3, s2, s1, s0, s7, \
+            in(pass_count + 1) ); \
+         STEP1_8W(n, 1, s5, s4, s3, s2, s1, s0, s7, s6, \
+            in(pass_count + 2) ); \
+         STEP1_8W(n, 1, s4, s3, s2, s1, s0, s7, s6, s5, \
+            in(pass_count + 3) ); \
+         STEP1_8W(n, 1, s3, s2, s1, s0, s7, s6, s5, s4, \
+            in(pass_count + 4) ); \
+         STEP1_8W(n, 1, s2, s1, s0, s7, s6, s5, s4, s3, \
+            in(pass_count + 5) ); \
+         STEP1_8W(n, 1, s1, s0, s7, s6, s5, s4, s3, s2, \
+            in(pass_count + 6) ); \
+         STEP1_8W(n, 1, s0, s7, s6, s5, s4, s3, s2, s1, \
+            in(pass_count + 7) ); \
         } \
   } while (0)

@@ -726,14 +793,14 @@ do { \
 static void
 haval_8way_init( haval_8way_context *sc, unsigned olen, unsigned passes )
 {
-   sc->s0 = m256_const1_32( 0x243F6A88UL );
-   sc->s1 = m256_const1_32( 0x85A308D3UL );
-   sc->s2 = m256_const1_32( 0x13198A2EUL );
-   sc->s3 = m256_const1_32( 0x03707344UL );
-   sc->s4 = m256_const1_32( 0xA4093822UL );
-   sc->s5 = m256_const1_32( 0x299F31D0UL );
-   sc->s6 = m256_const1_32( 0x082EFA98UL );
-   sc->s7 = m256_const1_32( 0xEC4E6C89UL );
+   sc->s0 = _mm256_set1_epi32( 0x243F6A88UL );
+   sc->s1 = _mm256_set1_epi32( 0x85A308D3UL );
+   sc->s2 = _mm256_set1_epi32( 0x13198A2EUL );
+   sc->s3 = _mm256_set1_epi32( 0x03707344UL );
+   sc->s4 = _mm256_set1_epi32( 0xA4093822UL );
+   sc->s5 = _mm256_set1_epi32( 0x299F31D0UL );
+   sc->s6 = _mm256_set1_epi32( 0x082EFA98UL );
+   sc->s7 = _mm256_set1_epi32( 0xEC4E6C89UL );
   sc->olen = olen;
   sc->passes = passes;
   sc->count_high = 0;
--- a/algo/haval/haval-hash-4way.h
+++ b/algo/haval/haval-hash-4way.h
@@ -68,7 +68,6 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
 #include "simd-utils.h"

 #define SPH_SIZE_haval256_5   256
@@ -77,7 +76,7 @@ typedef struct {
   __m128i buf[32];
   __m128i s0, s1, s2, s3, s4, s5, s6, s7;
   unsigned olen, passes;
-   sph_u32 count_high, count_low;
+   uint32_t count_high, count_low;
 } haval_4way_context;

 typedef haval_4way_context haval256_5_4way_context;
--- a/algo/haval/sph-haval.h
+++ b/algo/haval/sph-haval.h
@@ -66,7 +66,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for HAVAL-128/3.
--- a/algo/heavy/sph_hefty1.c
+++ b/algo/heavy/sph_hefty1.c
@@ -1,382 +0,0 @@
-/*
- * HEFTY1 cryptographic hash function
- *
- * Copyright (c) 2014, dbcc14 <BM-NBx4AKznJuyem3dArgVY8MGyABpihRy5>
- * 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 COPYRIGHT HOLDERS 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 COPYRIGHT OWNER 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.
- *
- * The views and conclusions contained in the software and documentation are those
- * of the authors and should not be interpreted as representing official policies,
- * either expressed or implied, of the FreeBSD Project.
- */
-
-#include <assert.h>
-#include <string.h>
-
-#ifdef _MSC_VER
-#define inline __inline
-#endif
-
-#include "sph_hefty1.h"
-
-#define Min(A, B) (A <= B ? A : B)
-#define RoundFunc(ctx, A, B, C, D, E, F, G, H, W, K)                    \
-    {                                                                   \
-        /* To thwart parallelism, Br modifies itself each time it's     \
-         * called.  This also means that calling it in different        \
-         * orders yeilds different results.  In C the order of          \
-         * evaluation of function arguments and + operands are          \
-         * unspecified (and depends on the compiler), so we must make   \
-         * the order of Br calls explicit.                              \
-         */                                                             \
-        uint32_t brG = Br(ctx, G);                                      \
-        uint32_t tmp1 = Ch(E, Br(ctx, F), brG) + H + W + K;             \
-        uint32_t tmp2 = tmp1 + Sigma1(Br(ctx, E));                      \
-        uint32_t brC = Br(ctx, C);                                      \
-        uint32_t brB = Br(ctx, B);                                      \
-        uint32_t tmp3 = Ma(Br(ctx, A), brB, brC);                       \
-        uint32_t tmp4 = tmp3 + Sigma0(Br(ctx, A));                      \
-        H = G;                                                          \
-        G = F;                                                          \
-        F = E;                                                          \
-        E = D + Br(ctx, tmp2);                                          \
-        D = C;                                                          \
-        C = B;                                                          \
-        B = A;                                                          \
-        A = tmp2 + tmp4;                                                \
-    }                                                                   \
-
-/* Nothing up my sleeve constants */
-const static uint32_t K[64] = {
-    0x428a2f98UL, 0x71374491UL, 0xb5c0fbcfUL, 0xe9b5dba5UL,
-    0x3956c25bUL, 0x59f111f1UL, 0x923f82a4UL, 0xab1c5ed5UL,
-    0xd807aa98UL, 0x12835b01UL, 0x243185beUL, 0x550c7dc3UL,
-    0x72be5d74UL, 0x80deb1feUL, 0x9bdc06a7UL, 0xc19bf174UL,
-    0xe49b69c1UL, 0xefbe4786UL, 0x0fc19dc6UL, 0x240ca1ccUL,
-    0x2de92c6fUL, 0x4a7484aaUL, 0x5cb0a9dcUL, 0x76f988daUL,
-    0x983e5152UL, 0xa831c66dUL, 0xb00327c8UL, 0xbf597fc7UL,
-    0xc6e00bf3UL, 0xd5a79147UL, 0x06ca6351UL, 0x14292967UL,
-    0x27b70a85UL, 0x2e1b2138UL, 0x4d2c6dfcUL, 0x53380d13UL,
-    0x650a7354UL, 0x766a0abbUL, 0x81c2c92eUL, 0x92722c85UL,
-    0xa2bfe8a1UL, 0xa81a664bUL, 0xc24b8b70UL, 0xc76c51a3UL,
-    0xd192e819UL, 0xd6990624UL, 0xf40e3585UL, 0x106aa070UL,
-    0x19a4c116UL, 0x1e376c08UL, 0x2748774cUL, 0x34b0bcb5UL,
-    0x391c0cb3UL, 0x4ed8aa4aUL, 0x5b9cca4fUL, 0x682e6ff3UL,
-    0x748f82eeUL, 0x78a5636fUL, 0x84c87814UL, 0x8cc70208UL,
-    0x90befffaUL, 0xa4506cebUL, 0xbef9a3f7UL, 0xc67178f2UL
-};
-
-/* Initial hash values */
-const static uint32_t H[HEFTY1_STATE_WORDS] = {
-    0x6a09e667UL,
-    0xbb67ae85UL,
-    0x3c6ef372UL,
-    0xa54ff53aUL,
-    0x510e527fUL,
-    0x9b05688cUL,
-    0x1f83d9abUL,
-    0x5be0cd19UL
-};
-
-static inline uint32_t Rr(uint32_t X, uint8_t n)
-{
-    return (X >> n) | (X << (32 - n));
-}
-
-static inline uint32_t Ch(uint32_t E, uint32_t F, uint32_t G)
-{
-    return (E & F) ^ (~E & G);
-}
-
-static inline uint32_t Sigma1(uint32_t E)
-{
-    return Rr(E, 6) ^ Rr(E, 11) ^ Rr(E, 25);
-}
-
-static inline uint32_t sigma1(uint32_t X)
-{
-    return Rr(X, 17) ^ Rr(X, 19) ^ (X >> 10);
-}
-
-static inline uint32_t Ma(uint32_t A, uint32_t B, uint32_t C)
-{
-    return (A & B) ^ (A & C) ^ (B & C);
-}
-
-static inline uint32_t Sigma0(uint32_t A)
-{
-    return Rr(A, 2) ^ Rr(A, 13) ^ Rr(A, 22);
-}
-
-static inline uint32_t sigma0(uint32_t X)
-{
-    return Rr(X, 7) ^ Rr(X, 18) ^ (X >> 3);
-}
-
-static inline uint32_t Reverse32(uint32_t n)
-{
-    #if BYTE_ORDER == LITTLE_ENDIAN
-        return n << 24 | (n & 0x0000ff00) << 8 | (n & 0x00ff0000) >> 8 | n >> 24;
-    #else
-        return n;
-    #endif
-}
-
-static inline uint64_t Reverse64(uint64_t n)
-{
-    #if BYTE_ORDER == LITTLE_ENDIAN
-        uint32_t a = n >> 32;
-        uint32_t b = (n << 32) >> 32;
-
-        return (uint64_t)Reverse32(b) << 32 | Reverse32(a);
-    #else
-        return n;
-    #endif
-}
-
-/* Smoosh byte into nibble */
-static inline uint8_t Smoosh4(uint8_t X)
-{
-    return (X >> 4) ^ (X & 0xf);
-}
-
-/* Smoosh 32-bit word into 2-bits */
-static inline uint8_t Smoosh2(uint32_t X)
-{
-    uint16_t w = (X >> 16) ^ (X & 0xffff);
-    uint8_t n = Smoosh4((w >> 8) ^ (w & 0xff));
-    return (n >> 2) ^ (n & 0x3);
-}
-
-static void Mangle(uint32_t *S)
-{
-    uint32_t *R = S;
-    uint32_t *C = &S[1];
-
-    uint8_t r0 = Smoosh4(R[0] >> 24);
-    uint8_t r1 = Smoosh4(R[0] >> 16);
-    uint8_t r2 = Smoosh4(R[0] >> 8);
-    uint8_t r3 = Smoosh4(R[0] & 0xff);
-
-    int i;
-
-    /* Diffuse */
-    uint32_t tmp = 0;
-    for (i = 0; i < HEFTY1_SPONGE_WORDS - 1; i++) {
-        uint8_t r = Smoosh2(tmp);
-        switch (r) {
-        case 0:
-            C[i] ^= Rr(R[0], i + r0);
-            break;
-        case 1:
-            C[i] += Rr(~R[0], i + r1);
-            break;
-        case 2:
-            C[i] &= Rr(~R[0], i + r2);
-            break;
-        case 3:
-            C[i] ^= Rr(R[0], i + r3);
-            break;
-        }
-        tmp ^= C[i];
-    }
-
-    /* Compress */
-    tmp = 0;
-    for (i = 0; i < HEFTY1_SPONGE_WORDS - 1; i++)
-        if (i % 2)
-            tmp ^= C[i];
-        else
-            tmp += C[i];
-    R[0] ^= tmp;
-}
-
-static void Absorb(uint32_t *S, uint32_t X)
-{
-    uint32_t *R = S;
-    R[0] ^= X;
-    Mangle(S);
-}
-
-static uint32_t Squeeze(uint32_t *S)
-{
-    uint32_t Y = S[0];
-    Mangle(S);
-    return Y;
-}
-
-/* Branch, compress and serialize function */
-static inline uint32_t Br(HEFTY1_CTX *ctx, uint32_t X)
-{
-    uint32_t R = Squeeze(ctx->sponge);
-
-    uint8_t r0 = R >> 8;
-    uint8_t r1 = R & 0xff;
-
-    uint32_t Y = 1 << (r0 % 32);
-
-    switch (r1 % 4)
-    {
-    case 0:
-        /* Do nothing */
-        break;
-    case 1:
-        return X & ~Y;
-    case 2:
-        return X | Y;
-    case 3:
-        return X ^ Y;
-    }
-
-    return X;
-}
-
-static void HashBlock(HEFTY1_CTX *ctx)
-{
-    uint32_t A, B, C, D, E, F, G, H;
-    uint32_t W[HEFTY1_BLOCK_BYTES];
-
-    assert(ctx);
-
-    A = ctx->h[0];
-    B = ctx->h[1];
-    C = ctx->h[2];
-    D = ctx->h[3];
-    E = ctx->h[4];
-    F = ctx->h[5];
-    G = ctx->h[6];
-    H = ctx->h[7];
-
-    int t = 0;
-    for (; t < 16; t++) {
-        W[t] = Reverse32(((uint32_t *)&ctx->block[0])[t]); /* To host byte order */
-        Absorb(ctx->sponge, W[t] ^ K[t]);
-    }
-
-    for (t = 0; t < 16; t++) {
-        Absorb(ctx->sponge, D ^ H);
-        RoundFunc(ctx, A, B, C, D, E, F, G, H, W[t], K[t]);
-    }
-    for (t = 16; t < 64; t++) {
-        Absorb(ctx->sponge, H + D);
-        W[t] = sigma1(W[t - 2]) + W[t - 7] + sigma0(W[t - 15]) + W[t - 16];
-        RoundFunc(ctx, A, B, C, D, E, F, G, H, W[t], K[t]);
-    }
-
-    ctx->h[0] += A;
-    ctx->h[1] += B;
-    ctx->h[2] += C;
-    ctx->h[3] += D;
-    ctx->h[4] += E;
-    ctx->h[5] += F;
-    ctx->h[6] += G;
-    ctx->h[7] += H;
-
-    A = 0;
-    B = 0;
-    C = 0;
-    D = 0;
-    E = 0;
-    F = 0;
-    G = 0;
-    H = 0;
-
-    memset(W, 0, sizeof(W));
-}
-
-/* Public interface */
-
-void HEFTY1_Init(HEFTY1_CTX *ctx)
-{
-    assert(ctx);
-
-    memcpy(ctx->h, H, sizeof(ctx->h));
-    memset(ctx->block, 0, sizeof(ctx->block));
-    ctx->written = 0;
-    memset(ctx->sponge, 0, sizeof(ctx->sponge));
-}
-
-void HEFTY1_Update(HEFTY1_CTX *ctx, const void *buf, size_t len)
-{
-    assert(ctx);
-
-    uint64_t read = 0;
-    while (len) {
-        size_t end = (size_t)(ctx->written % HEFTY1_BLOCK_BYTES);
-        size_t count = Min(len, HEFTY1_BLOCK_BYTES - end);
-        memcpy(&ctx->block[end], &((unsigned char *)buf)[read], count);
-        len -= count;
-        read += count;
-        ctx->written += count;
-        if (!(ctx->written % HEFTY1_BLOCK_BYTES))
-            HashBlock(ctx);
-    }
-}
-
-void HEFTY1_Final(unsigned char *digest, HEFTY1_CTX *ctx)
-{
-    assert(digest);
-    assert(ctx);
-
-    /* Pad message (FIPS 180 Section 5.1.1) */
-    size_t used = (size_t)(ctx->written % HEFTY1_BLOCK_BYTES);
-    ctx->block[used++] = 0x80; /* Append 1 to end of message */
-    if (used > HEFTY1_BLOCK_BYTES - 8) {
-        /* We have already written into the last 64bits, so
-         * we must continue into the next block. */
-        memset(&ctx->block[used], 0, HEFTY1_BLOCK_BYTES - used);
-        HashBlock(ctx);
-        used = 0; /* Create a new block (below) */
-    }
-
-    /* All remaining bits to zero */
-    memset(&ctx->block[used], 0, HEFTY1_BLOCK_BYTES - 8 - used);
-
-    /* The last 64bits encode the length (in network byte order) */
-    uint64_t *len = (uint64_t *)&ctx->block[HEFTY1_BLOCK_BYTES - 8];
-    *len = Reverse64(ctx->written*8);
-
-    HashBlock(ctx);
-
-    /* Convert back to network byte order */
-    int i = 0;
-    for (; i < HEFTY1_STATE_WORDS; i++)
-        ctx->h[i] = Reverse32(ctx->h[i]);
-
-    memcpy(digest, ctx->h, sizeof(ctx->h));
-    memset(ctx, 0, sizeof(HEFTY1_CTX));
-}
-
-unsigned char* HEFTY1(const unsigned char *buf, size_t len, unsigned char *digest)
-{
-    HEFTY1_CTX ctx;
-    static unsigned char m[HEFTY1_DIGEST_BYTES];
-
-    if (!digest)
-        digest = m;
-
-    HEFTY1_Init(&ctx);
-    HEFTY1_Update(&ctx, buf, len);
-    HEFTY1_Final(digest, &ctx);
-
-    return digest;
-}
--- a/algo/heavy/sph_hefty1.h
+++ b/algo/heavy/sph_hefty1.h
@@ -1,66 +0,0 @@
-/*
- * HEFTY1 cryptographic hash function
- *
- * Copyright (c) 2014, dbcc14 <BM-NBx4AKznJuyem3dArgVY8MGyABpihRy5>
- * 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 COPYRIGHT HOLDERS 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 COPYRIGHT OWNER 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.
- *
- * The views and conclusions contained in the software and documentation are those
- * of the authors and should not be interpreted as representing official policies,
- * either expressed or implied, of the FreeBSD Project.
- */
-
-#ifndef __HEFTY1_H__
-#define __HEFTY1_H__
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#ifndef WIN32
-#include <sys/types.h>
-#endif
-
-#include <inttypes.h>
-
-#define HEFTY1_DIGEST_BYTES 32
-#define HEFTY1_BLOCK_BYTES 64
-#define HEFTY1_STATE_WORDS 8
-#define HEFTY1_SPONGE_WORDS 4
-
-typedef struct HEFTY1_CTX {
-    uint32_t h[HEFTY1_STATE_WORDS];
-    uint8_t  block[HEFTY1_BLOCK_BYTES];
-    uint64_t written;
-    uint32_t sponge[HEFTY1_SPONGE_WORDS];
-} HEFTY1_CTX;
-
-void HEFTY1_Init(HEFTY1_CTX *cxt);
-void HEFTY1_Update(HEFTY1_CTX *cxt, const void *data, size_t len);
-void HEFTY1_Final(unsigned char *digest, HEFTY1_CTX *cxt);
-unsigned char* HEFTY1(const unsigned char *data, size_t len, unsigned char *digest);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif /* __HEFTY1_H__ */
--- a/algo/jh/jh-hash-4way.c
+++ b/algo/jh/jh-hash-4way.c
@@ -49,12 +49,11 @@ extern "C"{

 #define Sb_8W(x0, x1, x2, x3, c) \
 do { \
-   __m512i cc = _mm512_set1_epi64( c ); \
-    x3 = mm512_not( x3 ); \
+    const __m512i cc = _mm512_set1_epi64( c ); \
    x0 = mm512_xorandnot( x0, x2, cc ); \
    tmp = mm512_xorand( cc, x0, x1 ); \
-    x0 = mm512_xorand( x0, x2, x3 ); \
-    x3 = mm512_xorandnot( x3, x1, x2 ); \
+    x0 = mm512_xorandnot( x0, x3, x2 ); \
+    x3 = _mm512_ternarylogic_epi64( x3, x1, x2, 0x2d ); /* ~x3 ^ (~x1 & x2) */\
    x1 = mm512_xorand( x1, x0, x2 ); \
    x2 = mm512_xorandnot( x2, x3, x0 ); \
    x0 = mm512_xoror( x0, x1, x3 ); \
@@ -77,19 +76,31 @@ do { \

 #endif

+#if defined(__AVX512VL__)
+//TODO enable for AVX10_256, not used with AVX512VL
+
+#define notxorandnot( a, b, c ) \
+   _mm256_ternarylogic_epi64( a, b, c, 0x2d )
+
+#else
+
+#define notxorandnot( a, b, c ) \
+   _mm256_xor_si256( mm256_not( a ), _mm256_andnot_si256( b, c ) )
+
+#endif
+
 #define Sb(x0, x1, x2, x3, c) \
 do { \
-   __m256i cc = _mm256_set1_epi64x( c ); \
-    x3 = mm256_not( x3 ); \
-    x0 = _mm256_xor_si256( x0, _mm256_andnot_si256( x2, cc ) ); \
-    tmp = _mm256_xor_si256( cc, _mm256_and_si256( x0, x1 ) ); \
-    x0 = _mm256_xor_si256( x0, _mm256_and_si256( x2, x3 ) ); \
-    x3 = _mm256_xor_si256( x3, _mm256_andnot_si256( x1, x2 ) ); \
-    x1 = _mm256_xor_si256( x1, _mm256_and_si256( x0, x2 ) ); \
-    x2 = _mm256_xor_si256( x2, _mm256_andnot_si256( x3, x0 ) ); \
-    x0 = _mm256_xor_si256( x0, _mm256_or_si256( x1, x3 ) ); \
-    x3 = _mm256_xor_si256( x3, _mm256_and_si256( x1, x2 ) ); \
-    x1 = _mm256_xor_si256( x1, _mm256_and_si256( tmp, x0 ) ); \
+    const __m256i cc = _mm256_set1_epi64x( c ); \
+    x0 = mm256_xorandnot( x0, x2, cc ); \
+    tmp = mm256_xorand( cc, x0, x1 ); \
+    x0 = mm256_xorandnot( x0, x3, x2 ); \
+    x3 = notxorandnot( x3, x1, x2 ); \
+    x1 = mm256_xorand( x1, x0, x2 ); \
+    x2 = mm256_xorandnot( x2, x3, x0 ); \
+    x0 = mm256_xoror( x0, x1, x3 ); \
+    x3 = mm256_xorand( x3, x1, x2 ); \
+    x1 = mm256_xorand( x1, tmp, x0 ); \
    x2 = _mm256_xor_si256( x2, tmp ); \
 } while (0)

@@ -97,11 +108,11 @@ do { \
 do { \
    x4 = _mm256_xor_si256( x4, x1 ); \
    x5 = _mm256_xor_si256( x5, x2 ); \
-    x6 = _mm256_xor_si256( x6, _mm256_xor_si256( x3, x0 ) ); \
+    x6 = mm256_xor3( x6, x3, x0 ); \
    x7 = _mm256_xor_si256( x7, x0 ); \
    x0 = _mm256_xor_si256( x0, x5 ); \
    x1 = _mm256_xor_si256( x1, x6 ); \
-    x2 = _mm256_xor_si256( x2, _mm256_xor_si256( x7, x4 ) ); \
+    x2 = mm256_xor3( x2, x7, x4 ); \
    x3 = _mm256_xor_si256( x3, x4 ); \
 } while (0)

@@ -324,12 +335,12 @@ do { \
 } while (0)


-#define W80(x)   Wz_8W(x, m512_const1_64( 0x5555555555555555 ),  1 )
-#define W81(x)   Wz_8W(x, m512_const1_64( 0x3333333333333333 ),  2 )
-#define W82(x)   Wz_8W(x, m512_const1_64( 0x0F0F0F0F0F0F0F0F ),  4 )
-#define W83(x)   Wz_8W(x, m512_const1_64( 0x00FF00FF00FF00FF ),  8 ) 
-#define W84(x)   Wz_8W(x, m512_const1_64( 0x0000FFFF0000FFFF ), 16 )
-#define W85(x)   Wz_8W(x, m512_const1_64( 0x00000000FFFFFFFF ), 32 )
+#define W80(x)   Wz_8W(x, _mm512_set1_epi64( 0x5555555555555555 ),  1 )
+#define W81(x)   Wz_8W(x, _mm512_set1_epi64( 0x3333333333333333 ),  2 )
+#define W82(x)   Wz_8W(x, _mm512_set1_epi64( 0x0F0F0F0F0F0F0F0F ),  4 )
+#define W83(x)   Wz_8W(x, _mm512_set1_epi64( 0x00FF00FF00FF00FF ),  8 ) 
+#define W84(x)   Wz_8W(x, _mm512_set1_epi64( 0x0000FFFF0000FFFF ), 16 )
+#define W85(x)   Wz_8W(x, _mm512_set1_epi64( 0x00000000FFFFFFFF ), 32 )
 #define W86(x) \
 do { \
   __m512i t = x ## h; \
@@ -353,12 +364,12 @@ do { \
   x ## l = _mm256_or_si256( _mm256_and_si256((x ## l >> (n)), (c)), t ); \
 } while (0)

-#define W0(x)   Wz(x, m256_const1_64( 0x5555555555555555 ),  1 )
-#define W1(x)   Wz(x, m256_const1_64( 0x3333333333333333 ),  2 )
-#define W2(x)   Wz(x, m256_const1_64( 0x0F0F0F0F0F0F0F0F ),  4 )
-#define W3(x)   Wz(x, m256_const1_64( 0x00FF00FF00FF00FF ),  8 ) 
-#define W4(x)   Wz(x, m256_const1_64( 0x0000FFFF0000FFFF ), 16 )
-#define W5(x)   Wz(x, m256_const1_64( 0x00000000FFFFFFFF ), 32 )
+#define W0(x)   Wz(x, _mm256_set1_epi64x( 0x5555555555555555 ),  1 )
+#define W1(x)   Wz(x, _mm256_set1_epi64x( 0x3333333333333333 ),  2 )
+#define W2(x)   Wz(x, _mm256_set1_epi64x( 0x0F0F0F0F0F0F0F0F ),  4 )
+#define W3(x)   Wz(x, _mm256_set1_epi64x( 0x00FF00FF00FF00FF ),  8 ) 
+#define W4(x)   Wz(x, _mm256_set1_epi64x( 0x0000FFFF0000FFFF ), 16 )
+#define W5(x)   Wz(x, _mm256_set1_epi64x( 0x00000000FFFFFFFF ), 32 )
 #define W6(x) \
 do { \
   __m256i t = x ## h; \
@@ -625,22 +636,22 @@ static const sph_u64 IV512[] = {
 void jh256_8way_init( jh_8way_context *sc )
 {
    // bswapped IV256
-    sc->H[ 0] = m512_const1_64( 0xebd3202c41a398eb );
-    sc->H[ 1] = m512_const1_64( 0xc145b29c7bbecd92 );
-    sc->H[ 2] = m512_const1_64( 0xfac7d4609151931c );
-    sc->H[ 3] = m512_const1_64( 0x038a507ed6820026 );
-    sc->H[ 4] = m512_const1_64( 0x45b92677269e23a4 );
-    sc->H[ 5] = m512_const1_64( 0x77941ad4481afbe0 );
-    sc->H[ 6] = m512_const1_64( 0x7a176b0226abb5cd );
-    sc->H[ 7] = m512_const1_64( 0xa82fff0f4224f056 );
-    sc->H[ 8] = m512_const1_64( 0x754d2e7f8996a371 );
-    sc->H[ 9] = m512_const1_64( 0x62e27df70849141d );
-    sc->H[10] = m512_const1_64( 0x948f2476f7957627 );
-    sc->H[11] = m512_const1_64( 0x6c29804757b6d587 );
-    sc->H[12] = m512_const1_64( 0x6c0d8eac2d275e5c );
-    sc->H[13] = m512_const1_64( 0x0f7a0557c6508451 );
-    sc->H[14] = m512_const1_64( 0xea12247067d3e47b );
-    sc->H[15] = m512_const1_64( 0x69d71cd313abe389 );
+    sc->H[ 0] = _mm512_set1_epi64( 0xebd3202c41a398eb );
+    sc->H[ 1] = _mm512_set1_epi64( 0xc145b29c7bbecd92 );
+    sc->H[ 2] = _mm512_set1_epi64( 0xfac7d4609151931c );
+    sc->H[ 3] = _mm512_set1_epi64( 0x038a507ed6820026 );
+    sc->H[ 4] = _mm512_set1_epi64( 0x45b92677269e23a4 );
+    sc->H[ 5] = _mm512_set1_epi64( 0x77941ad4481afbe0 );
+    sc->H[ 6] = _mm512_set1_epi64( 0x7a176b0226abb5cd );
+    sc->H[ 7] = _mm512_set1_epi64( 0xa82fff0f4224f056 );
+    sc->H[ 8] = _mm512_set1_epi64( 0x754d2e7f8996a371 );
+    sc->H[ 9] = _mm512_set1_epi64( 0x62e27df70849141d );
+    sc->H[10] = _mm512_set1_epi64( 0x948f2476f7957627 );
+    sc->H[11] = _mm512_set1_epi64( 0x6c29804757b6d587 );
+    sc->H[12] = _mm512_set1_epi64( 0x6c0d8eac2d275e5c );
+    sc->H[13] = _mm512_set1_epi64( 0x0f7a0557c6508451 );
+    sc->H[14] = _mm512_set1_epi64( 0xea12247067d3e47b );
+    sc->H[15] = _mm512_set1_epi64( 0x69d71cd313abe389 );
    sc->ptr = 0;
    sc->block_count = 0;
 }
@@ -648,22 +659,22 @@ void jh256_8way_init( jh_8way_context *sc )
 void jh512_8way_init( jh_8way_context *sc )
 {
    // bswapped IV512
-    sc->H[ 0] = m512_const1_64( 0x17aa003e964bd16f );
-    sc->H[ 1] = m512_const1_64( 0x43d5157a052e6a63 );
-    sc->H[ 2] = m512_const1_64( 0x0bef970c8d5e228a );
-    sc->H[ 3] = m512_const1_64( 0x61c3b3f2591234e9 );
-    sc->H[ 4] = m512_const1_64( 0x1e806f53c1a01d89 );
-    sc->H[ 5] = m512_const1_64( 0x806d2bea6b05a92a );
-    sc->H[ 6] = m512_const1_64( 0xa6ba7520dbcc8e58 );
-    sc->H[ 7] = m512_const1_64( 0xf73bf8ba763a0fa9 );
-    sc->H[ 8] = m512_const1_64( 0x694ae34105e66901 );
-    sc->H[ 9] = m512_const1_64( 0x5ae66f2e8e8ab546 );
-    sc->H[10] = m512_const1_64( 0x243c84c1d0a74710 );
-    sc->H[11] = m512_const1_64( 0x99c15a2db1716e3b );
-    sc->H[12] = m512_const1_64( 0x56f8b19decf657cf );
-    sc->H[13] = m512_const1_64( 0x56b116577c8806a7 );
-    sc->H[14] = m512_const1_64( 0xfb1785e6dffcc2e3 );
-    sc->H[15] = m512_const1_64( 0x4bdd8ccc78465a54 );
+    sc->H[ 0] = _mm512_set1_epi64( 0x17aa003e964bd16f );
+    sc->H[ 1] = _mm512_set1_epi64( 0x43d5157a052e6a63 );
+    sc->H[ 2] = _mm512_set1_epi64( 0x0bef970c8d5e228a );
+    sc->H[ 3] = _mm512_set1_epi64( 0x61c3b3f2591234e9 );
+    sc->H[ 4] = _mm512_set1_epi64( 0x1e806f53c1a01d89 );
+    sc->H[ 5] = _mm512_set1_epi64( 0x806d2bea6b05a92a );
+    sc->H[ 6] = _mm512_set1_epi64( 0xa6ba7520dbcc8e58 );
+    sc->H[ 7] = _mm512_set1_epi64( 0xf73bf8ba763a0fa9 );
+    sc->H[ 8] = _mm512_set1_epi64( 0x694ae34105e66901 );
+    sc->H[ 9] = _mm512_set1_epi64( 0x5ae66f2e8e8ab546 );
+    sc->H[10] = _mm512_set1_epi64( 0x243c84c1d0a74710 );
+    sc->H[11] = _mm512_set1_epi64( 0x99c15a2db1716e3b );
+    sc->H[12] = _mm512_set1_epi64( 0x56f8b19decf657cf );
+    sc->H[13] = _mm512_set1_epi64( 0x56b116577c8806a7 );
+    sc->H[14] = _mm512_set1_epi64( 0xfb1785e6dffcc2e3 );
+    sc->H[15] = _mm512_set1_epi64( 0x4bdd8ccc78465a54 );
    sc->ptr = 0;
    sc->block_count = 0;
 }
@@ -722,7 +733,7 @@ jh_8way_close( jh_8way_context *sc, unsigned ub, unsigned n, void *dst,
   size_t numz, u;
   uint64_t l0, l1;

-   buf[0] = m512_const1_64( 0x80ULL );
+   buf[0] = _mm512_set1_epi64( 0x80ULL );

   if ( sc->ptr == 0 )
       numz = 48;
@@ -773,22 +784,22 @@ jh512_8way_close(void *cc, void *dst)
 void jh256_4way_init( jh_4way_context *sc )
 {
    // bswapped IV256
-    sc->H[ 0] = m256_const1_64( 0xebd3202c41a398eb );
-    sc->H[ 1] = m256_const1_64( 0xc145b29c7bbecd92 );
-    sc->H[ 2] = m256_const1_64( 0xfac7d4609151931c );
-    sc->H[ 3] = m256_const1_64( 0x038a507ed6820026 );
-    sc->H[ 4] = m256_const1_64( 0x45b92677269e23a4 );
-    sc->H[ 5] = m256_const1_64( 0x77941ad4481afbe0 );
-    sc->H[ 6] = m256_const1_64( 0x7a176b0226abb5cd );
-    sc->H[ 7] = m256_const1_64( 0xa82fff0f4224f056 );
-    sc->H[ 8] = m256_const1_64( 0x754d2e7f8996a371 );
-    sc->H[ 9] = m256_const1_64( 0x62e27df70849141d );
-    sc->H[10] = m256_const1_64( 0x948f2476f7957627 );
-    sc->H[11] = m256_const1_64( 0x6c29804757b6d587 );
-    sc->H[12] = m256_const1_64( 0x6c0d8eac2d275e5c );
-    sc->H[13] = m256_const1_64( 0x0f7a0557c6508451 );
-    sc->H[14] = m256_const1_64( 0xea12247067d3e47b );
-    sc->H[15] = m256_const1_64( 0x69d71cd313abe389 );
+    sc->H[ 0] = _mm256_set1_epi64x( 0xebd3202c41a398eb );
+    sc->H[ 1] = _mm256_set1_epi64x( 0xc145b29c7bbecd92 );
+    sc->H[ 2] = _mm256_set1_epi64x( 0xfac7d4609151931c );
+    sc->H[ 3] = _mm256_set1_epi64x( 0x038a507ed6820026 );
+    sc->H[ 4] = _mm256_set1_epi64x( 0x45b92677269e23a4 );
+    sc->H[ 5] = _mm256_set1_epi64x( 0x77941ad4481afbe0 );
+    sc->H[ 6] = _mm256_set1_epi64x( 0x7a176b0226abb5cd );
+    sc->H[ 7] = _mm256_set1_epi64x( 0xa82fff0f4224f056 );
+    sc->H[ 8] = _mm256_set1_epi64x( 0x754d2e7f8996a371 );
+    sc->H[ 9] = _mm256_set1_epi64x( 0x62e27df70849141d );
+    sc->H[10] = _mm256_set1_epi64x( 0x948f2476f7957627 );
+    sc->H[11] = _mm256_set1_epi64x( 0x6c29804757b6d587 );
+    sc->H[12] = _mm256_set1_epi64x( 0x6c0d8eac2d275e5c );
+    sc->H[13] = _mm256_set1_epi64x( 0x0f7a0557c6508451 );
+    sc->H[14] = _mm256_set1_epi64x( 0xea12247067d3e47b );
+    sc->H[15] = _mm256_set1_epi64x( 0x69d71cd313abe389 );
    sc->ptr = 0;
    sc->block_count = 0;
 }
@@ -796,22 +807,22 @@ void jh256_4way_init( jh_4way_context *sc )
 void jh512_4way_init( jh_4way_context *sc )
 {
    // bswapped IV512
-    sc->H[ 0] = m256_const1_64( 0x17aa003e964bd16f );
-    sc->H[ 1] = m256_const1_64( 0x43d5157a052e6a63 );
-    sc->H[ 2] = m256_const1_64( 0x0bef970c8d5e228a );
-    sc->H[ 3] = m256_const1_64( 0x61c3b3f2591234e9 );
-    sc->H[ 4] = m256_const1_64( 0x1e806f53c1a01d89 );
-    sc->H[ 5] = m256_const1_64( 0x806d2bea6b05a92a );
-    sc->H[ 6] = m256_const1_64( 0xa6ba7520dbcc8e58 );
-    sc->H[ 7] = m256_const1_64( 0xf73bf8ba763a0fa9 );
-    sc->H[ 8] = m256_const1_64( 0x694ae34105e66901 );
-    sc->H[ 9] = m256_const1_64( 0x5ae66f2e8e8ab546 );
-    sc->H[10] = m256_const1_64( 0x243c84c1d0a74710 );
-    sc->H[11] = m256_const1_64( 0x99c15a2db1716e3b );
-    sc->H[12] = m256_const1_64( 0x56f8b19decf657cf );
-    sc->H[13] = m256_const1_64( 0x56b116577c8806a7 );
-    sc->H[14] = m256_const1_64( 0xfb1785e6dffcc2e3 );
-    sc->H[15] = m256_const1_64( 0x4bdd8ccc78465a54 );
+    sc->H[ 0] = _mm256_set1_epi64x( 0x17aa003e964bd16f );
+    sc->H[ 1] = _mm256_set1_epi64x( 0x43d5157a052e6a63 );
+    sc->H[ 2] = _mm256_set1_epi64x( 0x0bef970c8d5e228a );
+    sc->H[ 3] = _mm256_set1_epi64x( 0x61c3b3f2591234e9 );
+    sc->H[ 4] = _mm256_set1_epi64x( 0x1e806f53c1a01d89 );
+    sc->H[ 5] = _mm256_set1_epi64x( 0x806d2bea6b05a92a );
+    sc->H[ 6] = _mm256_set1_epi64x( 0xa6ba7520dbcc8e58 );
+    sc->H[ 7] = _mm256_set1_epi64x( 0xf73bf8ba763a0fa9 );
+    sc->H[ 8] = _mm256_set1_epi64x( 0x694ae34105e66901 );
+    sc->H[ 9] = _mm256_set1_epi64x( 0x5ae66f2e8e8ab546 );
+    sc->H[10] = _mm256_set1_epi64x( 0x243c84c1d0a74710 );
+    sc->H[11] = _mm256_set1_epi64x( 0x99c15a2db1716e3b );
+    sc->H[12] = _mm256_set1_epi64x( 0x56f8b19decf657cf );
+    sc->H[13] = _mm256_set1_epi64x( 0x56b116577c8806a7 );
+    sc->H[14] = _mm256_set1_epi64x( 0xfb1785e6dffcc2e3 );
+    sc->H[15] = _mm256_set1_epi64x( 0x4bdd8ccc78465a54 );
    sc->ptr = 0;
    sc->block_count = 0;
 }
@@ -870,7 +881,7 @@ jh_4way_close( jh_4way_context *sc, unsigned ub, unsigned n, void *dst,
   size_t numz, u;
   uint64_t l0, l1;

-   buf[0] = m256_const1_64( 0x80ULL );
+   buf[0] = _mm256_set1_epi64x( 0x80ULL );

   if ( sc->ptr == 0 )
       numz = 48;
--- a/algo/jh/sph_jh.h
+++ b/algo/jh/sph_jh.h
@@ -41,7 +41,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for JH-224.
--- a/algo/keccak/keccak-4way.c
+++ b/algo/keccak/keccak-4way.c
@@ -2,7 +2,6 @@
 #include <stdlib.h>
 #include <string.h>
 #include <stdint.h>
-#include "sph_keccak.h"
 #include "keccak-hash-4way.h"

 #if defined(KECCAK_8WAY)
@@ -49,7 +48,7 @@ int scanhash_keccak_8way( struct work *work, uint32_t max_nonce,
          }
      }
      *noncev = _mm512_add_epi32( *noncev,
-                                  m512_const1_64( 0x0000000800000000 ) );
+                                  _mm512_set1_epi64( 0x0000000800000000 ) );
      n += 8;

   } while ( (n < max_nonce-8) && !work_restart[thr_id].restart);
@@ -101,7 +100,7 @@ int scanhash_keccak_4way( struct work *work, uint32_t max_nonce,
          }
      }
      *noncev = _mm256_add_epi32( *noncev,
-                                  m256_const1_64( 0x0000000400000000 ) );
+                                  _mm256_set1_epi64x( 0x0000000400000000 ) );
      n += 4;
   } while ( (n < max_nonce-4) && !work_restart[thr_id].restart);
   pdata[19] = n;
--- a/algo/keccak/keccak-gate.c
+++ b/algo/keccak/keccak-gate.c
@@ -9,7 +9,7 @@ int hard_coded_eb = 1;
 bool register_keccak_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT | AVX512_OPT;
-  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
+  gate->gen_merkle_root = (void*)&sha256_gen_merkle_root;
  opt_target_factor = 128.0;
 #if defined (KECCAK_8WAY)
  gate->scanhash  = (void*)&scanhash_keccak_8way;
--- a/algo/keccak/keccak-hash-4way.c
+++ b/algo/keccak/keccak-hash-4way.c
@@ -76,7 +76,7 @@ static const uint64_t RC[] = {
 #define XOR64              XOR
 #define AND64(d, a, b)     (d = _mm512_and_si512(a,b))
 #define OR64(d, a, b)      (d = _mm512_or_si512(a,b))
-#define NOT64(d, s)        (d = _mm512_xor_si512(s,m512_neg1))
+#define NOT64(d, s)        (d = mm512_not( s ) )
 #define ROL64(d, v, n)     (d = mm512_rol_64(v, n))
 #define XOROR(d, a, b, c)  (d = mm512_xoror(a, b, c))
 #define XORAND(d, a, b, c) (d = mm512_xorand(a, b, c))
@@ -180,15 +180,15 @@ static void keccak64_8way_close( keccak64_ctx_m512i *kc, void *dst,
    if ( kc->ptr == (lim - 8) )
    {
        const uint64_t t = eb | 0x8000000000000000;
-        u.tmp[0] = m512_const1_64( t );
+        u.tmp[0] = _mm512_set1_epi64( t );
        j = 8;
    }
    else
    {
        j = lim - kc->ptr;
-        u.tmp[0] = m512_const1_64( eb );
+        u.tmp[0] = _mm512_set1_epi64( eb );
        memset_zero_512( u.tmp + 1, (j>>3) - 2 );
-        u.tmp[ (j>>3) - 1] = m512_const1_64( 0x8000000000000000 );
+        u.tmp[ (j>>3) - 1] = _mm512_set1_epi64( 0x8000000000000000 );
    }
    keccak64_8way_core( kc, u.tmp, j, lim );
    /* Finalize the "lane complement" */
@@ -262,10 +262,10 @@ keccak512_8way_close(void *cc, void *dst)
 #define XOR64              XOR
 #define AND64(d, a, b)     (d = _mm256_and_si256(a,b))
 #define OR64(d, a, b)      (d = _mm256_or_si256(a,b))
-#define NOT64(d, s)      (d = _mm256_xor_si256(s,m256_neg1))
+#define NOT64(d, s)        (d = mm256_not( s ) )
 #define ROL64(d, v, n)     (d = mm256_rol_64(v, n))
-#define XOROR(d, a, b, c) (d = _mm256_xor_si256(a, _mm256_or_si256(b, c)))
-#define XORAND(d, a, b, c) (d = _mm256_xor_si256(a, _mm256_and_si256(b, c)))
+#define XOROR(d, a, b, c)  (d = mm256_xoror( a, b, c ) )
+#define XORAND(d, a, b, c) (d = mm256_xorand( a, b, c ) )
 #define XOR3( d, a, b, c ) (d = mm256_xor3( a, b, c ))

 #include "keccak-macros.c"
@@ -368,15 +368,15 @@ static void keccak64_close( keccak64_ctx_m256i *kc, void *dst, size_t byte_len,
    if ( kc->ptr == (lim - 8) )
    {
        const uint64_t t = eb | 0x8000000000000000;
-        u.tmp[0] = m256_const1_64( t );
+        u.tmp[0] = _mm256_set1_epi64x( t );
        j = 8;
    }
    else
    {
        j = lim - kc->ptr;
-        u.tmp[0] = m256_const1_64( eb );
+        u.tmp[0] = _mm256_set1_epi64x( eb );
        memset_zero_256( u.tmp + 1, (j>>3) - 2 );
-        u.tmp[ (j>>3) - 1] = m256_const1_64( 0x8000000000000000 );
+        u.tmp[ (j>>3) - 1] = _mm256_set1_epi64x( 0x8000000000000000 );
    }
    keccak64_core( kc, u.tmp, j, lim );
    /* Finalize the "lane complement" */
--- a/algo/keccak/keccak-hash-4way.h
+++ b/algo/keccak/keccak-hash-4way.h
@@ -1,45 +1,6 @@
-/* $Id: sph_keccak.h 216 2010-06-08 09:46:57Z tp $ */
-/**
- * Keccak interface. This is the interface for Keccak with the
- * recommended parameters for SHA-3, with output lengths 224, 256,
- * 384 and 512 bits.
- *
- * ==========================(LICENSE BEGIN)============================
- *
- * Copyright (c) 2007-2010  Projet RNRT SAPHIR
- * 
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- * 
- * The above copyright notice and this permission notice shall be
- * included in all copies or substantial portions of the Software.
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * ===========================(LICENSE END)=============================
- *
- * @file     sph_keccak.h
- * @author   Thomas Pornin <thomas.pornin@cryptolog.com>
- */
-
 #ifndef KECCAK_HASH_4WAY_H__
 #define KECCAK_HASH_4WAY_H__

-#ifdef __cplusplus
-extern "C"{
-#endif
-
 #ifdef  __AVX2__

 #include <stddef.h>
@@ -100,8 +61,4 @@ void keccak512_4way_addbits_and_close(

 #endif

-#ifdef __cplusplus
-}
-#endif
-
 #endif
--- a/algo/keccak/sha3d-4way.c
+++ b/algo/keccak/sha3d-4way.c
@@ -2,7 +2,6 @@
 #include <stdlib.h>
 #include <string.h>
 #include <stdint.h>
-#include "sph_keccak.h"
 #include "keccak-hash-4way.h"

 #if defined(KECCAK_8WAY)
@@ -56,7 +55,7 @@ int scanhash_sha3d_8way( struct work *work, uint32_t max_nonce,
          }
      }
      *noncev = _mm512_add_epi32( *noncev,
-                                  m512_const1_64( 0x0000000800000000 ) );
+                                  _mm512_set1_epi64( 0x0000000800000000 ) );
      n += 8;

   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
@@ -115,7 +114,7 @@ int scanhash_sha3d_4way( struct work *work, uint32_t max_nonce,
          }
      }
      *noncev = _mm256_add_epi32( *noncev,
-                                  m256_const1_64( 0x0000000400000000 ) );
+                                  _mm256_set1_epi64x( 0x0000000400000000 ) );
      n += 4;
   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
--- a/algo/keccak/sph_keccak.h
+++ b/algo/keccak/sph_keccak.h
@@ -41,7 +41,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for Keccak-224.
--- a/algo/lanehash/lane.h
+++ b/algo/lanehash/lane.h
@@ -23,7 +23,6 @@
 #define LANE_H

 #include <string.h>
-//#include "algo/sha/sha3-defs.h"
 #include <stdint.h>

 typedef unsigned char BitSequence;
--- a/algo/luffa/luffa-hash-2way.c
+++ b/algo/luffa/luffa-hash-2way.c
--- a/algo/luffa/luffa-hash-2way.h
+++ b/algo/luffa/luffa-hash-2way.h
@@ -23,7 +23,7 @@
 #if defined(__AVX2__)

 #include <immintrin.h>
-#include "algo/sha/sha3-defs.h"
+//#include "algo/sha/sha3-defs.h"
 #include "simd-utils.h"

 /* The length of digests*/
@@ -54,7 +54,7 @@
 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

 typedef struct {
-    uint32 buffer[8*4];
+    uint32_t buffer[8*4];
    __m512i chainv[10];   /* Chaining values */
    int hashbitlen;
    int rembytes;
@@ -82,7 +82,7 @@ int luffa512_4way_update_close( luffa_4way_context *state, void *output,
 #endif

 typedef struct {
-    uint32 buffer[8*2];
+    uint32_t buffer[8*2];
    __m256i chainv[10];   /* Chaining values */
    int hashbitlen;
    int rembytes;
--- a/algo/luffa/luffa_for_sse2.c
+++ b/algo/luffa/luffa_for_sse2.c
@@ -19,96 +19,122 @@
 */

 #include <string.h>
-#include <emmintrin.h>
 #include "simd-utils.h"
 #include "luffa_for_sse2.h"

+#define cns(i)  ( ( (__m128i*)CNS_INIT)[i] )
+
+#define ADD_CONSTANT( a, b, c0 ,c1 ) \
+    a = _mm_xor_si128( a, c0 ); \
+    b = _mm_xor_si128( b, c1 ); \
+
+#if defined(__AVX512VL__)
+//TODO enable for AVX10_512 AVX10_256
+
+#define MULT2( a0, a1 ) \
+{ \
+  __m128i b = _mm_xor_si128( a0, \
+                      _mm_maskz_shuffle_epi32( 0xb, a1, 0x10 ) ); \
+  a0 = _mm_alignr_epi8( a1, b, 4 ); \
+  a1 = _mm_alignr_epi8( b, a1, 4 ); \
+}
+
+#elif defined(__SSE4_1__)
+
 #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) ); \
-  a1 = _mm_or_si128( _mm_srli_si128(a1,4), _mm_slli_si128(b,12) );  \
+  __m128i b = _mm_xor_si128( a0, \
+                      _mm_shuffle_epi32( mm128_mask_32( a1, 0xe ), 0x10 ) ); \
+  a0 = _mm_alignr_epi8( a1, b, 4 ); \
+  a1 = _mm_alignr_epi8( b, a1, 4 ); \
 } while(0)

-/*
-static inline __m256i mult2_avx2( a )
-{ 
-   __m128 a0, a0, b;
-   a0 = mm128_extractlo_256( a );
-   a1 = mm128_extracthi_256( a );
-   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) );
-   a1 = _mm_or_si128( _mm_srli_si128(a1,4), _mm_slli_si128(b,12) );
-   return mm256_concat_128( a1, a0 );
+#else
+
+#define MULT2( a0, a1 ) do \
+{ \
+  __m128i b = _mm_xor_si128( a0, \
+                      _mm_shuffle_epi32( _mm_and_si128( a1, MASK ), 0x10 ) ); \
+  a0 = _mm_or_si128( _mm_srli_si128(  b, 4 ), _mm_slli_si128( a1, 12 ) ); \
+  a1 = _mm_or_si128( _mm_srli_si128( a1, 4 ), _mm_slli_si128(  b, 12 ) ); \
+} while(0)
+
+#endif
+
+#if defined(__AVX512VL__)
+//TODO enable for AVX10_512 AVX10_256
+
+#define SUBCRUMB( a0, a1, a2, a3 ) \
+{ \
+    __m128i t = a0; \
+    a0 = mm128_xoror( a3, a0, a1 ); \
+    a2 = _mm_xor_si128( a2, a3 ); \
+    a1 = _mm_ternarylogic_epi64( a1, a3, t, 0x87 ); /* a1 xnor (a3 & t) */ \
+    a3 = mm128_xorand( a2, a3, t ); \
+    a2 = mm128_xorand( a1, a2, a0 ); \
+    a1 = _mm_or_si128( a1, a3 ); \
+    a3 = _mm_xor_si128( a3, a2 ); \
+    t  = _mm_xor_si128( t, a1 ); \
+    a2 = _mm_and_si128( a2, a1 ); \
+    a1 = mm128_xnor( a1, a0 ); \
+    a0 = t; \
 }
-*/

-#define STEP_PART(x,c,t)\
-    SUBCRUMB(*x,*(x+1),*(x+2),*(x+3),*t);\
-    SUBCRUMB(*(x+5),*(x+6),*(x+7),*(x+4),*t);\
-    MIXWORD(*x,*(x+4),*t,*(t+1));\
-    MIXWORD(*(x+1),*(x+5),*t,*(t+1));\
-    MIXWORD(*(x+2),*(x+6),*t,*(t+1));\
-    MIXWORD(*(x+3),*(x+7),*t,*(t+1));\
-    ADD_CONSTANT(*x, *(x+4), *c, *(c+1));
+#else

-#define STEP_PART2(a0,a1,t0,t1,c0,c1,tmp0,tmp1)\
-    a1 = _mm_shuffle_epi32(a1,147);\
-    t0 = _mm_load_si128(&a1);\
-    a1 = _mm_unpacklo_epi32(a1,a0);\
-    t0 = _mm_unpackhi_epi32(t0,a0);\
-    t1 = _mm_shuffle_epi32(t0,78);\
-    a0 = _mm_shuffle_epi32(a1,78);\
-    SUBCRUMB(t1,t0,a0,a1,tmp0);\
-    t0 = _mm_unpacklo_epi32(t0,t1);\
-    a1 = _mm_unpacklo_epi32(a1,a0);\
-    a0 = _mm_load_si128(&a1);\
-    a0 = _mm_unpackhi_epi64(a0,t0);\
-    a1 = _mm_unpacklo_epi64(a1,t0);\
-    a1 = _mm_shuffle_epi32(a1,57);\
-    MIXWORD(a0,a1,tmp0,tmp1);\
-    ADD_CONSTANT(a0,a1,c0,c1);
+#define SUBCRUMB( a0, a1, a2, a3 ) \
+{ \
+    __m128i t = a0; \
+    a0 = _mm_or_si128( a0, a1 ); \
+    a2 = _mm_xor_si128( a2, a3 ); \
+    a1 = mm128_not( a1 ); \
+    a0 = _mm_xor_si128( a0, a3 ); \
+    a3 = _mm_and_si128( a3, t ); \
+    a1 = _mm_xor_si128( a1, a3 ); \
+    a3 = _mm_xor_si128( a3, a2 ); \
+    a2 = _mm_and_si128( a2, a0 ); \
+    a0 = mm128_not( a0 ); \
+    a2 = _mm_xor_si128( a2, a1 ); \
+    a1 = _mm_or_si128(  a1, a3 ); \
+    t  = _mm_xor_si128( t , a1 ); \
+    a3 = _mm_xor_si128( a3, a2 ); \
+    a2 = _mm_and_si128( a2, a1 ); \
+    a1 = _mm_xor_si128( a1, a0 ); \
+    a0 = t; \
+}

-#define SUBCRUMB(a0,a1,a2,a3,t)\
-    t  = _mm_load_si128(&a0);\
-    a0 = _mm_or_si128(a0,a1);\
-    a2 = _mm_xor_si128(a2,a3);\
-    a1 = _mm_andnot_si128(a1,ALLONE);\
-    a0 = _mm_xor_si128(a0,a3);\
-    a3 = _mm_and_si128(a3,t);\
-    a1 = _mm_xor_si128(a1,a3);\
-    a3 = _mm_xor_si128(a3,a2);\
-    a2 = _mm_and_si128(a2,a0);\
-    a0 = _mm_andnot_si128(a0,ALLONE);\
-    a2 = _mm_xor_si128(a2,a1);\
-    a1 = _mm_or_si128(a1,a3);\
-    t  = _mm_xor_si128(t,a1);\
-    a3 = _mm_xor_si128(a3,a2);\
-    a2 = _mm_and_si128(a2,a1);\
-    a1 = _mm_xor_si128(a1,a0);\
-    a0 = _mm_load_si128(&t);\
+#endif

-#define MIXWORD(a,b,t1,t2)\
-    b = _mm_xor_si128(a,b);\
-    t1 = _mm_slli_epi32(a,2);\
-    t2 = _mm_srli_epi32(a,30);\
-    a = _mm_or_si128(t1,t2);\
-    a = _mm_xor_si128(a,b);\
-    t1 = _mm_slli_epi32(b,14);\
-    t2 = _mm_srli_epi32(b,18);\
-    b = _mm_or_si128(t1,t2);\
-    b = _mm_xor_si128(a,b);\
-    t1 = _mm_slli_epi32(a,10);\
-    t2 = _mm_srli_epi32(a,22);\
-    a = _mm_or_si128(t1,t2);\
-    a = _mm_xor_si128(a,b);\
-    t1 = _mm_slli_epi32(b,1);\
-    t2 = _mm_srli_epi32(b,31);\
-    b = _mm_or_si128(t1,t2);
+#define MIXWORD( a, b ) \
+    b = _mm_xor_si128( a, b ); \
+    a = _mm_xor_si128( b, mm128_rol_32( a, 2 ) ); \
+    b = _mm_xor_si128( a, mm128_rol_32( b, 14 ) ); \
+    a = _mm_xor_si128( b, mm128_rol_32( a, 10 ) ); \
+    b = mm128_rol_32( b, 1 );

-#define ADD_CONSTANT(a,b,c0,c1)\
-    a = _mm_xor_si128(a,c0);\
-    b = _mm_xor_si128(b,c1);\
+#define STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, c0, c1 ) \
+    SUBCRUMB( x0, x1, x2, x3 ); \
+    SUBCRUMB( x5, x6, x7, x4 ); \
+    MIXWORD( x0, x4 ); \
+    MIXWORD( x1, x5 ); \
+    MIXWORD( x2, x6 ); \
+    MIXWORD( x3, x7 ); \
+    ADD_CONSTANT( x0, x4, c0, c1 );
+
+#define STEP_PART2( a0, a1, t0, t1, c0, c1 ) \
+    t0 = _mm_shuffle_epi32( a1, 147 ); \
+    a1 = _mm_unpacklo_epi32( t0, a0 ); \
+    t0 = _mm_unpackhi_epi32( t0, a0 ); \
+    t1 = _mm_shuffle_epi32( t0, 78 ); \
+    a0 = _mm_shuffle_epi32( a1, 78 ); \
+    SUBCRUMB( t1, t0, a0, a1 ); \
+    t0 = _mm_unpacklo_epi32( t0, t1 ); \
+    a1 = _mm_unpacklo_epi32( a1, a0 ); \
+    a0 = _mm_unpackhi_epi64( a1, t0 ); \
+    a1 = _mm_unpacklo_epi64( a1, t0 ); \
+    a1 = _mm_shuffle_epi32( a1, 57 ); \
+    MIXWORD( a0, a1 ); \
+    ADD_CONSTANT( a0, a1, c0, c1 );

 #define NMLTOM768(r0,r1,r2,s0,s1,s2,s3,p0,p1,p2,q0,q1,q2,q3)\
    s2 = _mm_load_si128(&r1);\
@@ -169,32 +195,22 @@ static inline __m256i mult2_avx2( a )
    q1 = _mm_load_si128(&p1);\

 #define NMLTOM1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\
-    s1 = _mm_load_si128(&r3);\
-    q1 = _mm_load_si128(&p3);\
-    s3 = _mm_load_si128(&r3);\
-    q3 = _mm_load_si128(&p3);\
-    s1 = _mm_unpackhi_epi32(s1,r2);\
-    q1 = _mm_unpackhi_epi32(q1,p2);\
-    s3 = _mm_unpacklo_epi32(s3,r2);\
-    q3 = _mm_unpacklo_epi32(q3,p2);\
-    s0 = _mm_load_si128(&s1);\
-    q0 = _mm_load_si128(&q1);\
-    s2 = _mm_load_si128(&s3);\
-    q2 = _mm_load_si128(&q3);\
-    r3 = _mm_load_si128(&r1);\
-    p3 = _mm_load_si128(&p1);\
-    r1 = _mm_unpacklo_epi32(r1,r0);\
-    p1 = _mm_unpacklo_epi32(p1,p0);\
-    r3 = _mm_unpackhi_epi32(r3,r0);\
-    p3 = _mm_unpackhi_epi32(p3,p0);\
-    s0 = _mm_unpackhi_epi64(s0,r3);\
-    q0 = _mm_unpackhi_epi64(q0,p3);\
-    s1 = _mm_unpacklo_epi64(s1,r3);\
-    q1 = _mm_unpacklo_epi64(q1,p3);\
-    s2 = _mm_unpackhi_epi64(s2,r1);\
-    q2 = _mm_unpackhi_epi64(q2,p1);\
-    s3 = _mm_unpacklo_epi64(s3,r1);\
-    q3 = _mm_unpacklo_epi64(q3,p1);
+    s1 = _mm_unpackhi_epi32( r3, r2 ); \
+    q1 = _mm_unpackhi_epi32( p3, p2 ); \
+    s3 = _mm_unpacklo_epi32( r3, r2 ); \
+    q3 = _mm_unpacklo_epi32( p3, p2 ); \
+    r3 = _mm_unpackhi_epi32( r1, r0 ); \
+    r1 = _mm_unpacklo_epi32( r1, r0 ); \
+    p3 = _mm_unpackhi_epi32( p1, p0 ); \
+    p1 = _mm_unpacklo_epi32( p1, p0 ); \
+    s0 = _mm_unpackhi_epi64( s1, r3 ); \
+    q0 = _mm_unpackhi_epi64( q1 ,p3 ); \
+    s1 = _mm_unpacklo_epi64( s1, r3 ); \
+    q1 = _mm_unpacklo_epi64( q1, p3 ); \
+    s2 = _mm_unpackhi_epi64( s3, r1 ); \
+    q2 = _mm_unpackhi_epi64( q3, p1 ); \
+    s3 = _mm_unpacklo_epi64( s3, r1 ); \
+    q3 = _mm_unpacklo_epi64( q3, p1 );

 #define MIXTON1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3)\
    NMLTOM1024(r0,r1,r2,r3,s0,s1,s2,s3,p0,p1,p2,p3,q0,q1,q2,q3);
@@ -255,17 +271,18 @@ static const uint32 CNS_INIT[128] __attribute((aligned(16))) = {


 __m128i CNS128[32];
-__m128i ALLONE;
+#if !defined(__SSE4_1__)
 __m128i MASK;
+#endif

 HashReturn init_luffa(hashState_luffa *state, int hashbitlen)
 {
    int i;
    state->hashbitlen = hashbitlen;
+#if !defined(__SSE4_1__)
    /* set the lower 32 bits to '1' */
    MASK= _mm_set_epi32(0x00000000, 0x00000000, 0x00000000, 0xffffffff);
-    /* set all bits to '1' */
-    ALLONE = _mm_set_epi32(0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff);
+#endif
    /* set the 32-bit round constant values to the 128-bit data field */
    for ( i=0; i<32; i++ )
        CNS128[i] = _mm_load_si128( (__m128i*)&CNS_INIT[i*4] );
@@ -297,8 +314,7 @@ HashReturn update_luffa( hashState_luffa *state, const BitSequence *data,
      // remaining data bytes
      casti_m128i( state->buffer, 0 ) = mm128_bswap_32( cast_m128i( data ) );
      // padding of partial block
-      casti_m128i( state->buffer, 1 ) =
-            _mm_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 );
+      casti_m128i( state->buffer, 1 ) =  _mm_set_epi32( 0, 0, 0, 0x80000000 );
    }

    return SUCCESS;
@@ -316,8 +332,7 @@ HashReturn final_luffa(hashState_luffa *state, BitSequence *hashval)
    else
    {
      // empty pad block, constant data
-     rnd512( state, _mm_setzero_si128(),
-                       _mm_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 ) );
+     rnd512( state, _mm_setzero_si128(), _mm_set_epi32( 0, 0, 0, 0x80000000 ) );
    }

    finalization512(state, (uint32*) hashval);
@@ -345,11 +360,11 @@ 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_i128(  0x80000000 ),
+       rnd512( state, mm128_mov64_128(  0x80000000 ),
                      mm128_bswap_32( cast_m128i( data ) ) );
    else
       // empty pad block
-       rnd512( state, m128_zero, m128_const_i128( 0x80000000 ) );
+       rnd512( state, m128_zero, mm128_mov64_128( 0x80000000 ) );

    finalization512( state, (uint32*) output );
    if ( state->hashbitlen > 512 )
@@ -365,10 +380,10 @@ int luffa_full( hashState_luffa *state, BitSequence* output, int hashbitlen,
 // Optimized for integrals of 16 bytes, good for 64 and 80 byte len
    int i;
    state->hashbitlen = hashbitlen;
+#if !defined(__SSE4_1__)
    /* set the lower 32 bits to '1' */
    MASK= _mm_set_epi32(0x00000000, 0x00000000, 0x00000000, 0xffffffff);
-    /* set all bits to '1' */
-    ALLONE = _mm_set_epi32(0xffffffff, 0xffffffff, 0xffffffff, 0xffffffff);
+#endif
    /* set the 32-bit round constant values to the 128-bit data field */
    for ( i=0; i<32; i++ )
        CNS128[i] = _mm_load_si128( (__m128i*)&CNS_INIT[i*4] );
@@ -394,11 +409,11 @@ 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_i128( 0x80000000 ),
+       rnd512( state, mm128_mov64_128( 0x80000000 ),
                      mm128_bswap_32( cast_m128i( data ) ) );
    else
       // empty pad block
-       rnd512( state, m128_zero, m128_const_i128( 0x80000000 ) );
+       rnd512( state, m128_zero, mm128_mov64_128( 0x80000000 ) );

    finalization512( state, (uint32*) output );
    if ( state->hashbitlen > 512 )
@@ -414,163 +429,119 @@ int luffa_full( hashState_luffa *state, BitSequence* output, int hashbitlen,

 static void rnd512( hashState_luffa *state, __m128i msg1, __m128i msg0 )
 {
-    __m128i t[2];
+    __m128i t0, t1;
    __m128i *chainv = state->chainv;
-    __m128i tmp[2];
-    __m128i x[8];
+    __m128i x0, x1, x2, x3, x4, x5, x6, x7; 

-    t[0] = chainv[0];
-    t[1] = chainv[1];
+    t0 = mm128_xor3( chainv[0], chainv[2], chainv[4] );
+    t1 = mm128_xor3( chainv[1], chainv[3], chainv[5] );
+    t0 = mm128_xor3( t0, chainv[6], chainv[8] );
+    t1 = mm128_xor3( t1, chainv[7], chainv[9] );

-    t[0] = _mm_xor_si128( t[0], chainv[2] );
-    t[1] = _mm_xor_si128( t[1], chainv[3] );
-    t[0] = _mm_xor_si128( t[0], chainv[4] );
-    t[1] = _mm_xor_si128( t[1], chainv[5] );
-    t[0] = _mm_xor_si128( t[0], chainv[6] );
-    t[1] = _mm_xor_si128( t[1], chainv[7] );
-    t[0] = _mm_xor_si128( t[0], chainv[8] );
-    t[1] = _mm_xor_si128( t[1], chainv[9] );
-
-    MULT2( t[0], t[1] );
+    MULT2( t0, t1 );

    msg0 = _mm_shuffle_epi32( msg0, 27 );
    msg1 = _mm_shuffle_epi32( msg1, 27 );

-    chainv[0] = _mm_xor_si128( chainv[0], t[0] );
-    chainv[1] = _mm_xor_si128( chainv[1], t[1] );
-    chainv[2] = _mm_xor_si128( chainv[2], t[0] );
-    chainv[3] = _mm_xor_si128( chainv[3], t[1] );
-    chainv[4] = _mm_xor_si128( chainv[4], t[0] );
-    chainv[5] = _mm_xor_si128( chainv[5], t[1] );
-    chainv[6] = _mm_xor_si128( chainv[6], t[0] );
-    chainv[7] = _mm_xor_si128( chainv[7], t[1] );
-    chainv[8] = _mm_xor_si128( chainv[8], t[0] );
-    chainv[9] = _mm_xor_si128( chainv[9], t[1] );
+    chainv[0] = _mm_xor_si128( chainv[0], t0 );
+    chainv[1] = _mm_xor_si128( chainv[1], t1 );
+    chainv[2] = _mm_xor_si128( chainv[2], t0 );
+    chainv[3] = _mm_xor_si128( chainv[3], t1 );
+    chainv[4] = _mm_xor_si128( chainv[4], t0 );
+    chainv[5] = _mm_xor_si128( chainv[5], t1 );
+    chainv[6] = _mm_xor_si128( chainv[6], t0 );
+    chainv[7] = _mm_xor_si128( chainv[7], t1 );
+    chainv[8] = _mm_xor_si128( chainv[8], t0 );
+    chainv[9] = _mm_xor_si128( chainv[9], t1 );

-    t[0] = chainv[0];
-    t[1] = chainv[1];
+    t0 = chainv[0];
+    t1 = chainv[1];

    MULT2( chainv[0], chainv[1]);
-
    chainv[0] = _mm_xor_si128( chainv[0], chainv[2] );
    chainv[1] = _mm_xor_si128( chainv[1], chainv[3] );

    MULT2( chainv[2], chainv[3]);
-
    chainv[2] = _mm_xor_si128(chainv[2], chainv[4]);
    chainv[3] = _mm_xor_si128(chainv[3], chainv[5]);

    MULT2( chainv[4], chainv[5]);
-
    chainv[4] = _mm_xor_si128(chainv[4], chainv[6]);
    chainv[5] = _mm_xor_si128(chainv[5], chainv[7]);

    MULT2( chainv[6], chainv[7]);
-
    chainv[6] = _mm_xor_si128(chainv[6], chainv[8]);
    chainv[7] = _mm_xor_si128(chainv[7], chainv[9]);

    MULT2( chainv[8], chainv[9]);
-
-    chainv[8] = _mm_xor_si128( chainv[8], t[0] );
-    chainv[9] = _mm_xor_si128( chainv[9], t[1] );
-
-    t[0] = chainv[8];
-    t[1] = chainv[9];
+    t0 = chainv[8] = _mm_xor_si128( chainv[8], t0 );
+    t1 = chainv[9] = _mm_xor_si128( chainv[9], t1 );

    MULT2( chainv[8], chainv[9]);
-
    chainv[8] = _mm_xor_si128( chainv[8], chainv[6] );
    chainv[9] = _mm_xor_si128( chainv[9], chainv[7] );

    MULT2( chainv[6], chainv[7]);
-
    chainv[6] = _mm_xor_si128( chainv[6], chainv[4] );
    chainv[7] = _mm_xor_si128( chainv[7], chainv[5] );

    MULT2( chainv[4], chainv[5]);
-
    chainv[4] = _mm_xor_si128( chainv[4], chainv[2] );
    chainv[5] = _mm_xor_si128( chainv[5], chainv[3] );

    MULT2( chainv[2], chainv[3] );
-
    chainv[2] = _mm_xor_si128( chainv[2], chainv[0] );
    chainv[3] = _mm_xor_si128( chainv[3], chainv[1] );

    MULT2( chainv[0], chainv[1] );
-
-    chainv[0] = _mm_xor_si128( _mm_xor_si128( chainv[0], t[0] ), msg0 );
-    chainv[1] = _mm_xor_si128( _mm_xor_si128( chainv[1], t[1] ), msg1 );
+    chainv[0] = _mm_xor_si128( _mm_xor_si128( chainv[0], t0 ), msg0 );
+    chainv[1] = _mm_xor_si128( _mm_xor_si128( chainv[1], t1 ), msg1 );

    MULT2( msg0, msg1);
-
    chainv[2] = _mm_xor_si128( chainv[2], msg0 );
    chainv[3] = _mm_xor_si128( chainv[3], msg1 );

    MULT2( msg0, msg1);
-
    chainv[4] = _mm_xor_si128( chainv[4], msg0 );
    chainv[5] = _mm_xor_si128( chainv[5], msg1 );

    MULT2( msg0, msg1);
-
    chainv[6] = _mm_xor_si128( chainv[6], msg0 );
    chainv[7] = _mm_xor_si128( chainv[7], msg1 );

    MULT2( msg0, msg1);
-
    chainv[8] = _mm_xor_si128( chainv[8], msg0 );
    chainv[9] = _mm_xor_si128( chainv[9], msg1 );

    MULT2( msg0, msg1);
+    chainv[3] = mm128_rol_32( chainv[3], 1 );    
+    chainv[5] = mm128_rol_32( chainv[5], 2 );
+    chainv[7] = mm128_rol_32( chainv[7], 3 );
+    chainv[9] = mm128_rol_32( chainv[9], 4 );
    
-    chainv[3] = _mm_or_si128( _mm_slli_epi32(chainv[3], 1),
-                              _mm_srli_epi32(chainv[3], 31) );
-    chainv[5] = _mm_or_si128( _mm_slli_epi32(chainv[5], 2),
-                              _mm_srli_epi32(chainv[5], 30) );
-    chainv[7] = _mm_or_si128( _mm_slli_epi32(chainv[7], 3),
-                              _mm_srli_epi32(chainv[7], 29) );
-    chainv[9] = _mm_or_si128( _mm_slli_epi32(chainv[9], 4),
-                              _mm_srli_epi32(chainv[9], 28) );
+    NMLTOM1024( chainv[0], chainv[2], chainv[4], chainv[6], x0, x1, x2, x3,
+                chainv[1], chainv[3], chainv[5], chainv[7], x4, x5, x6, x7 );

+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 0), cns( 1) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 2), cns( 3) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 4), cns( 5) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 6), cns( 7) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns( 8), cns( 9) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(10), cns(11) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(12), cns(13) );
+    STEP_PART( x0, x1, x2, x3, x4, x5, x6, x7, cns(14), cns(15) );
    
-    NMLTOM1024( chainv[0], chainv[2], chainv[4], chainv[6],
-                x[0], x[1], x[2], x[3],
-                chainv[1],chainv[3],chainv[5],chainv[7],
-                x[4], x[5], x[6], x[7] );
+    MIXTON1024( x0, x1, x2, x3, chainv[0], chainv[2], chainv[4], chainv[6],
+                x4, x5, x6, x7, chainv[1], chainv[3], chainv[5], chainv[7]);

-    STEP_PART( &x[0], &CNS128[ 0], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[ 2], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[ 4], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[ 6], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[ 8], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[10], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[12], &tmp[0] );
-    STEP_PART( &x[0], &CNS128[14], &tmp[0] );
-
-    MIXTON1024( x[0], x[1], x[2], x[3],
-                chainv[0], chainv[2], chainv[4],chainv[6],
-                x[4], x[5], x[6], x[7],
-                chainv[1],chainv[3],chainv[5],chainv[7]);
-
-    /* Process last 256-bit block */
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[16], CNS128[17],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[18], CNS128[19],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[20], CNS128[21],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[22], CNS128[23],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[24], CNS128[25],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[26], CNS128[27],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[28], CNS128[29],
-                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t[0], t[1], CNS128[30], CNS128[31],
-                tmp[0], tmp[1] );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(16), cns(17) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(18), cns(19) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(20), cns(21) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(22), cns(23) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(24), cns(25) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(26), cns(27) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(28), cns(29) );
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(30), cns(31) );
 }


@@ -579,51 +550,6 @@ static void rnd512( hashState_luffa *state, __m128i msg1, __m128i msg0 )
 /* state: hash context    */
 /* b[8]: hash values      */

-#if defined (__AVX2__)
-
-static void finalization512( hashState_luffa *state, uint32 *b )
-{
-    uint32   hash[8] __attribute((aligned(64)));
-    __m256i* chainv = (__m256i*)state->chainv;
-    __m256i  t;
-    const __m128i zero = m128_zero;
-    const __m256i shuff_bswap32 = m256_const_64( 0x1c1d1e1f18191a1b,
-                                                 0x1415161710111213,
-                                                 0x0c0d0e0f08090a0b,
-                                                 0x0405060700010203 );
-
-    rnd512( state, zero, zero );
-
-    t = chainv[0];
-    t = _mm256_xor_si256( t, chainv[1] );
-    t = _mm256_xor_si256( t, chainv[2] );
-    t = _mm256_xor_si256( t, chainv[3] );
-    t = _mm256_xor_si256( t, chainv[4] );
-
-    t = _mm256_shuffle_epi32( t, 27 );
-
-    _mm256_store_si256( (__m256i*)hash, t );
-
-    casti_m256i( b, 0 ) = _mm256_shuffle_epi8(
-                                 casti_m256i( hash, 0 ), shuff_bswap32 );
-
-    rnd512( state, zero, zero );
-
-    t = chainv[0];
-    t = _mm256_xor_si256( t, chainv[1] );
-    t = _mm256_xor_si256( t, chainv[2] );
-    t = _mm256_xor_si256( t, chainv[3] );
-    t = _mm256_xor_si256( t, chainv[4] );
-    t = _mm256_shuffle_epi32( t, 27 );
-
-    _mm256_store_si256( (__m256i*)hash, t );
-
-    casti_m256i( b, 1 ) = _mm256_shuffle_epi8( 
-                                 casti_m256i( hash, 0 ), shuff_bswap32 );
-}
-
-#else
-
 static void finalization512( hashState_luffa *state, uint32 *b )
 {
    uint32 hash[8] __attribute((aligned(64)));
@@ -676,6 +602,5 @@ static void finalization512( hashState_luffa *state, uint32 *b )
    casti_m128i( b, 2 ) = mm128_bswap_32( casti_m128i( hash, 0 ) );
    casti_m128i( b, 3 ) = mm128_bswap_32( casti_m128i( hash, 1 ) );
 }
-#endif

 /***************************************************/
--- a/algo/luffa/luffa_for_sse2.h
+++ b/algo/luffa/luffa_for_sse2.h
@@ -22,7 +22,7 @@
 */

 #include <emmintrin.h>
-#include "algo/sha/sha3-defs.h"
+#include "compat/sha3-defs.h"
 /* The length of digests*/
 #define DIGEST_BIT_LEN_224 224
 #define DIGEST_BIT_LEN_256 256
--- a/algo/luffa/sph_luffa.h
+++ b/algo/luffa/sph_luffa.h
@@ -41,7 +41,7 @@ extern "C"{
 #endif

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for Luffa-224.
--- a/algo/lyra2/allium-4way.c
+++ b/algo/lyra2/allium-4way.c
@@ -48,7 +48,7 @@ static void allium_16way_hash( void *state, const void *midstate_vars,
   uint32_t hash15[8] __attribute__ ((aligned (32)));
   allium_16way_ctx_holder ctx __attribute__ ((aligned (64)));

-   blake256_16way_final_rounds_le( vhash, midstate_vars, midhash, block );
+   blake256_16way_final_rounds_le( vhash, midstate_vars, midhash, block, 14 );

   dintrlv_16x32( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
                  hash8, hash9, hash10, hash11, hash12, hash13, hash14, hash15,
@@ -212,12 +212,12 @@ int scanhash_allium_16way( struct work *work, uint32_t max_nonce,
   const uint32_t last_nonce = max_nonce - 16;
   const int thr_id = mythr->id;
   const bool bench = opt_benchmark;
-   const __m512i sixteen = m512_const1_32( 16 );
+   const __m512i sixteen = _mm512_set1_epi32( 16 );

   if ( bench ) ( (uint32_t*)ptarget )[7] = 0x0000ff;

   // Prehash first block.
-   blake256_transform_le( phash, pdata, 512, 0 );
+   blake256_transform_le( phash, pdata, 512, 0, 14 );

   // Interleave hash for second block prehash.
   block0_hash[0] = _mm512_set1_epi32( phash[0] );
@@ -230,25 +230,13 @@ int scanhash_allium_16way( struct work *work, uint32_t max_nonce,
   block0_hash[7] = _mm512_set1_epi32( phash[7] );

   // Build vectored second block, interleave last 16 bytes of data using
-   // unique nonces, add padding.
+   // unique nonces.
   block_buf[ 0] = _mm512_set1_epi32( pdata[16] );
   block_buf[ 1] = _mm512_set1_epi32( pdata[17] );
   block_buf[ 2] = _mm512_set1_epi32( pdata[18] );
   block_buf[ 3] =
             _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 );
-   block_buf[ 4] = m512_const1_32( 0x80000000 );
-   block_buf[ 5] =
-   block_buf[ 6] = 
-   block_buf[ 7] = 
-   block_buf[ 8] = 
-   block_buf[ 9] = 
-   block_buf[10] = 
-   block_buf[11] = 
-   block_buf[12] = m512_zero;
-   block_buf[13] = m512_one_32;
-   block_buf[14] = m512_zero;
-   block_buf[15] = m512_const1_32( 80*8 );

   // Partialy prehash second block without touching nonces in block_buf[3].
   blake256_16way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
@@ -298,7 +286,7 @@ static void allium_8way_hash( void *hash, const void *midstate_vars,
   uint64_t *hash7 = (uint64_t*)hash+28;
   allium_8way_ctx_holder ctx __attribute__ ((aligned (64))); 

-   blake256_8way_final_rounds_le( vhashA, midstate_vars, midhash, block );
+   blake256_8way_final_rounds_le( vhashA, midstate_vars, midhash, block, 14 );

   dintrlv_8x32( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
                 vhashA, 256 );
@@ -410,10 +398,10 @@ int scanhash_allium_8way( struct work *work, uint32_t max_nonce,
   uint32_t n = first_nonce;
   const int thr_id = mythr->id;  
   const bool bench = opt_benchmark;
-   const __m256i eight = m256_const1_32( 8 );
+   const __m256i eight = _mm256_set1_epi32( 8 );

   // Prehash first block
-   blake256_transform_le( phash, pdata, 512, 0 );
+   blake256_transform_le( phash, pdata, 512, 0, 14 );

   block0_hash[0] = _mm256_set1_epi32( phash[0] );
   block0_hash[1] = _mm256_set1_epi32( phash[1] );
@@ -425,24 +413,12 @@ int scanhash_allium_8way( struct work *work, uint32_t max_nonce,
   block0_hash[7] = _mm256_set1_epi32( phash[7] );

   // Build vectored second block, interleave last 16 bytes of data using
-   // unique nonces and add padding.
+   // unique nonces.
   block_buf[ 0] = _mm256_set1_epi32( pdata[16] );
   block_buf[ 1] = _mm256_set1_epi32( pdata[17] );
   block_buf[ 2] = _mm256_set1_epi32( pdata[18] );
-   block_buf[ 3] =
-            _mm256_set_epi32( n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n+ 1, n );
-   block_buf[ 4] = m256_const1_32( 0x80000000 );
-   block_buf[ 5] =
-   block_buf[ 6] =
-   block_buf[ 7] =
-   block_buf[ 8] =
-   block_buf[ 9] =
-   block_buf[10] =
-   block_buf[11] =
-   block_buf[12] = m256_zero;
-   block_buf[13] = m256_one_32;
-   block_buf[14] = m256_zero;
-   block_buf[15] = m256_const1_32( 80*8 );
+   block_buf[ 3] = _mm256_set_epi32( n+ 7, n+ 6, n+ 5, n+ 4,
+                                     n+ 3, n+ 2, n+ 1, n );

   // Partialy prehash second block without touching nonces
   blake256_8way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
--- a/algo/lyra2/lyra2.h
+++ b/algo/lyra2/lyra2.h
@@ -21,9 +21,8 @@
 #define LYRA2_H_

 #include <stdint.h>
-#include "algo/sha/sha3-defs.h"

-//typedef unsigned char byte;
+typedef unsigned char byte;

 //Block length required so Blake2's Initialization Vector (IV) is not overwritten (THIS SHOULD NOT BE MODIFIED)
 #define BLOCK_LEN_BLAKE2_SAFE_INT64 8                                   //512 bits (=64 bytes, =8 uint64_t)
--- a/algo/lyra2/lyra2rev2-4way.c
+++ b/algo/lyra2/lyra2rev2-4way.c
@@ -75,7 +75,7 @@ void lyra2rev2_16way_hash( void *state, const void *input )
   keccak256_8way_close( &ctx.keccak, vhash );

   dintrlv_8x64( hash8,  hash9,  hash10,  hash11,
-                 hash12, hash13, hash14, hash5, vhash, 256 );
+                 hash12, hash13, hash14, hash15, vhash, 256 );

   cubehash_full( &ctx.cube, (byte*) hash0,  256, (const byte*) hash0,  32 );
   cubehash_full( &ctx.cube, (byte*) hash1,  256, (const byte*) hash1,  32 );
@@ -203,7 +203,7 @@ int scanhash_lyra2rev2_16way( struct work *work, const uint32_t max_nonce,
             submit_solution( work, lane_hash, mythr );
         }
      }
-      *noncev = _mm512_add_epi32( *noncev, m512_const1_32( 16 ) );
+      *noncev = _mm512_add_epi32( *noncev, _mm512_set1_epi32( 16 ) );
      n += 16;
   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
@@ -345,7 +345,7 @@ int scanhash_lyra2rev2_8way( struct work *work, const uint32_t max_nonce,
             submit_solution( work, lane_hash, mythr );
         }
      }
-      *noncev = _mm256_add_epi32( *noncev, m256_const1_32( 8 ) );
+      *noncev = _mm256_add_epi32( *noncev, _mm256_set1_epi32( 8 ) );
      n += 8;
   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
--- a/algo/lyra2/lyra2rev2.c
+++ b/algo/lyra2/lyra2rev2.c
@@ -4,7 +4,6 @@

 #include <memory.h>
 #include "algo/blake/sph_blake.h"
-#include "algo/cubehash/sph_cubehash.h"
 #include "algo/keccak/sph_keccak.h"
 #include "algo/skein/sph_skein.h"
 #include "algo/bmw/sph_bmw.h"
--- a/algo/lyra2/lyra2rev3-4way.c
+++ b/algo/lyra2/lyra2rev3-4way.c
@@ -287,7 +287,7 @@ int scanhash_lyra2rev3_8way( struct work *work, const uint32_t max_nonce,
             submit_solution( work, lane_hash, mythr );
         }
      }
-      *noncev = _mm256_add_epi32( *noncev, m256_const1_32( 8 ) );
+      *noncev = _mm256_add_epi32( *noncev, _mm256_set1_epi32( 8 ) );
      n += 8;
   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
   pdata[19] = n;
@@ -389,7 +389,7 @@ int scanhash_lyra2rev3_4way( struct work *work, const uint32_t max_nonce,
              submit_solution( work, lane_hash, mythr );
 	      }
      }
-      *noncev = _mm_add_epi32( *noncev, m128_const1_32( 4 ) );
+      *noncev = _mm_add_epi32( *noncev, _mm_set1_epi32( 4 ) );
      n += 4;
   } while ( (n < max_nonce-4) && !work_restart[thr_id].restart);
   pdata[19] = n;
--- a/algo/lyra2/lyra2rev3.c
+++ b/algo/lyra2/lyra2rev3.c
@@ -4,7 +4,6 @@

 #include <memory.h>
 #include "algo/blake/sph_blake.h"
-#include "algo/cubehash/sph_cubehash.h"
 #include "algo/bmw/sph_bmw.h"
 #include "algo/cubehash/cubehash_sse2.h" 
 //#include "lyra2.h"
--- a/algo/lyra2/lyra2z-4way.c
+++ b/algo/lyra2/lyra2z-4way.c
@@ -35,7 +35,7 @@ static void lyra2z_16way_hash( void *state, const void *midstate_vars,
    uint32_t hash14[8] __attribute__ ((aligned (32)));
    uint32_t hash15[8] __attribute__ ((aligned (32)));

-    blake256_16way_final_rounds_le( vhash, midstate_vars, midhash, block );
+    blake256_16way_final_rounds_le( vhash, midstate_vars, midhash, block, 14 );

    dintrlv_16x32( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
              hash8, hash9, hash10, hash11 ,hash12, hash13, hash14, hash15,
@@ -103,12 +103,12 @@ int scanhash_lyra2z_16way( struct work *work, uint32_t max_nonce,
   const uint32_t last_nonce = max_nonce - 16;
   const int thr_id = mythr->id;
   const bool bench = opt_benchmark;
-   const __m512i sixteen = m512_const1_32( 16 );
+   const __m512i sixteen = _mm512_set1_epi32( 16 );

   if ( bench ) ( (uint32_t*)ptarget )[7] = 0x0000ff;

   // Prehash first block
-   blake256_transform_le( phash, pdata, 512, 0 );
+   blake256_transform_le( phash, pdata, 512, 0, 14 );

   block0_hash[0] = _mm512_set1_epi32( phash[0] );
   block0_hash[1] = _mm512_set1_epi32( phash[1] );
@@ -120,25 +120,13 @@ int scanhash_lyra2z_16way( struct work *work, uint32_t max_nonce,
   block0_hash[7] = _mm512_set1_epi32( phash[7] );

   // Build vectored second block, interleave last 16 bytes of data using
-   // unique nonces and add padding.
+   // unique nonces.
   block_buf[ 0] = _mm512_set1_epi32( pdata[16] );
   block_buf[ 1] = _mm512_set1_epi32( pdata[17] );
   block_buf[ 2] = _mm512_set1_epi32( pdata[18] );
   block_buf[ 3] =
             _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 );
-   block_buf[ 4] = m512_const1_32( 0x80000000 );
-   block_buf[ 5] =
-   block_buf[ 6] =
-   block_buf[ 7] =
-   block_buf[ 8] =
-   block_buf[ 9] =
-   block_buf[10] =
-   block_buf[11] =
-   block_buf[12] = m512_zero;
-   block_buf[13] = m512_one_32;
-   block_buf[14] = m512_zero;
-   block_buf[15] = m512_const1_32( 80*8 );

   // Partialy prehash second block without touching nonces in block_buf[3].
   blake256_16way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
@@ -182,7 +170,7 @@ static void lyra2z_8way_hash( void *state, const void *midstate_vars,
     uint32_t hash7[8] __attribute__ ((aligned (32)));
     uint32_t vhash[8*8] __attribute__ ((aligned (64)));

-     blake256_8way_final_rounds_le( vhash, midstate_vars, midhash, block );
+     blake256_8way_final_rounds_le( vhash, midstate_vars, midhash, block, 14 );

     dintrlv_8x32( hash0, hash1, hash2, hash3,
                   hash4, hash5, hash6, hash7, vhash, 256 );
@@ -225,10 +213,10 @@ int scanhash_lyra2z_8way( struct work *work, uint32_t max_nonce,
   uint32_t n = first_nonce;
   const int thr_id = mythr->id;
   const bool bench = opt_benchmark;
-   const __m256i eight = m256_const1_32( 8 );
+   const __m256i eight = _mm256_set1_epi32( 8 );

   // Prehash first block
-   blake256_transform_le( phash, pdata, 512, 0 );
+   blake256_transform_le( phash, pdata, 512, 0, 14 );

   block0_hash[0] = _mm256_set1_epi32( phash[0] );
   block0_hash[1] = _mm256_set1_epi32( phash[1] );
@@ -240,24 +228,12 @@ int scanhash_lyra2z_8way( struct work *work, uint32_t max_nonce,
   block0_hash[7] = _mm256_set1_epi32( phash[7] );

   // Build vectored second block, interleave last 16 bytes of data using
-   // unique nonces and add padding.
+   // unique nonces.
   block_buf[ 0] = _mm256_set1_epi32( pdata[16] );
   block_buf[ 1] = _mm256_set1_epi32( pdata[17] );
   block_buf[ 2] = _mm256_set1_epi32( pdata[18] );
   block_buf[ 3] =
            _mm256_set_epi32( n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n +1, n );
-   block_buf[ 4] = m256_const1_32( 0x80000000 );
-   block_buf[ 5] =
-   block_buf[ 6] =
-   block_buf[ 7] =
-   block_buf[ 8] =
-   block_buf[ 9] =
-   block_buf[10] =
-   block_buf[11] =
-   block_buf[12] = m256_zero;
-   block_buf[13] = m256_one_32;
-   block_buf[14] = m256_zero;
-   block_buf[15] = m256_const1_32( 80*8 );

   // Partialy prehash second block without touching nonces
   blake256_8way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
@@ -352,7 +328,7 @@ int scanhash_lyra2z_4way( struct work *work, uint32_t max_nonce,
           submit_solution( work, lane_hash, mythr );
        }
      }
-      *noncev = _mm_add_epi32( *noncev, m128_const1_32( 4 ) );
+      *noncev = _mm_add_epi32( *noncev, _mm_set1_epi32( 4 ) );
      n += 4;
   } while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );

--- a/algo/lyra2/lyra2z330.c
+++ b/algo/lyra2/lyra2z330.c
@@ -3,7 +3,7 @@
 #include "lyra2.h"
 #include "simd-utils.h"

-__thread uint64_t* lyra2z330_wholeMatrix;
+static __thread uint64_t* lyra2z330_wholeMatrix;

 void lyra2z330_hash(void *state, const void *input, uint32_t height)
 {
--- a/algo/lyra2/sponge-2way.c
+++ b/algo/lyra2/sponge-2way.c
@@ -85,9 +85,9 @@ inline void absorbBlockBlake2Safe_2way( uint64_t *State, const uint64_t *In,

  state0 = 
  state1 = m512_zero;
-  state2 = m512_const4_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
+  state2 = _mm512_set4_epi64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
                              0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
-  state3 = m512_const4_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
+  state3 = _mm512_set4_epi64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
                              0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );

  for ( int i = 0; i < nBlocks; i++ )
--- a/algo/lyra2/sponge.c
+++ b/algo/lyra2/sponge.c
@@ -41,16 +41,16 @@
 inline void initState( uint64_t State[/*16*/] )
 {

-   /*
+/*
 #if defined (__AVX2__)

  __m256i* state = (__m256i*)State;
  const __m256i zero = m256_zero; 
  state[0] = zero;
  state[1] = zero;
-  state[2] = m256_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
+  state[2] = _mm256_set_epi64x( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
                                0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
-  state[3] = m256_const_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
+  state[3] = _mm256_set_epi64x( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
                                0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );

 #elif defined (__SSE2__)
@@ -62,10 +62,10 @@ inline void initState( uint64_t State[/*16*/] )
  state[1] = zero;
  state[2] = zero;
  state[3] = zero;
-  state[4] = m128_const_64( 0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
-  state[5] = m128_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL );
-  state[6] = m128_const_64( 0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );
-  state[7] = m128_const_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL );
+  state[4] = _mm_set_epi64x( 0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
+  state[5] = _mm_set_epi64x( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL );
+  state[6] = _mm_set_epi64x( 0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );
+  state[7] = _mm_set_epi64x( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL );

 #else
    //First 512 bis are zeros
@@ -271,9 +271,9 @@ inline void absorbBlockBlake2Safe( uint64_t *State, const uint64_t *In,

  state0 = 
  state1 = m256_zero;
-  state2 = m256_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
+  state2 = _mm256_set_epi64x( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
                              0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
-  state3 = m256_const_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
+  state3 = _mm256_set_epi64x( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
                              0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );

  for ( int i = 0; i < nBlocks; i++ )
@@ -299,10 +299,10 @@ inline void absorbBlockBlake2Safe( uint64_t *State, const uint64_t *In,
  state1 =
  state2 =
  state3 = m128_zero;
-  state4 = m128_const_64( 0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
-  state5 = m128_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL );
-  state6 = m128_const_64( 0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );
-  state7 = m128_const_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL );
+  state4 = _mm_set_epi64x( 0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
+  state5 = _mm_set_epi64x( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL );
+  state6 = _mm_set_epi64x( 0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );
+  state7 = _mm_set_epi64x( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL );

  for ( int i = 0; i < nBlocks; i++ )
  { 
--- a/algo/lyra2/sponge.h
+++ b/algo/lyra2/sponge.h
@@ -43,27 +43,29 @@ static const uint64_t blake2b_IV[8] =
  0x1f83d9abfb41bd6bULL, 0x5be0cd19137e2179ULL
 };

-/*Blake2b's rotation*/
-static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
-    return ( w >> c ) | ( w << ( 64 - c ) );
-}
-
-// serial data is only 32 bytes so AVX2 is the limit for that dimension.
-// However, 2 way parallel looks trivial to code for AVX512 except for
-// a data dependency with rowa.
-
 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

 #define G2W_4X64(a,b,c,d) \
   a = _mm512_add_epi64( a, b ); \
-   d = mm512_ror_64( _mm512_xor_si512( d, a ), 32 ); \
+   d = _mm512_ror_epi64( _mm512_xor_si512( d, a ), 32 ); \
   c = _mm512_add_epi64( c, d ); \
-   b = mm512_ror_64( _mm512_xor_si512( b, c ), 24 ); \
+   b = _mm512_ror_epi64( _mm512_xor_si512( b, c ), 24 ); \
   a = _mm512_add_epi64( a, b ); \
-   d = mm512_ror_64( _mm512_xor_si512( d, a ), 16 ); \
+   d = _mm512_ror_epi64( _mm512_xor_si512( d, a ), 16 ); \
   c = _mm512_add_epi64( c, d ); \
-   b = mm512_ror_64( _mm512_xor_si512( b, c ), 63 );
+   b = _mm512_ror_epi64( _mm512_xor_si512( b, c ), 63 );

+#define LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
+   G2W_4X64( s0, s1, s2, s3 ); \
+   s0 = mm512_shufll256_64( s0 ); \
+   s3 = mm512_swap256_128( s3); \
+   s2 = mm512_shuflr256_64( s2 ); \
+   G2W_4X64( s0, s1, s2, s3 ); \
+   s0 = mm512_shuflr256_64( s0 ); \
+   s3 = mm512_swap256_128( s3 ); \
+   s2 = mm512_shufll256_64( s2 ); 
+
+/*
 #define LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
   G2W_4X64( s0, s1, s2, s3 ); \
   s3 = mm512_shufll256_64( s3 ); \
@@ -73,6 +75,7 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
   s3 = mm512_shuflr256_64( s3 ); \
   s1 = mm512_shufll256_64( s1 ); \
   s2 = mm512_swap256_128( s2 ); 
+*/

 #define LYRA_12_ROUNDS_2WAY_AVX512( s0, s1, s2, s3 ) \
   LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
@@ -88,23 +91,32 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
   LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
   LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 )

-
 #endif  // AVX512

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

-// process 4 columns in parallel
-// returns void, updates all args
 #define G_4X64(a,b,c,d) \
   a = _mm256_add_epi64( a, b ); \
-   d = mm256_ror_64( _mm256_xor_si256( d, a ), 32 ); \
+   d = mm256_swap64_32( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi64( c, d ); \
-   b = mm256_ror_64( _mm256_xor_si256( b, c ), 24 ); \
+   b = mm256_shuflr64_24( _mm256_xor_si256( b, c ) ); \
   a = _mm256_add_epi64( a, b ); \
-   d = mm256_ror_64( _mm256_xor_si256( d, a ), 16 ); \
+   d = mm256_shuflr64_16( _mm256_xor_si256( d, a ) ); \
   c = _mm256_add_epi64( c, d ); \
   b = mm256_ror_64( _mm256_xor_si256( b, c ), 63 );

+// Pivot about s1 instead of s0 reduces latency.
+#define LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
+   G_4X64( s0, s1, s2, s3 ); \
+   s0 = mm256_shufll_64( s0 ); \
+   s3 = mm256_swap_128( s3); \
+   s2 = mm256_shuflr_64( s2 ); \
+   G_4X64( s0, s1, s2, s3 ); \
+   s0 = mm256_shuflr_64( s0 ); \
+   s3 = mm256_swap_128( s3 ); \
+   s2 = mm256_shufll_64( s2 );
+
+/*
 #define LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
   G_4X64( s0, s1, s2, s3 ); \
   s3 = mm256_shufll_64( s3 ); \
@@ -114,6 +126,7 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
   s3 = mm256_shuflr_64( s3 ); \
   s1 = mm256_shufll_64( s1 ); \
   s2 = mm256_swap_128( s2 );
+*/

 #define LYRA_12_ROUNDS_AVX2( s0, s1, s2, s3 ) \
   LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
@@ -137,23 +150,34 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
 // returns void, all args updated
 #define G_2X64(a,b,c,d) \
   a = _mm_add_epi64( a, b ); \
-   d = mm128_ror_64( _mm_xor_si128( d, a), 32 ); \
+   d = mm128_swap64_32( _mm_xor_si128( d, a) ); \
   c = _mm_add_epi64( c, d ); \
-   b = mm128_ror_64( _mm_xor_si128( b, c ), 24 ); \
+   b = mm128_shuflr64_24( _mm_xor_si128( b, c ) ); \
   a = _mm_add_epi64( a, b ); \
-   d = mm128_ror_64( _mm_xor_si128( d, a ), 16 ); \
+   d = mm128_shuflr64_16( _mm_xor_si128( d, a ) ); \
   c = _mm_add_epi64( c, d ); \
   b = mm128_ror_64( _mm_xor_si128( b, c ), 63 );

 #define LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
+{ \
+   __m128i t; \
   G_2X64( s0, s2, s4, s6 ); \
   G_2X64( s1, s3, s5, s7 ); \
-   mm128_vrol256_64( s6, s7 ); \
-   mm128_vror256_64( s2, s3 ); \
+   t = mm128_alignr_64( s7, s6, 1 ); \
+   s6 = mm128_alignr_64( s6, s7, 1 ); \
+   s7 = t; \
+   t = mm128_alignr_64( s2, s3, 1 ); \
+   s2 =  mm128_alignr_64( s3, s2, 1 ); \
+   s3 = t; \
   G_2X64( s0, s2, s5, s6 ); \
   G_2X64( s1, s3, s4, s7 ); \
-   mm128_vror256_64( s6, s7 ); \
-   mm128_vrol256_64( s2, s3 );
+   t = mm128_alignr_64( s6, s7, 1 ); \
+   s6 = mm128_alignr_64( s7, s6, 1 ); \
+   s7 = t; \
+   t = mm128_alignr_64( s3, s2, 1 ); \
+   s2 =  mm128_alignr_64( s2, s3, 1 ); \
+   s3 = t; \
+} 

 #define LYRA_12_ROUNDS_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
   LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
@@ -171,8 +195,13 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){

 #endif // AVX2 else SSE2

-// Scalar
-//Blake2b's G function
+/*
+// Scalar, not used.
+
+static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
+    return ( w >> c ) | ( w << ( 64 - c ) );
+}
+
 #define G(r,i,a,b,c,d) \
  do { \
    a = a + b; \
@@ -185,8 +214,6 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
    b = rotr64(b ^ c, 63); \
  } while(0)

-
-/*One Round of the Blake2b's compression function*/
 #define ROUND_LYRA(r)  \
    G(r,0,v[ 0],v[ 4],v[ 8],v[12]); \
    G(r,1,v[ 1],v[ 5],v[ 9],v[13]); \
@@ -196,6 +223,7 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
    G(r,5,v[ 1],v[ 6],v[11],v[12]); \
    G(r,6,v[ 2],v[ 7],v[ 8],v[13]); \
    G(r,7,v[ 3],v[ 4],v[ 9],v[14]);
+*/

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

--- a/algo/panama/sph_panama.h
+++ b/algo/panama/sph_panama.h
@@ -58,7 +58,7 @@
 #define SPH_PANAMA_H__

 #include <stddef.h>
-#include "algo/sha/sph_types.h"
+#include "compat/sph_types.h"

 /**
 * Output size (in bits) for PANAMA.
--- a/algo/quark/anime-4way.c
+++ b/algo/quark/anime-4way.c
@@ -15,7 +15,8 @@

 #if defined (ANIME_8WAY)

-typedef struct {
+union _anime_8way_context_overlay
+{
    blake512_8way_context   blake;
    bmw512_8way_context     bmw;
 #if defined(__VAES__)
@@ -26,23 +27,9 @@ typedef struct {
    jh512_8way_context      jh;
    skein512_8way_context   skein;
    keccak512_8way_context  keccak;
-} anime_8way_ctx_holder;
+} __attribute__ ((aligned (64)));

-anime_8way_ctx_holder anime_8way_ctx __attribute__ ((aligned (64)));
-
-void init_anime_8way_ctx()
-{
-     blake512_8way_init( &anime_8way_ctx.blake );
-     bmw512_8way_init( &anime_8way_ctx.bmw );
-#if defined(__VAES__)
-     groestl512_4way_init( &anime_8way_ctx.groestl, 64 );
-#else
-     init_groestl( &anime_8way_ctx.groestl, 64 );
-#endif
-     skein512_8way_init( &anime_8way_ctx.skein );
-     jh512_8way_init( &anime_8way_ctx.jh );
-     keccak512_8way_init( &anime_8way_ctx.keccak );
-}
+typedef union _anime_8way_context_overlay anime_8way_context_overlay;

 void anime_8way_hash( void *state, const void *input )
 {
@@ -64,18 +51,15 @@ void anime_8way_hash( void *state, const void *input )
    __m512i* vhA = (__m512i*)vhashA;
    __m512i* vhB = (__m512i*)vhashB;
    __m512i* vhC = (__m512i*)vhashC;
-    const __m512i bit3_mask = m512_const1_64( 8 );
-    const __m512i zero = _mm512_setzero_si512();
+    const __m512i bit3_mask = _mm512_set1_epi64( 8 );
    __mmask8 vh_mask;
-    anime_8way_ctx_holder ctx;
-    memcpy( &ctx, &anime_8way_ctx, sizeof(anime_8way_ctx) );
+    anime_8way_context_overlay ctx __attribute__ ((aligned (64)));

    bmw512_8way_full( &ctx.bmw, vhash, input, 80 );

    blake512_8way_full( &ctx.blake, vhash, vhash, 64 );

-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ),
-                                       zero );
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );

 #if defined(__VAES__)

@@ -152,8 +136,7 @@ void anime_8way_hash( void *state, const void *input )
    jh512_8way_update( &ctx.jh, vhash, 64 );
    jh512_8way_close( &ctx.jh, vhash );

-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ),
-                                       zero );
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );

    if ( ( vh_mask & 0xff ) != 0xff )
       blake512_8way_full( &ctx.blake, vhashA, vhash, 64 );
@@ -168,8 +151,7 @@ void anime_8way_hash( void *state, const void *input )

    skein512_8way_full( &ctx.skein, vhash, vhash, 64 );

-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ), 
-                                       zero );
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );

    if ( ( vh_mask & 0xff ) != 0xff )
    {
@@ -227,7 +209,7 @@ int scanhash_anime_8way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm512_add_epi32( *noncev,
-                                   m512_const1_64( 0x0000000800000000 ) );
+                                   _mm512_set1_epi64( 0x0000000800000000 ) );
       n += 8;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
    pdata[19] = n;
@@ -237,14 +219,20 @@ int scanhash_anime_8way( struct work *work, uint32_t max_nonce,

 #elif defined (ANIME_4WAY)

-typedef struct {
+union _anime_4way_context_overlay
+{
    blake512_4way_context  blake;
    bmw512_4way_context    bmw;
    hashState_groestl      groestl;
    jh512_4way_context     jh;
    skein512_4way_context  skein;
    keccak512_4way_context keccak;
-} anime_4way_ctx_holder;
+#if defined(__VAES__)
+    groestl512_2way_context groestl2;
+#endif
+} __attribute__ ((aligned (64)));
+
+typedef union _anime_4way_context_overlay anime_4way_context_overlay;

 void anime_4way_hash( void *state, const void *input )
 {
@@ -260,9 +248,9 @@ void anime_4way_hash( void *state, const void *input )
    __m256i* vhB = (__m256i*)vhashB;
    __m256i vh_mask;
    int h_mask;
-    const __m256i bit3_mask = m256_const1_64( 8 );
+    const __m256i bit3_mask = _mm256_set1_epi64x( 8 );
    const __m256i zero = _mm256_setzero_si256();
-    anime_4way_ctx_holder ctx;
+    anime_4way_context_overlay ctx __attribute__ ((aligned (64)));

    bmw512_4way_init( &ctx.bmw );
    bmw512_4way_update( &ctx.bmw, input, 80 );
@@ -293,6 +281,17 @@ void anime_4way_hash( void *state, const void *input )

    mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask );

+#if defined(__VAES__)
+
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+   groestl512_2way_full( &ctx.groestl2, vhashA, vhashA, 64 );
+   groestl512_2way_full( &ctx.groestl2, vhashB, vhashB, 64 );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+
   dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );

    groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -302,6 +301,8 @@ void anime_4way_hash( void *state, const void *input )

    intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );

+#endif
+
    jh512_4way_init( &ctx.jh );
    jh512_4way_update( &ctx.jh, vhash, 64 );
    jh512_4way_close( &ctx.jh, vhash );
@@ -387,7 +388,7 @@ int scanhash_anime_4way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm256_add_epi32( *noncev,
-                                   m256_const1_64( 0x0000000400000000 ) );
+                                   _mm256_set1_epi64x( 0x0000000400000000 ) );
       n += 4;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
    pdata[19] = n;
--- a/algo/quark/hmq1725-4way.c
+++ b/algo/quark/hmq1725-4way.c
@@ -13,6 +13,7 @@
 #include "algo/cubehash/cubehash_sse2.h"
 #include "algo/simd/nist.h"
 #include "algo/shavite/sph_shavite.h"
+#include "algo/shavite/shavite-hash-2way.h"
 #include "algo/simd/simd-hash-2way.h"
 #include "algo/echo/aes_ni/hash_api.h"
 #include "algo/hamsi/hamsi-hash-4way.h"
@@ -20,7 +21,7 @@
 #include "algo/shabal/shabal-hash-4way.h"
 #include "algo/whirlpool/sph_whirlpool.h"
 #include "algo/haval/haval-hash-4way.h"
-#include "algo/sha/sha-hash-4way.h"
+#include "algo/sha/sha512-hash.h"
 #if defined(__VAES__)
  #include "algo/groestl/groestl512-hash-4way.h"
  #include "algo/shavite/shavite-hash-4way.h"
@@ -74,7 +75,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   uint32_t hash7 [16]    __attribute__ ((aligned (32)));
   hmq1725_8way_context_overlay ctx __attribute__ ((aligned (64)));
   __mmask8 vh_mask;
-   const __m512i vmask = m512_const1_64( 24 );
+   const __m512i vmask = _mm512_set1_epi64( 24 );
   const uint32_t mask = 24;
   __m512i* vh  = (__m512i*)vhash;
   __m512i* vhA = (__m512i*)vhashA;
@@ -98,8 +99,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3,
                           hash4, hash5, hash6,  hash7 );

-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

   // A
 #if defined(__VAES__)
@@ -154,8 +154,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   keccak512_8way_update( &ctx.keccak, vhash, 64 );
   keccak512_8way_close( &ctx.keccak, vhash );

-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

   // A
   if ( ( vh_mask & 0xff ) != 0xff )
@@ -174,8 +173,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   cube_4way_full( &ctx.cube, vhashB, 512, vhashB, 64 );

   rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );
-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

   if ( likely( ( vh_mask & 0xff ) != 0xff ) )
   {
@@ -223,8 +221,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   simd512_4way_full( &ctx.simd, vhashB, vhashB, 64 );

   rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );
-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );
   dintrlv_8x64_512( hash0, hash1, hash2, hash3,
                     hash4, hash5, hash6, hash7, vhash );
   // 4x32 for haval
@@ -302,8 +299,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)

   blake512_8way_full( &ctx.blake, vhash, vhash, 64 );

-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

   // A
 #if defined(__VAES__)
@@ -374,8 +370,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)

   intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3,
                           hash4, hash5, hash6, hash7 );
-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

     // A   
 #if defined(__VAES__)
@@ -455,8 +450,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)

   intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3,
                           hash4, hash5, hash6, hash7 );
-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );

   if ( hash0[0] & mask )
      fugue512_full( &ctx.fugue, hash0, hash0, 64 );
@@ -520,8 +514,7 @@ extern void hmq1725_8way_hash(void *state, const void *input)
   sha512_8way_update( &ctx.sha512, vhash, 64 );
   sha512_8way_close( &ctx.sha512, vhash );

-   vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], vmask ),
-                                       m512_zero );
+   vh_mask = _mm512_testn_epi64_mask( vh[0], vmask );
   dintrlv_8x64_512( hash0, hash1, hash2, hash3,
                     hash4, hash5, hash6, hash7, vhash );

@@ -600,7 +593,7 @@ int scanhash_hmq1725_8way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm512_add_epi32( *noncev,
-                                   m512_const1_64( 0x0000000800000000 ) );
+                                   _mm512_set1_epi64( 0x0000000800000000 ) );
       n += 8;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );

@@ -625,6 +618,7 @@ union _hmq1725_4way_context_overlay
    cube_2way_context       cube2;
    sph_shavite512_context  shavite;
    hashState_sd            sd;
+    shavite512_2way_context shavite2;
    simd_2way_context       simd;
    hashState_echo          echo;
    hamsi512_4way_context   hamsi;
@@ -633,6 +627,10 @@ union _hmq1725_4way_context_overlay
    sph_whirlpool_context   whirlpool;
    sha512_4way_context     sha512;
    haval256_5_4way_context haval;
+#if defined(__VAES__)
+    groestl512_2way_context groestl2;
+    echo_2way_context       echo2;
+#endif    
 } __attribute__ ((aligned (64)));

 typedef union _hmq1725_4way_context_overlay hmq1725_4way_context_overlay;
@@ -649,7 +647,7 @@ extern void hmq1725_4way_hash(void *state, const void *input)
   hmq1725_4way_context_overlay ctx __attribute__ ((aligned (64)));
   __m256i vh_mask;     
   int h_mask;
-   const __m256i vmask = m256_const1_64( 24 );
+   const __m256i vmask = _mm256_set1_epi64x( 24 );
   const uint32_t mask = 24;
   __m256i* vh  = (__m256i*)vhash;
   __m256i* vhA = (__m256i*)vhashA;
@@ -750,15 +748,10 @@ extern void hmq1725_4way_hash(void *state, const void *input)

    mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask );

-    dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+    rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );

-    shavite512_full( &ctx.shavite, hash0, hash0, 64 );
-    shavite512_full( &ctx.shavite, hash1, hash1, 64 );
-    shavite512_full( &ctx.shavite, hash2, hash2, 64 );
-    shavite512_full( &ctx.shavite, hash3, hash3, 64 );
-
-    intrlv_2x128_512( vhashA, hash0, hash1 );
-    intrlv_2x128_512( vhashB, hash2, hash3 );
+    shavite512_2way_full( &ctx.shavite2, vhashA, vhashA, 64 );
+    shavite512_2way_full( &ctx.shavite2, vhashB, vhashB, 64 );

    simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
    simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
@@ -795,6 +788,17 @@ extern void hmq1725_4way_hash(void *state, const void *input)

    mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask );

+#if defined(__VAES__)
+
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+   echo_2way_full( &ctx.echo2, vhashA, 512, vhashA, 64 );
+   echo_2way_full( &ctx.echo2, vhashB, 512, vhashB, 64 );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+    
    dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
    
    echo_full( &ctx.echo, (BitSequence *)hash0, 512,
@@ -808,6 +812,8 @@ extern void hmq1725_4way_hash(void *state, const void *input)

    intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );

+#endif
+
    blake512_4way_full( &ctx.blake, vhash, vhash, 64 );

    dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
@@ -939,6 +945,17 @@ extern void hmq1725_4way_hash(void *state, const void *input)

   mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask );

+#if defined(__VAES__)
+
+   rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
+
+   groestl512_2way_full( &ctx.groestl2, vhashA, vhashA, 64 );
+   groestl512_2way_full( &ctx.groestl2, vhashB, vhashB, 64 );
+
+   rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
+
+#else
+   
   dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );

   groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -948,6 +965,8 @@ extern void hmq1725_4way_hash(void *state, const void *input)

   intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );

+#endif
+
   sha512_4way_init( &ctx.sha512 ); 
   sha512_4way_update( &ctx.sha512, vhash, 64 );
   sha512_4way_close( &ctx.sha512, vhash ); 
@@ -1022,7 +1041,7 @@ int scanhash_hmq1725_4way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm256_add_epi32( *noncev,
-                                   m256_const1_64( 0x0000000400000000 ) );
+                                   _mm256_set1_epi64x( 0x0000000400000000 ) );
       n += 4;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
    pdata[19] = n;
--- a/algo/quark/quark-4way.c
+++ b/algo/quark/quark-4way.c
@@ -67,8 +67,7 @@ void quark_8way_hash( void *state, const void *input )
    __mmask8 vh_mask;
    quark_8way_ctx_holder ctx;
    const uint32_t mask = 8;
-    const __m512i bit3_mask = m512_const1_64( mask );
-    const __m512i zero = _mm512_setzero_si512();
+    const __m512i bit3_mask = _mm512_set1_epi64( mask );

    memcpy( &ctx, &quark_8way_ctx, sizeof(quark_8way_ctx) );

@@ -76,9 +75,7 @@ void quark_8way_hash( void *state, const void *input )

    bmw512_8way_full( &ctx.bmw, vhash, vhash, 64 );
    
-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ),
-                                       zero );
-
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );
    
 #if defined(__VAES__)

@@ -154,8 +151,7 @@ void quark_8way_hash( void *state, const void *input )
    jh512_8way_update( &ctx.jh, vhash, 64 );
    jh512_8way_close( &ctx.jh, vhash );

-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ),
-                                       zero );
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );

    if ( ( vh_mask & 0xff ) != 0xff )
       blake512_8way_full( &ctx.blake, vhashA, vhash, 64 );
@@ -169,8 +165,7 @@ void quark_8way_hash( void *state, const void *input )

    skein512_8way_full( &ctx.skein, vhash, vhash, 64 );

-    vh_mask = _mm512_cmpeq_epi64_mask( _mm512_and_si512( vh[0], bit3_mask ),
-                                       zero );
+    vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask );

    if ( ( vh_mask & 0xff ) != 0xff )
    {
@@ -229,7 +224,7 @@ int scanhash_quark_8way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm512_add_epi32( *noncev,
-                                  m512_const1_64( 0x0000000800000000 ) );
+                                  _mm512_set1_epi64( 0x0000000800000000 ) );
       n += 8;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );

@@ -276,7 +271,7 @@ void quark_4way_hash( void *state, const void *input )
    __m256i vh_mask;
    int h_mask;
    quark_4way_ctx_holder ctx;
-    const __m256i bit3_mask = m256_const1_64( 8 );
+    const __m256i bit3_mask = _mm256_set1_epi64x( 8 );
    const __m256i zero = _mm256_setzero_si256();

    memcpy( &ctx, &quark_4way_ctx, sizeof(quark_4way_ctx) );
@@ -402,7 +397,7 @@ int scanhash_quark_4way( struct work *work, uint32_t max_nonce,
          }
       }
       *noncev = _mm256_add_epi32( *noncev,
-                                  m256_const1_64( 0x0000000400000000 ) );
+                                  _mm256_set1_epi64x( 0x0000000400000000 ) );
       n += 4;
    } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );

--- a/algo/radiogatun/sph_radiogatun.c
+++ b/algo/radiogatun/sph_radiogatun.c
--- a/algo/radiogatun/sph_radiogatun.h
+++ b/algo/radiogatun/sph_radiogatun.h
@@ -1,186 +0,0 @@
-/* $Id: sph_radiogatun.h 226 2010-06-16 17:28:08Z tp $ */
-/**
- * RadioGatun interface.
- *
- * RadioGatun has been published in: G. Bertoni, J. Daemen, M. Peeters
- * and G. Van Assche, "RadioGatun, a belt-and-mill hash function",
- * presented at the Second Cryptographic Hash Workshop, Santa Barbara,
- * August 24-25, 2006. The main Web site, containing that article, the
- * reference code and some test vectors, appears to be currently located
- * at the following URL: http://radiogatun.noekeon.org/
- *
- * The presentation article does not specify endianness or padding. The
- * reference code uses the following conventions, which we also apply
- * here:
- * <ul>
- * <li>The input message is an integral number of sequences of three
- * words. Each word is either a 32-bit of 64-bit word (depending on
- * the version of RadioGatun).</li>
- * <li>Input bytes are decoded into words using little-endian
- * convention.</li>
- * <li>Padding consists of a single bit of value 1, using little-endian
- * convention within bytes (i.e. for a byte-oriented input, a single
- * byte of value 0x01 is appended), then enough bits of value 0 to finish
- * the current block.</li>
- * <li>Output consists of 256 bits. Successive output words are encoded
- * with little-endian convention.</li>
- * </ul>
- * These conventions are very close to those we use for PANAMA, which is
- * a close ancestor or RadioGatun.
- *
- * RadioGatun is actually a family of functions, depending on some
- * internal parameters. We implement here two functions, with a "belt
- * length" of 13, a "belt width" of 3, and a "mill length" of 19. The
- * RadioGatun[32] version uses 32-bit words, while the RadioGatun[64]
- * variant uses 64-bit words.
- *
- * Strictly speaking, the name "RadioGatun" should use an acute accent
- * on the "u", which we omitted here to keep strict ASCII-compatibility
- * of this file.
- *
- * ==========================(LICENSE BEGIN)============================
- *
- * Copyright (c) 2007-2010  Projet RNRT SAPHIR
- *
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sublicense, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice shall be
- * included in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
- * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
- * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
- * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
- * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
- * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * ===========================(LICENSE END)=============================
- *
- * @file     sph_radiogatun.h
- * @author   Thomas Pornin <thomas.pornin@cryptolog.com>
- */
-
-#ifndef SPH_RADIOGATUN_H__
-#define SPH_RADIOGATUN_H__
-
-#include <stddef.h>
-#include "algo/sha/sph_types.h"
-
-/**
- * Output size (in bits) for RadioGatun[32].
- */
-#define SPH_SIZE_radiogatun32   256
-
-/**
- * This structure is a context for RadioGatun[32] computations: it
- * contains intermediate values and some data from the last entered
- * block. Once a RadioGatun[32] computation has been performed, the
- * context can be reused for another computation.
- *
- * The contents of this structure are private. A running RadioGatun[32]
- * computation can be cloned by copying the context (e.g. with a
- * simple <code>memcpy()</code>).
- */
-typedef struct {
-#ifndef DOXYGEN_IGNORE
-	unsigned char data[156];   /* first field, for alignment */
-	unsigned data_ptr;
-	sph_u32 a[19], b[39];
-#endif
-} sph_radiogatun32_context;
-
-/**
- * Initialize a RadioGatun[32] context. This process performs no
- * memory allocation.
- *
- * @param cc   the RadioGatun[32] context (pointer to a
- *             <code>sph_radiogatun32_context</code>)
- */
-void sph_radiogatun32_init(void *cc);
-
-/**
- * Process some data bytes. It is acceptable that <code>len</code> is zero
- * (in which case this function does nothing).
- *
- * @param cc     the RadioGatun[32] context
- * @param data   the input data
- * @param len    the input data length (in bytes)
- */
-void sph_radiogatun32(void *cc, const void *data, size_t len);
-
-/**
- * Terminate the current RadioGatun[32] computation and output the
- * result into the provided buffer. The destination buffer must be wide
- * enough to accomodate the result (32 bytes). The context is
- * automatically reinitialized.
- *
- * @param cc    the RadioGatun[32] context
- * @param dst   the destination buffer
- */
-void sph_radiogatun32_close(void *cc, void *dst);
-
-#if SPH_64
-
-/**
- * Output size (in bits) for RadioGatun[64].
- */
-#define SPH_SIZE_radiogatun64   256
-
-/**
- * This structure is a context for RadioGatun[64] computations: it
- * contains intermediate values and some data from the last entered
- * block. Once a RadioGatun[64] computation has been performed, the
- * context can be reused for another computation.
- *
- * The contents of this structure are private. A running RadioGatun[64]
- * computation can be cloned by copying the context (e.g. with a
- * simple <code>memcpy()</code>).
- */
-typedef struct {
-#ifndef DOXYGEN_IGNORE
-	unsigned char data[312];   /* first field, for alignment */
-	unsigned data_ptr;
-	sph_u64 a[19], b[39];
-#endif
-} sph_radiogatun64_context;
-
-/**
- * Initialize a RadioGatun[64] context. This process performs no
- * memory allocation.
- *
- * @param cc   the RadioGatun[64] context (pointer to a
- *             <code>sph_radiogatun64_context</code>)
- */
-void sph_radiogatun64_init(void *cc);
-
-/**
- * Process some data bytes. It is acceptable that <code>len</code> is zero
- * (in which case this function does nothing).
- *
- * @param cc     the RadioGatun[64] context
- * @param data   the input data
- * @param len    the input data length (in bytes)
- */
-void sph_radiogatun64(void *cc, const void *data, size_t len);
-
-/**
- * Terminate the current RadioGatun[64] computation and output the
- * result into the provided buffer. The destination buffer must be wide
- * enough to accomodate the result (32 bytes). The context is
- * automatically reinitialized.
- *
- * @param cc    the RadioGatun[64] context
- * @param dst   the destination buffer
- */
-void sph_radiogatun64_close(void *cc, void *dst);
-
-#endif
-
-#endif
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Jay D Dee	be88afc349	v3.23.2	2023-09-21 12:34:06 -04:00
Jay D Dee	d6b5750362	v3.23.1	2023-09-13 11:48:52 -04:00
Jay D Dee	4378d2f841	v3.23.0	2023-08-30 20:15:48 -04:00
Jay D Dee	57a6b7b58b	v3.22.3	2023-06-14 11:07:40 -04:00
Jay D Dee	de564ccbde	v3.22.2	2023-04-06 13:38:37 -04:00
Jay D Dee	fcd7727b0d	v3.22.1	2023-03-24 18:29:42 -04:00
Jay D Dee	3dd6787531	v3.22.0	2023-03-21 17:12:51 -04:00
Jay D Dee	cae1ce2ab7	v3.21.5	2023-03-15 12:27:04 -04:00
Jay D Dee	7a91c41d74	v3.21.4	2023-03-13 14:54:38 -04:00
Jay D Dee	c6bc9d67fb	v3.21.3 Unreleased	2023-03-13 03:20:13 -04:00
Jay D Dee	b339450898	v3.21.3	2023-03-11 14:54:49 -05:00
Jay D Dee	fb93160641	v3.21.2	2023-03-03 12:38:31 -05:00
Jay D Dee	520d4d5384	v3.21.1	2023-02-08 22:11:05 -05:00
Jay D Dee	da7030faa8	v3.21.0	2022-12-21 13:09:14 -05:00
Jay D Dee	bd84f199fe	v3.20.3	2022-10-21 23:12:18 -04:00
Jay D Dee	58030e2788	v3.20.2	2022-08-01 20:21:05 -04:00