v3.9.10

v3.9.9.1
v3.9.9
2025-09-17 23:44:27 +00:00 · 2019-11-22 20:29:18 -05:00 · 2019-10-24 14:11:26 -04:00 · 2019-10-10 19:58:34 -04:00 · 2019-10-01 14:17:36 -04:00 · 2019-09-26 22:37:26 -04:00
111 changed files with 7907 additions and 4215 deletions
--- a/Makefile.am
+++ b/Makefile.am
@@ -18,7 +18,6 @@ dist_man_MANS	= cpuminer.1
 cpuminer_SOURCES = \
  cpu-miner.c \
  util.c \
-  uint256.cpp \
  api.c \
  sysinfos.c \
  algo-gate-api.c\
@@ -263,6 +262,8 @@ cpuminer_SOURCES = \
  algo/x16/x16r-gate.c \
  algo/x16/x16r.c \
  algo/x16/x16r-4way.c \
+  algo/x16/x16rv2.c \
+  algo/x16/x16rv2-4way.c \
  algo/x16/x16rt.c \
  algo/x16/x16rt-4way.c \
  algo/x16/hex.c \
@@ -281,7 +282,9 @@ cpuminer_SOURCES = \
  algo/yescrypt/yescrypt.c \
  algo/yescrypt/sha256_Y.c \
  algo/yescrypt/yescrypt-best.c \
-  algo/yespower/yespower.c \
+  algo/yespower/yespower-gate.c \
+  algo/yespower/yespower-blake2b.c \
+  algo/yespower/crypto/blake2b-yp.c \
  algo/yespower/sha256_p.c \
  algo/yespower/yespower-opt.c

--- a/README.md
+++ b/README.md
@@ -24,7 +24,7 @@ Requirements

 1. A x86_64 architecture CPU with a minimum of SSE2 support. This includes
 Intel Core2 and newer and AMD equivalents. In order to take advantage of AES_NI
-optimizations a CPU with AES_NI is required. This includes Intel Westbridge
+optimizations a CPU with AES_NI is required. This includes Intel Westmere
 and newer and AMD equivalents. Further optimizations are available on some
 algoritms for CPUs with AVX and AVX2, Sandybridge and Haswell respectively.

@@ -87,10 +87,12 @@ Supported Algorithms
                          neoscrypt     NeoScrypt(128, 2, 1)
                          nist5         Nist5
                          pentablake    Pentablake
-                          phi1612       phi, LUX coin (original algo)
-                          phi2          LUX coin (new algo)
+                          phi1612       phi
+                          phi2          Luxcoin (LUX)
+                          phi2-lux      identical to phi2
                          pluck         Pluck:128 (Supcoin)
                          polytimos     Ninja
+                          power2b       MicroBitcoin (MBC)
                          quark         Quark
                          qubit         Qubit
                          scrypt        scrypt(1024, 1, 1) (default)
@@ -120,9 +122,10 @@ Supported Algorithms
                          x13sm3        hsr (Hshare)
                          x14           X14
                          x15           X15
-                          x16r          Ravencoin (RVN)
+                          x16r          
+                          x16rv2        Ravencoin (RVN)
                          x16rt         Gincoin (GIN)
-                          x16rt_veil    Veil (VEIL)
+                          x16rt-veil    Veil (VEIL)
                          x16s          Pigeoncoin (PGN)
                          x17
                          x21s
@@ -133,6 +136,7 @@ Supported Algorithms
                          yescryptr32   WAVI
                          yespower      Cryply
                          yespowerr16   Yenten (YTN)
+                          yespower-b2b  generic yespower + blake2b
                          zr5           Ziftr

 Errata
@@ -155,14 +159,17 @@ Benchmark testing does not work for x11evo.
 Bugs
 ----

-Users are encouraged to post their bug reports on the Bitcoin Talk
-forum at:
+Users are encouraged to post their bug reports using git issues or on the
+Bitcoin Talk forum or opening an issue in git:

 https://bitcointalk.org/index.php?topic=1326803.0

-All problem reports must be accompanied by a proper definition.
+https://github.com/JayDDee/cpuminer-opt/issues
+
+All problem reports must be accompanied by a proper problem definition.
 This should include how the problem occurred, the command line and
-output from the miner showing the startup and any errors.
+output from the miner showing the startup messages and any errors.
+A history is also useful, ie did it work before.

 Donations
 ---------
@@ -170,10 +177,6 @@ Donations
 cpuminer-opt has no fees of any kind but donations are accepted.

 BTC: 12tdvfF7KmAsihBXQXynT6E6th2c2pByTT
- ETH: 0x72122edabcae9d3f57eab0729305a425f6fef6d0
- LTC: LdUwoHJnux9r9EKqFWNvAi45kQompHk6e8
- BCH: 1QKYkB6atn4P7RFozyziAXLEnurwnUM1cQ
- BTG: GVUyECtRHeC5D58z9F3nGGfVQndwnsPnHQ

 Happy mining!

--- a/76
+++ b/76
@@ -1,14 +1,6 @@
 cpuminer-opt is a console program run from the command line using the
 keyboard, not the mouse.

-cpuminer-opt now supports HW SHA acceleration available on AMD Ryzen CPUs.
-This feature requires recent SW including GCC version 5 or higher and
-openssl version 1.1 or higher. It may also require using "-march=znver1"
-compile flag.
-
-cpuminer-opt is a console program, if you're using a mouse you're doing it
-wrong.
-
 Security warning
 ----------------

@@ -34,10 +26,78 @@ Intel Core2 or newer, or AMD Steamroller or newer CPU. ARM CPUs are not
 supported.

 64 bit Linux or Windows operating system. Apple and Android are not supported.
+FreeBSD YMMV.

 Change Log
 ----------

+v3.9.10
+
+Faster X* algos with AVX2.
+Small improvements to summary stats report.
+
+v3.9.9.1
+
+Fixed a day1 bug that could cause the miner to idle for up to 2 minutes
+under certain circumstances.
+
+Redesigned summary stats report now includes session statistics.
+
+More robust handling of statistics to reduce corruption.
+
+Removed --hide-diff option.
+
+Better handling of cpu-affinity with more than 64 CPUs.
+
+v3.9.9
+
+Added power2b algo for MicroBitcoin.
+Added generic yespower-b2b (yespower + blake2b) algo to be used with
+the parameters introduced in v3.9.7 for yespower & yescrypt.
+Display additional info when a share is rejected.
+Some low level enhancements and minor tweaking of log output.
+RELEASE_NOTES (this file) and README.md  added to Windows release package.
+
+v3.9.8.1
+
+Summary log report will be generated on stratum diff change or after 5 minutes,
+whichever comes first, to prevent incorrect data in the report.
+
+Removed phi2-lux alias (introduced in v3.9.8) due to Luxcoin's planned fork
+to a new algo. The new Luxcoin algo is not supported by cpuminer-opt.
+Until the fork Luxcoin can be mined using phi2 algo.
+
+--hide-diff option is deprecated and has no effect. It will be removed in a
+future release.
+
+v3.9.8
+
+Changes to log output to provide data more relevant to actual mining
+performance.
+phi2 can now handle pools with a mix of coins that use and don't use roots.
+phi2-lux added as an alias for phi2 as they are identical except for roots.
+Add x16rv2 algo for Ravencoin fork.
+
+v3.9.7
+
+Command line option changes:
+
+"-R" is no longer used as a shortcut for "--retry-pause", users must
+use the long option.
+
+New options:
+
+-N, --param-n: set the N parameter for yescrypt, yespower or scrypt algos
+-R, --param-r: set the R parameter for yescrypt or yespower algos, scrypt is
+     hardcoded with R=1
+-K, --param-key: set the client key/pers parameter for yescrypt/yespower algos.
+
+These options can be used to mine yescrypt or yespower variations using
+the generic yescrypt or yespower algo name and specifying the parameters
+manually. They can even be used to mine variations that aren't formally
+supported by a unique algo name. Existing algos can continue to to be mined
+using their original name without parameters.
+
 v3.9.6.2

 New algo blake2b.
--- a/algo-gate-api.c
+++ b/algo-gate-api.c
@@ -116,13 +116,10 @@ void init_algo_gate( algo_gate_t* gate )
   gate->get_nonceptr            = (void*)&std_get_nonceptr;
   gate->work_decode             = (void*)&std_le_work_decode;
   gate->decode_extra_data       = (void*)&do_nothing;
-   gate->wait_for_diff           = (void*)&std_wait_for_diff;
-   gate->get_max64               = (void*)&get_max64_0x1fffffLL;
   gate->gen_merkle_root         = (void*)&sha256d_gen_merkle_root;
   gate->stratum_gen_work        = (void*)&std_stratum_gen_work;
   gate->build_stratum_request   = (void*)&std_le_build_stratum_request;
   gate->malloc_txs_request      = (void*)&std_malloc_txs_request;
-   gate->set_target              = (void*)&std_set_target;
   gate->submit_getwork_result   = (void*)&std_le_submit_getwork_result;
   gate->build_block_header      = (void*)&std_build_block_header;
   gate->build_extraheader       = (void*)&std_build_extraheader;
@@ -205,6 +202,7 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    case ALGO_PHI2:          register_phi2_algo          ( gate ); break;
    case ALGO_PLUCK:         register_pluck_algo         ( gate ); break;
    case ALGO_POLYTIMOS:     register_polytimos_algo     ( gate ); break;
+    case ALGO_POWER2B:       register_power2b_algo       ( gate ); break;
    case ALGO_QUARK:         register_quark_algo         ( gate ); break;
    case ALGO_QUBIT:         register_qubit_algo         ( gate ); break;
    case ALGO_SCRYPT:        register_scrypt_algo        ( gate ); break;
@@ -234,6 +232,7 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    case ALGO_X14:           register_x14_algo           ( gate ); break;
    case ALGO_X15:           register_x15_algo           ( gate ); break;
    case ALGO_X16R:          register_x16r_algo          ( gate ); break;
+    case ALGO_X16RV2:        register_x16rv2_algo        ( gate ); break;
    case ALGO_X16RT:         register_x16rt_algo         ( gate ); break;
    case ALGO_X16RT_VEIL:    register_x16rt_veil_algo    ( gate ); break;
    case ALGO_X16S:          register_x16s_algo          ( gate ); break;
@@ -251,6 +250,7 @@ bool register_algo_gate( int algo, algo_gate_t *gate )
    case ALGO_YESCRYPTR32:   register_yescryptr32_algo   ( gate ); break;
    case ALGO_YESPOWER:      register_yespower_algo      ( gate ); break;
    case ALGO_YESPOWERR16:   register_yespowerr16_algo   ( gate ); break;
+    case ALGO_YESPOWER_B2B:  register_yespower_b2b_algo  ( gate ); break;
    case ALGO_ZR5:           register_zr5_algo           ( gate ); break;
   default:
      applog(LOG_ERR,"FAIL: algo_gate registration failed, unknown algo %s.\n", algo_names[opt_algo] );
@@ -276,7 +276,7 @@ bool register_json_rpc2( algo_gate_t *gate )
  applog(LOG_WARNING,"supported by cpuminer-opt. Shares submitted will");
  applog(LOG_WARNING,"likely be rejected. Proceed at your own risk.\n");

-  gate->wait_for_diff           = (void*)&do_nothing;
+//  gate->wait_for_diff           = (void*)&do_nothing;
  gate->get_new_work            = (void*)&jr2_get_new_work;
  gate->get_nonceptr            = (void*)&jr2_get_nonceptr;
  gate->stratum_gen_work        = (void*)&jr2_stratum_gen_work;
@@ -337,7 +337,6 @@ const char* const algo_alias_map[][2] =
  { "myriad",            "myr-gr"       },
  { "neo",               "neoscrypt"    },
  { "phi",               "phi1612"      },
-//  { "sia",               "blake2b"      },
  { "sib",               "x11gost"      },
  { "timetravel8",       "timetravel"   },
  { "veil",              "x16rt-veil"   },
@@ -365,40 +364,3 @@ void get_algo_alias( char** algo_or_alias )
 #undef ALIAS
 #undef PROPER

-bool submit_solution( struct work *work, void *hash,
-                      struct thr_info *thr )
-{
-     work_set_target_ratio( work, hash );
-     if ( submit_work( thr, work ) )
-     {
-         if ( !opt_quiet )
-            applog( LOG_BLUE, "Share %d submitted by thread %d, job %s.",
-                    accepted_share_count + rejected_share_count + 1,
-                    thr->id, work->job_id );
-         return true;
-     }
-     else
-          applog( LOG_WARNING, "Failed to submit share." );
-     return false;
-}
-
-bool submit_lane_solution( struct work *work, void *hash,
-                           struct thr_info *thr, int lane )
-{
-     work_set_target_ratio( work, hash );
-     if ( submit_work( thr, work ) )
-     {
-         if ( !opt_quiet )
-//            applog( LOG_BLUE, "Share %d submitted by thread %d, lane %d.",
-//                    accepted_share_count + rejected_share_count + 1,
-//                    thr->id, lane );
-            applog( LOG_BLUE, "Share %d submitted by thread %d, lane %d, job %s.",
-                    accepted_share_count + rejected_share_count + 1, thr->id,
-                    lane, work->job_id );
-         return true;
-     }
-     else
-          applog( LOG_WARNING, "Failed to submit share." );
-     return false;
-}
-
--- a/algo-gate-api.h
+++ b/algo-gate-api.h
@@ -35,7 +35,7 @@
 //    6. Determine if other non existant functions are required.
 //    That is determined by the need to add code in cpu-miner.c
 //    that applies only to the new algo. That is forbidden. All
-//    algo specific code must be in theh algo's file.
+//    algo specific code must be in the algo's file.
 //
 //    7. If new functions need to be added to the gate add the type
 //    to the structure, declare a null instance in this file and define
@@ -48,10 +48,10 @@
 //    instances as they are defined by default, or unsafe functions that
 //    are not needed by the algo.
 //
-//    9. Add an case entry to the switch/case in function register_gate
+//    9. Add a case entry to the switch/case in function register_gate
 //    in file algo-gate-api.c for the new algo.
 //
-//    10 If a new function type was defined add an entry to ini talgo_gate
+//    10 If a new function type was defined add an entry to init algo_gate
 //    to initialize the new function to its null instance described in step 7.
 //
 //    11. If the new algo has aliases add them to the alias array in
@@ -85,14 +85,16 @@

 typedef  uint32_t set_t;

-#define EMPTY_SET       0
-#define SSE2_OPT        1
-#define AES_OPT         2  
-#define SSE42_OPT       4
-#define AVX_OPT         8
-#define AVX2_OPT     0x10
-#define SHA_OPT      0x20
-#define AVX512_OPT   0x40
+#define EMPTY_SET        0
+#define SSE2_OPT         1
+#define AES_OPT          2  
+#define SSE42_OPT        4
+#define AVX_OPT          8   // Sandybridge
+#define AVX2_OPT      0x10   // Haswell
+#define SHA_OPT       0x20   // sha256 (Ryzen, Ice Lake)
+#define AVX512_OPT    0x40   // AVX512- F, VL, DQ, BW (Skylake-X)
+#define VAES_OPT      0x80   // VAES (Ice Lake)
+

 // return set containing all elements from sets a & b
 inline set_t set_union ( set_t a, set_t b ) { return a | b; }
@@ -108,14 +110,7 @@ inline bool set_excl ( set_t a, set_t b ) { return (a & b) == 0; }

 typedef struct
 {
-// special case, only one target, provides a callback for scanhash to
-// submit work with less overhead.
-// bool (*submit_work )             ( struct thr_info*, const struct work* );
-
 // mandatory functions, must be overwritten
-// Added a 5th arg for the thread_info structure to replace the int thr id
-// in the first arg. Both will co-exist during the trasition.
-//int ( *scanhash ) ( int, struct work*, uint32_t, uint64_t* );
 int ( *scanhash ) ( struct work*, uint32_t, uint64_t*, struct thr_info* );

 // optional unsafe, must be overwritten if algo uses function
@@ -123,28 +118,55 @@ void ( *hash )     ( void*, const void*, uint32_t ) ;
 void ( *hash_suw ) ( void*, const void* );

 //optional, safe to use default in most cases
+
+// Allocate thread local buffers and other initialization specific to miner
+// threads.
 bool ( *miner_thread_init )      ( int );
+
+// Generate global blockheader from stratum data.
 void ( *stratum_gen_work )       ( struct stratum_ctx*, struct work* );
+
+// Get thread local copy of blockheader with unique nonce.
 void ( *get_new_work )           ( struct work*, struct work*, int, uint32_t*,
                                   bool );
+
+// Return pointer to nonce in blockheader.
 uint32_t *( *get_nonceptr )      ( uint32_t* );
-void ( *decode_extra_data )      ( struct work*, uint64_t* );
-void ( *wait_for_diff )          ( struct stratum_ctx* );
-int64_t ( *get_max64 )           ();
+
+// Decode getwork blockheader
 bool ( *work_decode )            ( const json_t*, struct work* );
-void ( *set_target)              ( struct work*, double );
+
+// Extra getwork data
+void ( *decode_extra_data )      ( struct work*, uint64_t* );
+
 bool ( *submit_getwork_result )  ( CURL*, struct work* );
+
 void ( *gen_merkle_root )        ( char*, struct stratum_ctx* );
+
+// Increment extranonce
 void ( *build_extraheader )      ( struct work*, struct stratum_ctx* );
+
 void ( *build_block_header )     ( struct work*, uint32_t, uint32_t*,
-	                           uint32_t*, uint32_t, uint32_t );
+	                                uint32_t*, uint32_t, uint32_t );
+// Build mining.submit message
 void ( *build_stratum_request )  ( char*, struct work*, struct stratum_ctx* );
+
 char* ( *malloc_txs_request )    ( struct work* );
+
+// Big or little
 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 );
-void ( *resync_threads )         ( struct work* );
+
+// Diverge mining threads
 bool ( *do_this_thread )         ( int );
+
+// After do_this_thread
+void ( *resync_threads )         ( struct work* );
+
 json_t* (*longpoll_rpc_call)     ( CURL*, int*, char* );
 bool ( *stratum_handle_response )( json_t* );
 set_t optimizations;
@@ -193,23 +215,12 @@ void four_way_not_tested();
 // allways returns failure
 int null_scanhash();

-// Allow algos to submit from scanhash loop.
-bool submit_solution( struct work *work, void *hash,
-                      struct thr_info *thr );
-bool submit_lane_solution( struct work *work, void *hash,
-                          struct thr_info *thr, int lane );
-
- 
-bool submit_work( struct thr_info *thr, const struct work *work_in );
-
 // displays warning
 void null_hash    ();
 void null_hash_suw();

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

-void std_wait_for_diff();
-
 uint32_t *std_get_nonceptr( uint32_t *work_data );
 uint32_t *jr2_get_nonceptr( uint32_t *work_data );

@@ -224,25 +235,13 @@ void jr2_stratum_gen_work( struct stratum_ctx *sctx, struct work *work );
 void sha256d_gen_merkle_root( char *merkle_root, struct stratum_ctx *sctx );
 void SHA256_gen_merkle_root ( char *merkle_root, struct stratum_ctx *sctx );

-// pick your favorite or define your own
-int64_t get_max64_0x1fffffLL(); // default
-int64_t get_max64_0x40LL();
-int64_t get_max64_0x3ffff();
-int64_t get_max64_0x3fffffLL();
-int64_t get_max64_0x1ffff();
-int64_t get_max64_0xffffLL();
-
-void std_set_target(    struct work *work, double job_diff );
-void alt_set_target(    struct work* work, double job_diff );
-void scrypt_set_target( struct work *work, double job_diff );
-
 bool std_le_work_decode( const json_t *val, struct work *work );
 bool std_be_work_decode( const json_t *val, struct work *work );
-bool jr2_work_decode( const json_t *val, struct work *work );
+bool jr2_work_decode(    const json_t *val, struct work *work );

 bool std_le_submit_getwork_result( CURL *curl, struct work *work );
 bool std_be_submit_getwork_result( CURL *curl, struct work *work );
-bool jr2_submit_getwork_result( CURL *curl, struct work *work );
+bool jr2_submit_getwork_result(    CURL *curl, struct work *work );

 void std_le_build_stratum_request( char *req, struct work *work );
 void std_be_build_stratum_request( char *req, struct work *work );
@@ -256,8 +255,8 @@ 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 );
+	                          uint32_t *prevhash,  uint32_t *merkle_root,
+   	                       uint32_t ntime,      uint32_t nbits );

 void std_build_extraheader( struct work *work, struct stratum_ctx *sctx );

@@ -278,8 +277,8 @@ int std_get_work_data_size();
 // by calling the algo's register function.
 bool register_algo_gate( int algo, algo_gate_t *gate );

-// Override any default gate functions that are applicable and do any other
-// algo-specific initialization.
+// Called by algos toverride any default gate functions that are applicable
+// and do any other algo-specific initialization.
 // The register functions for all the algos can be declared here to reduce
 // compiler warnings but that's just more work for devs adding new algos.
 bool register_algo( algo_gate_t *gate );
@@ -292,5 +291,7 @@ bool register_json_rpc2( algo_gate_t *gate );
 // use this to call the hash function of an algo directly, ie util.c test.
 void exec_hash_function( int algo, void *output, const void *pdata );

-void get_algo_alias( char** algo_or_alias );
+// Validate a string as a known algo and alias, updates arg to proper
+// algo name if valid alias, NULL if invalid alias or algo.
+void get_algo_alias( char **algo_or_alias );

--- a/algo/argon2/argon2a/argon2a.c
+++ b/algo/argon2/argon2a/argon2a.c
@@ -74,19 +74,14 @@ int scanhash_argon2( struct work* work, uint32_t max_nonce,
 	return 0;
 }

-int64_t argon2_get_max64 ()
-{
-  return 0x1ffLL;
-}
-
 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->set_target      = (void*)&scrypt_set_target;
-  gate->get_max64       = (void*)&argon2_get_max64;
+  opt_target_factor = 65536.0;
+
  return true;
 };

--- a/algo/argon2/argon2d/argon2d-gate.c
+++ b/algo/argon2/argon2d/argon2d-gate.c
@@ -67,8 +67,8 @@ bool register_argon2d_crds_algo( algo_gate_t* gate )
 {
        gate->scanhash = (void*)&scanhash_argon2d_crds;
        gate->hash = (void*)&argon2d_crds_hash;
-        gate->set_target = (void*)&scrypt_set_target;
        gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
+        opt_target_factor = 65536.0;
        return true;
 }

@@ -135,8 +135,8 @@ bool register_argon2d_dyn_algo( algo_gate_t* gate )
 {
        gate->scanhash = (void*)&scanhash_argon2d_dyn;
        gate->hash = (void*)&argon2d_dyn_hash;
-        gate->set_target = (void*)&scrypt_set_target;
        gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
+        opt_target_factor = 65536.0;
        return true;
 }

@@ -179,14 +179,11 @@ int scanhash_argon2d4096( struct work *work, uint32_t max_nonce,
   return 0;
 }

-int64_t get_max64_0x1ff() { return 0x1ff; }
-
 bool register_argon2d4096_algo( algo_gate_t* gate )
 {
        gate->scanhash = (void*)&scanhash_argon2d4096;
-        gate->set_target = (void*)&scrypt_set_target;
-        gate->get_max64  = (void*)&get_max64_0x1ff;
        gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
+        opt_target_factor = 65536.0;
        return true;
 }

--- a/algo/blake/blake-gate.c
+++ b/algo/blake/blake-gate.c
@@ -1,18 +1,8 @@
 #include "blake-gate.h"

-int64_t blake_get_max64 ()
-{
-  return 0x7ffffLL;
-}
-
 bool register_blake_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT;
-  gate->get_max64 = (void*)&blake_get_max64;
-//#if defined (__AVX2__) && defined (FOUR_WAY)
-//   gate->optimizations = SSE2_OPT | AVX2_OPT;
-//  gate->scanhash  = (void*)&scanhash_blake_8way;
-//  gate->hash      = (void*)&blakehash_8way;
 #if defined(BLAKE_4WAY)
  four_way_not_tested();
  gate->scanhash  = (void*)&scanhash_blake_4way;
--- a/algo/blake/blake2b-gate.c
+++ b/algo/blake/blake2b-gate.c
@@ -1,13 +1,5 @@
 #include "blake2b-gate.h"

-/*
-// changed to get_max64_0x3fffffLL in cpuminer-multi-decred
-int64_t blake2s_get_max64 ()
-{
-   return 0x7ffffLL;
-}
-*/
-
 bool register_blake2b_algo( algo_gate_t* gate )
 {
 #if defined(BLAKE2B_4WAY)
@@ -17,7 +9,6 @@ bool register_blake2b_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_blake2b;
  gate->hash      = (void*)&blake2b_hash;
 #endif
-//  gate->get_max64 = (void*)&blake2s_get_max64;
  gate->optimizations =  AVX2_OPT;
  return true;
 };
--- a/algo/blake/blake2s-gate.c
+++ b/algo/blake/blake2s-gate.c
@@ -1,12 +1,5 @@
 #include "blake2s-gate.h"

-
-// changed to get_max64_0x3fffffLL in cpuminer-multi-decred
-int64_t blake2s_get_max64 ()
-{
-   return 0x7ffffLL;
-}
-
 bool register_blake2s_algo( algo_gate_t* gate )
 {
 #if defined(BLAKE2S_8WAY)
@@ -19,7 +12,6 @@ bool register_blake2s_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_blake2s;
  gate->hash      = (void*)&blake2s_hash;
 #endif
-  gate->get_max64 = (void*)&blake2s_get_max64;
  gate->optimizations = SSE2_OPT | AVX2_OPT;
  return true;
 };
--- a/algo/blake/blake2s.c
+++ b/algo/blake/blake2s.c
@@ -70,18 +70,3 @@ int scanhash_blake2s( struct work *work,

 	return 0;
 }
-/*
-// changed to get_max64_0x3fffffLL in cpuminer-multi-decred
-int64_t blake2s_get_max64 ()
-{
-   return 0x7ffffLL;
-}
-
-bool register_blake2s_algo( algo_gate_t* gate )
-{
-  gate->scanhash  = (void*)&scanhash_blake2s;
-  gate->hash      = (void*)&blake2s_hash;
-  gate->get_max64 = (void*)&blake2s_get_max64;
-  return true;
-};
-*/
--- a/algo/blake/blake512-hash-4way.c
+++ b/algo/blake/blake512-hash-4way.c
@@ -403,7 +403,9 @@ static const sph_u64 CB[16] = {
 	__m256i M[16]; \
 	__m256i V0, V1, V2, V3, V4, V5, V6, V7; \
 	__m256i V8, V9, VA, VB, VC, VD, VE, VF; \
-	unsigned r; \
+   const __m256i shuff_bswap64 = m256_const2_64( 0x08090a0b0c0d0e0f, \
+                                                 0x0001020304050607 ) \
+   unsigned r; \
 	V0 = H0; \
 	V1 = H1; \
 	V2 = H2; \
@@ -412,53 +414,53 @@ static const sph_u64 CB[16] = {
 	V5 = H5; \
 	V6 = H6; \
 	V7 = H7; \
-   V8 = _mm256_xor_si256( S0, _mm256_set_epi64x( CB0, CB0, CB0, CB0 ) ); \
-   V9 = _mm256_xor_si256( S1, _mm256_set_epi64x( CB1, CB1, CB1, CB1 ) ); \
-   VA = _mm256_xor_si256( S2, _mm256_set_epi64x( CB2, CB2, CB2, CB2 ) ); \
-   VB = _mm256_xor_si256( S3, _mm256_set_epi64x( CB3, CB3, CB3, CB3 ) ); \
-   VC = _mm256_xor_si256( _mm256_set_epi64x( T0, T0, T0, T0 ), \
-                          _mm256_set_epi64x( CB4, CB4, CB4, CB4 ) ); \
-   VD = _mm256_xor_si256( _mm256_set_epi64x( T0, T0, T0, T0 ), \
-                          _mm256_set_epi64x( CB5, CB5, CB5, CB5 ) ); \
-   VE = _mm256_xor_si256( _mm256_set_epi64x( T1, T1, T1, T1 ), \
-                          _mm256_set_epi64x( CB6, CB6, CB6, CB6 ) ); \
-   VF = _mm256_xor_si256( _mm256_set_epi64x( T1, T1, T1, T1 ), \
-                          _mm256_set_epi64x( CB7, CB7, CB7, CB7 ) ); \
-	M[0x0] = mm256_bswap_64( *(buf+0) ); \
-	M[0x1] = mm256_bswap_64( *(buf+1) ); \
-	M[0x2] = mm256_bswap_64( *(buf+2) ); \
-	M[0x3] = mm256_bswap_64( *(buf+3) ); \
-	M[0x4] = mm256_bswap_64( *(buf+4) ); \
-	M[0x5] = mm256_bswap_64( *(buf+5) ); \
-	M[0x6] = mm256_bswap_64( *(buf+6) ); \
-	M[0x7] = mm256_bswap_64( *(buf+7) ); \
-	M[0x8] = mm256_bswap_64( *(buf+8) ); \
-	M[0x9] = mm256_bswap_64( *(buf+9) ); \
-	M[0xA] = mm256_bswap_64( *(buf+10) ); \
-	M[0xB] = mm256_bswap_64( *(buf+11) ); \
-	M[0xC] = mm256_bswap_64( *(buf+12) ); \
-	M[0xD] = mm256_bswap_64( *(buf+13) ); \
-	M[0xE] = mm256_bswap_64( *(buf+14) ); \
-	M[0xF] = mm256_bswap_64( *(buf+15) ); \
+   V8 = _mm256_xor_si256( S0, _mm256_set1_epi64x( CB0 ) ); \
+   V9 = _mm256_xor_si256( S1, _mm256_set1_epi64x( CB1 ) ); \
+   VA = _mm256_xor_si256( S2, _mm256_set1_epi64x( CB2 ) ); \
+   VB = _mm256_xor_si256( S3, _mm256_set1_epi64x( CB3 ) ); \
+   VC = _mm256_xor_si256( _mm256_set1_epi64x( T0 ), \
+                          _mm256_set1_epi64x( CB4 ) ); \
+   VD = _mm256_xor_si256( _mm256_set1_epi64x( T0 ), \
+                          _mm256_set1_epi64x( CB5 ) ); \
+   VE = _mm256_xor_si256( _mm256_set1_epi64x( T1 ), \
+                          _mm256_set1_epi64x( CB6 ) ); \
+   VF = _mm256_xor_si256( _mm256_set1_epi64x( T1 ), \
+                          _mm256_set1_epi64x( CB7, CB7, CB7, CB7 ) ); \
+   M[0x0] = _mm256_shuffle_epi8( *(buf+ 0), shuff_bswap64 ); \
+	M[0x1] = _mm256_shuffle_epi8( *(buf+ 1), shuff_bswap64 ); \
+	M[0x2] = _mm256_shuffle_epi8( *(buf+ 2), shuff_bswap64 ); \
+	M[0x3] = _mm256_shuffle_epi8( *(buf+ 3), shuff_bswap64 ); \
+	M[0x4] = _mm256_shuffle_epi8( *(buf+ 4), shuff_bswap64 ); \
+	M[0x5] = _mm256_shuffle_epi8( *(buf+ 5), shuff_bswap64 ); \
+	M[0x6] = _mm256_shuffle_epi8( *(buf+ 6), shuff_bswap64 ); \
+	M[0x7] = _mm256_shuffle_epi8( *(buf+ 7), shuff_bswap64 ); \
+	M[0x8] = _mm256_shuffle_epi8( *(buf+ 8), shuff_bswap64 ); \
+	M[0x9] = _mm256_shuffle_epi8( *(buf+ 9), shuff_bswap64 ); \
+	M[0xA] = _mm256_shuffle_epi8( *(buf+10), shuff_bswap64 ); \
+	M[0xB] = _mm256_shuffle_epi8( *(buf+11), shuff_bswap64 ); \
+	M[0xC] = _mm256_shuffle_epi8( *(buf+12), shuff_bswap64 ); \
+	M[0xD] = _mm256_shuffle_epi8( *(buf+13), shuff_bswap64 ); \
+	M[0xE] = _mm256_shuffle_epi8( *(buf+14), shuff_bswap64 ); \
+	M[0xF] = _mm256_shuffle_epi8( *(buf+15), shuff_bswap64 ); \
 	for (r = 0; r < 16; r ++) \
 		ROUND_B_4WAY(r); \
-        H0 = _mm256_xor_si256( _mm256_xor_si256( \
+   H0 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S0, V0 ), V8 ), H0 ); \
-        H1 = _mm256_xor_si256( _mm256_xor_si256( \
+   H1 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S1, V1 ), V9 ), H1 ); \
-        H2 = _mm256_xor_si256( _mm256_xor_si256( \
+   H2 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S2, V2 ), VA ), H2 ); \
-        H3 = _mm256_xor_si256( _mm256_xor_si256( \
+   H3 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S3, V3 ), VB ), H3 ); \
-        H4 = _mm256_xor_si256( _mm256_xor_si256( \
+   H4 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S0, V4 ), VC ), H4 ); \
-        H5 = _mm256_xor_si256( _mm256_xor_si256( \
+   H5 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S1, V5 ), VD ), H5 ); \
-        H6 = _mm256_xor_si256( _mm256_xor_si256( \
+   H6 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S2, V6 ), VE ), H6 ); \
-        H7 = _mm256_xor_si256( _mm256_xor_si256( \
+   H7 = _mm256_xor_si256( _mm256_xor_si256( \
                    _mm256_xor_si256( S3, V7 ), VF ), H7 ); \
-	} while (0)
+} while (0)

 #else

@@ -491,8 +493,7 @@ static const sph_u64 CB[16] = {
                         m256_const1_64( CB6 ) );  \
  VF = _mm256_xor_si256( _mm256_set1_epi64x( T1 ), \
                         m256_const1_64( CB7 ) );  \
-  shuf_bswap64 = m256_const_64( 0x08090a0b0c0d0e0f, 0x0001020304050607, \
-                                0x08090a0b0c0d0e0f, 0x0001020304050607 ); \
+  shuf_bswap64 = m256_const2_64( 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 ); \
@@ -620,7 +621,7 @@ blake64_4way_close( blake_4way_big_context *sc,
   bit_len = ((unsigned)ptr << 3);
   z = 0x80 >> n;
   zz = ((ub & -z) | z) & 0xFF;
-   buf[ptr>>3] = _mm256_set_epi64x( zz, zz, zz, zz );
+   buf[ptr>>3] = _mm256_set1_epi64x( zz );
   tl = sc->T0 + bit_len;
   th = sc->T1;
   if (ptr == 0 )
--- a/algo/blake/blakecoin-gate.c
+++ b/algo/blake/blakecoin-gate.c
@@ -1,13 +1,6 @@
 #include "blakecoin-gate.h"
 #include <memory.h>

-// changed to get_max64_0x3fffffLL in cpuminer-multi-decred
-int64_t blakecoin_get_max64 ()
-{
-  return 0x7ffffLL;
-//  return 0x3fffffLL;
-}
-
 // vanilla uses default gen merkle root, otherwise identical to blakecoin
 bool register_vanilla_algo( algo_gate_t* gate )
 {
@@ -23,7 +16,6 @@ bool register_vanilla_algo( algo_gate_t* gate )
  gate->hash     = (void*)&blakecoinhash;
 #endif
  gate->optimizations = SSE42_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&blakecoin_get_max64;
  return true;
 }

--- a/algo/blake/blakecoin.c
+++ b/algo/blake/blakecoin.c
@@ -93,33 +93,3 @@ int scanhash_blakecoin( struct work *work, uint32_t max_nonce,
 	return 0;
 }

-/*
-void blakecoin_gen_merkle_root ( char* merkle_root, struct stratum_ctx* sctx )
-{
- SHA256( sctx->job.coinbase, (int)sctx->job.coinbase_size, merkle_root );
-}
-*/
-/*
-// changed to get_max64_0x3fffffLL in cpuminer-multi-decred
-int64_t blakecoin_get_max64 ()
-{
-  return 0x7ffffLL;
-}
-
-// vanilla uses default gen merkle root, otherwise identical to blakecoin
-bool register_vanilla_algo( algo_gate_t* gate )
-{
-    gate->scanhash = (void*)&scanhash_blakecoin;
-    gate->hash     = (void*)&blakecoinhash;
-    gate->get_max64 = (void*)&blakecoin_get_max64;
-    blakecoin_init( &blake_init_ctx );
-    return true;
-}
-
-bool register_blakecoin_algo( algo_gate_t* gate )
-{
-  register_vanilla_algo( gate );
-  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
-  return true;
-}
-*/
--- a/algo/blake/decred-gate.c
+++ b/algo/blake/decred-gate.c
@@ -38,7 +38,7 @@ void decred_decode_extradata( struct work* work, uint64_t* net_blocks )
   if (!have_longpoll && work->height > *net_blocks + 1)
   {
      char netinfo[64] = { 0 };
-      if (opt_showdiff && net_diff > 0.)
+      if ( net_diff > 0. )
      {
         if (net_diff != work->targetdiff)
            sprintf(netinfo, ", diff %.3f, target %.1f", net_diff,
@@ -116,7 +116,7 @@ void decred_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
   // 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];
+   sctx->block_height = g_work->data[32];
   //applog_hex(work->data, 180);
   //applog_hex(&work->data[36], 36);
 }
@@ -154,7 +154,6 @@ bool register_decred_algo( algo_gate_t* gate )
 #endif
  gate->optimizations = AVX2_OPT;
  gate->get_nonceptr          = (void*)&decred_get_nonceptr;
-  gate->get_max64             = (void*)&get_max64_0x3fffffLL;
  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;
--- a/algo/blake/decred.c
+++ b/algo/blake/decred.c
@@ -143,7 +143,7 @@ void decred_decode_extradata( struct work* work, uint64_t* net_blocks )
   if (!have_longpoll && work->height > *net_blocks + 1)
   {
      char netinfo[64] = { 0 };
-      if (opt_showdiff && net_diff > 0.)
+      if (net_diff > 0.)
      {
         if (net_diff != work->targetdiff)
 	    sprintf(netinfo, ", diff %.3f, target %.1f", net_diff,
@@ -269,7 +269,6 @@ bool register_decred_algo( algo_gate_t* gate )
  gate->scanhash              = (void*)&scanhash_decred;
  gate->hash                  = (void*)&decred_hash;
  gate->get_nonceptr          = (void*)&decred_get_nonceptr;
-  gate->get_max64             = (void*)&get_max64_0x3fffffLL;
  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;
--- a/algo/blake/pentablake-gate.c
+++ b/algo/blake/pentablake-gate.c
@@ -10,7 +10,6 @@ bool register_pentablake_algo( algo_gate_t* gate )
    gate->hash      = (void*)&pentablakehash;
 #endif
    gate->optimizations = AVX2_OPT;
-    gate->get_max64 = (void*)&get_max64_0x3ffff;
    return true;
 };

--- a/algo/bmw/bmw512-gate.c
+++ b/algo/bmw/bmw512-gate.c
@@ -1,12 +1,9 @@
 #include "bmw512-gate.h"

-int64_t bmw512_get_max64() { return 0x7ffffLL; }
-
 bool register_bmw512_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT;
-  gate->set_target      = (void*)&alt_set_target;
-  gate->get_max64       = (void*)&bmw512_get_max64;
+  opt_target_factor = 256.0;
 #if defined (BMW512_4WAY)
  gate->scanhash  = (void*)&scanhash_bmw512_4way;
  gate->hash      = (void*)&bmw512hash_4way;
--- a/algo/cryptonight/cryptolight.c
+++ b/algo/cryptonight/cryptolight.c
@@ -363,7 +363,6 @@ bool register_cryptolight_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_cryptolight;
  gate->hash      = (void*)&cryptolight_hash;
  gate->hash_suw  = (void*)&cryptolight_hash; 
-  gate->get_max64 = (void*)&get_max64_0x40LL;
  return true;
 };

--- a/algo/cryptonight/cryptonight-common.c
+++ b/algo/cryptonight/cryptonight-common.c
@@ -111,7 +111,6 @@ bool register_cryptonight_algo( algo_gate_t* gate )
  gate->scanhash         = (void*)&scanhash_cryptonight;
  gate->hash             = (void*)&cryptonight_hash;
  gate->hash_suw         = (void*)&cryptonight_hash_suw;  
-  gate->get_max64        = (void*)&get_max64_0x40LL;
  return true;
 };

@@ -123,7 +122,6 @@ bool register_cryptonightv7_algo( algo_gate_t* gate )
  gate->scanhash      = (void*)&scanhash_cryptonight;
  gate->hash          = (void*)&cryptonight_hash;
  gate->hash_suw      = (void*)&cryptonight_hash_suw;
-  gate->get_max64     = (void*)&get_max64_0x40LL;
  return true;
 };

--- a/algo/cubehash/cube-hash-2way.c
+++ b/algo/cubehash/cube-hash-2way.c
@@ -7,7 +7,7 @@

 // 2x128

-/*
+
 // The result of hashing 10 rounds of initial data which consists of params
 // zero padded.
 static const uint64_t IV256[] =
@@ -25,7 +25,7 @@ static const uint64_t IV512[] =
 0x148FE485FCD398D9, 0xB64445321B017BEF, 0x2FF5781C6A536159, 0x0DBADEA991FA7934,
 0xA5A70E75D65C8A2B, 0xBC796576B1C62456, 0xE7989AF11921C8F7, 0xD43E3B447795D246
 };
-*/
+

 static void transform_2way( cube_2way_context *sp )
 {
@@ -97,39 +97,30 @@ static void transform_2way( cube_2way_context *sp )
 int cube_2way_init( cube_2way_context *sp, int hashbitlen, int rounds,
                    int blockbytes )
 {
-    __m128i* h = (__m128i*)sp->h;
+    __m256i *h = (__m256i*)sp->h;
+    __m128i *iv = (__m128i*)( hashbitlen == 512 ? (__m128i*)IV512
+                                                : (__m128i*)IV256 );
    sp->hashlen   = hashbitlen/128;
    sp->blocksize = blockbytes/16;
    sp->rounds    = rounds;
    sp->pos       = 0;

-    if ( hashbitlen == 512 )
-    {
-
-       h[ 0] = m128_const_64( 0x4167D83E2D538B8B, 0x50F494D42AEA2A61 );
-       h[ 2] = m128_const_64( 0x50AC5695CC39968E, 0xC701CF8C3FEE2313 );
-       h[ 4] = m128_const_64( 0x825B453797CF0BEF, 0xA647A8B34D42C787 );
-       h[ 6] = m128_const_64( 0xA23911AED0E5CD33, 0xF22090C4EEF864D2 );
-       h[ 8] = m128_const_64( 0xB64445321B017BEF, 0x148FE485FCD398D9 );
-       h[10] = m128_const_64( 0x0DBADEA991FA7934, 0x2FF5781C6A536159 );
-       h[12] = m128_const_64( 0xBC796576B1C62456, 0xA5A70E75D65C8A2B );
-       h[14] = m128_const_64( 0xD43E3B447795D246, 0xE7989AF11921C8F7 );
-       h[1] = h[ 0];  h[ 3] = h[ 2]; h[ 5] = h[ 4]; h[ 7] = h[ 6];
-       h[9] = h[ 8];  h[11] = h[10]; h[13] = h[12]; h[15] = h[14];
-    }
-    else
-    {
-       h[ 0] = m128_const_64( 0x35481EAE63117E71, 0xCCD6F29FEA2BD4B4 );
-       h[ 2] = m128_const_64( 0xF4CC12BE7E624131, 0xE5D94E6322512D5B );
-       h[ 4] = m128_const_64( 0x3361DA8CD0720C35, 0x42AF2070C2D0B696 );
-       h[ 6] = m128_const_64( 0x40E5FBAB4680AC00, 0x8EF8AD8328CCECA4 );
-       h[ 8] = m128_const_64( 0xF0B266796C859D41, 0x6107FBD5D89041C3 );
-       h[10] = m128_const_64( 0x93CB628565C892FD, 0x5FA2560309392549 );
-       h[12] = m128_const_64( 0x85254725774ABFDD, 0x9E4B4E602AF2B5AE );
-       h[14] = m128_const_64( 0xD6032C0A9CDAF8AF, 0x4AB6AAD615815AEB );
-       h[1] = h[ 0];  h[ 3] = h[ 2]; h[ 5] = h[ 4]; h[ 7] = h[ 6];
-       h[9] = h[ 8];  h[11] = h[10]; h[13] = h[12]; h[15] = h[14];
-    }
+    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] = 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] );
    
    return 0;
 }
@@ -164,11 +155,11 @@ 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 ],
-                                _mm256_set_epi32( 0,0,0,0x80,  0,0,0,0x80 ) );
+                                   m256_const2_64( 0, 0x0000000000000080 ) );
    transform_2way( sp );

    sp->h[7] = _mm256_xor_si256( sp->h[7],
-		                 _mm256_set_epi32( 1,0,0,0,  1,0,0,0 ) );
+                                   m256_const2_64( 0x0000000100000000, 0 ) );

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

@@ -197,13 +188,13 @@ 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 ],
-                    _mm256_set_epi32( 0,0,0,0x80,  0,0,0,0x80 ) );
+                                    m256_const2_64( 0, 0x0000000000000080 ) );
    transform_2way( sp );

-    sp->h[7] = _mm256_xor_si256( sp->h[7], _mm256_set_epi32( 1,0,0,0,
-                                                             1,0,0,0 ) );
+    sp->h[7] = _mm256_xor_si256( sp->h[7],
+                                    m256_const2_64( 0x0000000100000000, 0 ) );

-    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/groestl/groestl.c
+++ b/algo/groestl/groestl.c
@@ -94,19 +94,13 @@ int scanhash_groestl( struct work *work, uint32_t max_nonce,
 	return 0;
 }

-void groestl_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool register_dmd_gr_algo( algo_gate_t* gate )
 {
    init_groestl_ctx();
    gate->optimizations   = SSE2_OPT | AES_OPT;
    gate->scanhash        = (void*)&scanhash_groestl;
    gate->hash            = (void*)&groestlhash;
-    gate->set_target      = (void*)&groestl_set_target;
-    gate->get_max64       = (void*)&get_max64_0x3ffff;
+    opt_target_factor = 256.0;
    return true;
 };

--- a/algo/groestl/myr-groestl.c
+++ b/algo/groestl/myr-groestl.c
@@ -88,15 +88,3 @@ int scanhash_myriad( struct work *work, uint32_t max_nonce,
 	*hashes_done = pdata[19] - first_nonce + 1;
 	return 0;
 }
-/*
-bool register_myriad_algo( algo_gate_t* gate )
-{
-    gate->optimizations = SSE2_OPT | AES_OPT;
-    init_myrgr_ctx();
-    gate->scanhash = (void*)&scanhash_myriad;
-    gate->hash     = (void*)&myriadhash;
-//    gate->hash_alt = (void*)&myriadhash;
-    gate->get_max64 = (void*)&get_max64_0x3ffff;
-    return true;
-};
-*/
--- a/algo/groestl/myrgr-gate.c
+++ b/algo/groestl/myrgr-gate.c
@@ -12,7 +12,6 @@ bool register_myriad_algo( algo_gate_t* gate )
  gate->hash      = (void*)&myriad_hash;
 #endif
  gate->optimizations = AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/hamsi/hamsi-hash-4way.c
+++ b/algo/hamsi/hamsi-hash-4way.c
@@ -32,8 +32,6 @@

 #include <stddef.h>
 #include <string.h>
-
-//#include "miner.h"
 #include "hamsi-hash-4way.h"

 #if defined(__AVX2__)
@@ -100,7 +98,7 @@ extern "C"{
 #endif

 //#include "hamsi-helper-4way.c"
-
+/*
 static const sph_u32 IV512[] = {
 	SPH_C32(0x73746565), SPH_C32(0x6c706172), SPH_C32(0x6b204172),
 	SPH_C32(0x656e6265), SPH_C32(0x72672031), SPH_C32(0x302c2062),
@@ -109,7 +107,7 @@ static const sph_u32 IV512[] = {
 	SPH_C32(0x65766572), SPH_C32(0x6c65652c), SPH_C32(0x2042656c),
 	SPH_C32(0x6769756d)
 };
-
+*/
 static const sph_u32 alpha_n[] = {
 	SPH_C32(0xff00f0f0), SPH_C32(0xccccaaaa), SPH_C32(0xf0f0cccc),
 	SPH_C32(0xff00aaaa), SPH_C32(0xccccaaaa), SPH_C32(0xf0f0ff00),
@@ -138,6 +136,7 @@ static const sph_u32 alpha_f[] = {
 	SPH_C32(0xcaf9f9c0), SPH_C32(0x0ff0639c)
 };

+
 // imported from hamsi helper

 /* Note: this table lists bits within each byte from least
@@ -529,49 +528,34 @@ static const sph_u32 T512[64][16] = {
 	  SPH_C32(0xe7e00a94) }
 };

+
 #define INPUT_BIG \
 do { \
-  const __m256i zero = _mm256_setzero_si256(); \
  __m256i db = *buf; \
-  const sph_u32 *tp = &T512[0][0]; \
-  m0 = zero; \
-  m1 = zero; \
-  m2 = zero; \
-  m3 = zero; \
-  m4 = zero; \
-  m5 = zero; \
-  m6 = zero; \
-  m7 = zero; \
+  const uint64_t *tp = (uint64_t*)&T512[0][0];  \
+  m0 = m1 = m2 = m3 = m4 = m5 = m6 = m7 = m256_zero; \
  for ( int u = 0; u < 64; u++ ) \
  { \
     __m256i dm = _mm256_and_si256( db, m256_one_64 ) ; \
     dm = mm256_negate_32( _mm256_or_si256( dm, \
                         _mm256_slli_epi64( dm, 32 ) ) ); \
     m0 = _mm256_xor_si256( m0, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0x1], tp[0x0], tp[0x1], tp[0x0], \
-                                    tp[0x1], tp[0x0], tp[0x1], tp[0x0] ) ) ); \
+                                          m256_const1_64( tp[0] ) ) ); \
     m1 = _mm256_xor_si256( m1, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0x3], tp[0x2], tp[0x3], tp[0x2], \
-                                    tp[0x3], tp[0x2], tp[0x3], tp[0x2] ) ) ); \
+                                          m256_const1_64( tp[1] ) ) ); \
     m2 = _mm256_xor_si256( m2, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0x5], tp[0x4], tp[0x5], tp[0x4], \
-                                    tp[0x5], tp[0x4], tp[0x5], tp[0x4] ) ) ); \
+                                          m256_const1_64( tp[2] ) ) ); \
     m3 = _mm256_xor_si256( m3, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0x7], tp[0x6], tp[0x7], tp[0x6], \
-                                    tp[0x7], tp[0x6], tp[0x7], tp[0x6] ) ) ); \
+                                          m256_const1_64( tp[3] ) ) ); \
     m4 = _mm256_xor_si256( m4, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0x9], tp[0x8], tp[0x9], tp[0x8], \
-                                    tp[0x9], tp[0x8], tp[0x9], tp[0x8] ) ) ); \
+                                          m256_const1_64( tp[4] ) ) ); \
     m5 = _mm256_xor_si256( m5, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0xB], tp[0xA], tp[0xB], tp[0xA], \
-                                    tp[0xB], tp[0xA], tp[0xB], tp[0xA] ) ) ); \
+                                          m256_const1_64( tp[5] ) ) ); \
     m6 = _mm256_xor_si256( m6, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0xD], tp[0xC], tp[0xD], tp[0xC], \
-                                    tp[0xD], tp[0xC], tp[0xD], tp[0xC] ) ) ); \
+                                          m256_const1_64( tp[6] ) ) ); \
     m7 = _mm256_xor_si256( m7, _mm256_and_si256( dm, \
-                  _mm256_set_epi32( tp[0xF], tp[0xE], tp[0xF], tp[0xE], \
-                                    tp[0xF], tp[0xE], tp[0xF], tp[0xE] ) ) ); \
-     tp += 0x10; \
+                                          m256_const1_64( tp[7] ) ) ); \
+     tp += 8; \
     db = _mm256_srli_epi64( db, 1 ); \
  } \
 } while (0)
@@ -662,55 +646,39 @@ do { \

 #define ROUND_BIG(rc, alpha) \
 do { \
-  __m256i t0, t1, t2, t3; \
-  s0 = _mm256_xor_si256( s0, _mm256_set_epi32( \
-        alpha[0x01] ^ (rc), alpha[0x00], alpha[0x01] ^ (rc), alpha[0x00], \
-        alpha[0x01] ^ (rc), alpha[0x00], alpha[0x01] ^ (rc), alpha[0x00] ) ); \
-  s1 = _mm256_xor_si256( s1, _mm256_set_epi32( \
-                     alpha[0x03], alpha[0x02], alpha[0x03], alpha[0x02], \
-                     alpha[0x03], alpha[0x02], alpha[0x03], alpha[0x02] ) ); \
-  s2 = _mm256_xor_si256( s2, _mm256_set_epi32( \
-                     alpha[0x05], alpha[0x04], alpha[0x05], alpha[0x04], \
-                     alpha[0x05], alpha[0x04], alpha[0x05], alpha[0x04] ) ); \
-  s3 = _mm256_xor_si256( s3, _mm256_set_epi32( \
-                     alpha[0x07], alpha[0x06], alpha[0x07], alpha[0x06], \
-                     alpha[0x07], alpha[0x06], alpha[0x07], alpha[0x06] ) ); \
-  s4 = _mm256_xor_si256( s4, _mm256_set_epi32( \
-                     alpha[0x09], alpha[0x08], alpha[0x09], alpha[0x08], \
-                     alpha[0x09], alpha[0x08], alpha[0x09], alpha[0x08] ) ); \
-  s5 = _mm256_xor_si256( s5, _mm256_set_epi32( \
-                     alpha[0x0B], alpha[0x0A], alpha[0x0B], alpha[0x0A], \
-                     alpha[0x0B], alpha[0x0A], alpha[0x0B], alpha[0x0A] ) ); \
-  s6 = _mm256_xor_si256( s6, _mm256_set_epi32( \
-                     alpha[0x0D], alpha[0x0C], alpha[0x0D], alpha[0x0C], \
-                     alpha[0x0D], alpha[0x0C], alpha[0x0D], alpha[0x0C] ) ); \
-  s7 = _mm256_xor_si256( s7, _mm256_set_epi32( \
-                     alpha[0x0F], alpha[0x0E], alpha[0x0F], alpha[0x0E], \
-                     alpha[0x0F], alpha[0x0E], alpha[0x0F], alpha[0x0E] ) ); \
-  s8 = _mm256_xor_si256( s8, _mm256_set_epi32( \
-                     alpha[0x11], alpha[0x10], alpha[0x11], alpha[0x10], \
-                     alpha[0x11], alpha[0x10], alpha[0x11], alpha[0x10] ) ); \
-  s9 = _mm256_xor_si256( s9, _mm256_set_epi32( \
-                     alpha[0x13], alpha[0x12], alpha[0x13], alpha[0x12], \
-                     alpha[0x13], alpha[0x12], alpha[0x13], alpha[0x12] ) ); \
-  sA = _mm256_xor_si256( sA, _mm256_set_epi32( \
-                     alpha[0x15], alpha[0x14], alpha[0x15], alpha[0x14], \
-                     alpha[0x15], alpha[0x14], alpha[0x15], alpha[0x14] ) ); \
-  sB = _mm256_xor_si256( sB, _mm256_set_epi32( \
-                     alpha[0x17], alpha[0x16], alpha[0x17], alpha[0x16], \
-                     alpha[0x17], alpha[0x16], alpha[0x17], alpha[0x16] ) ); \
-  sC = _mm256_xor_si256( sC, _mm256_set_epi32( \
-                     alpha[0x19], alpha[0x18], alpha[0x19], alpha[0x18], \
-                     alpha[0x19], alpha[0x18], alpha[0x19], alpha[0x18] ) ); \
-  sD = _mm256_xor_si256( sD, _mm256_set_epi32( \
-                     alpha[0x1B], alpha[0x1A], alpha[0x1B], alpha[0x1A], \
-                     alpha[0x1B], alpha[0x1A], alpha[0x1B], alpha[0x1A] ) ); \
-  sE = _mm256_xor_si256( sE, _mm256_set_epi32( \
-                     alpha[0x1D], alpha[0x1C], alpha[0x1D], alpha[0x1C], \
-                     alpha[0x1D], alpha[0x1C], alpha[0x1D], alpha[0x1C] ) ); \
-  sF = _mm256_xor_si256( sF, _mm256_set_epi32( \
-                     alpha[0x1F], alpha[0x1E], alpha[0x1F], alpha[0x1E], \
-                     alpha[0x1F], alpha[0x1E], alpha[0x1F], alpha[0x1E] ) ); \
+   __m256i t0, t1, t2, t3; \
+   s0 = _mm256_xor_si256( s0, m256_const1_64( \
+        ( ( (uint64_t)( (rc) ^ alpha[1] ) << 32 ) ) | (uint64_t)alpha[0] ) ); \
+   s1 = _mm256_xor_si256( s1, m256_const1_64( \
+        ( (uint64_t)alpha[ 3] << 32 ) | (uint64_t)alpha[ 2] ) ); \
+   s2 = _mm256_xor_si256( s2, m256_const1_64( \
+        ( (uint64_t)alpha[ 5] << 32 ) | (uint64_t)alpha[ 4] ) ); \
+   s3 = _mm256_xor_si256( s3, m256_const1_64( \
+        ( (uint64_t)alpha[ 7] << 32 ) | (uint64_t)alpha[ 6] ) ); \
+   s4 = _mm256_xor_si256( s4, m256_const1_64( \
+        ( (uint64_t)alpha[ 9] << 32 ) | (uint64_t)alpha[ 8] ) ); \
+   s5 = _mm256_xor_si256( s5, m256_const1_64( \
+        ( (uint64_t)alpha[11] << 32 ) | (uint64_t)alpha[10] ) ); \
+   s6 = _mm256_xor_si256( s6, m256_const1_64( \
+        ( (uint64_t)alpha[13] << 32 ) | (uint64_t)alpha[12] ) ); \
+   s7 = _mm256_xor_si256( s7, m256_const1_64( \
+        ( (uint64_t)alpha[15] << 32 ) | (uint64_t)alpha[14] ) ); \
+   s8 = _mm256_xor_si256( s8, m256_const1_64( \
+        ( (uint64_t)alpha[17] << 32 ) | (uint64_t)alpha[16] ) ); \
+   s9 = _mm256_xor_si256( s9, m256_const1_64( \
+        ( (uint64_t)alpha[19] << 32 ) | (uint64_t)alpha[18] ) ); \
+   sA = _mm256_xor_si256( sA, m256_const1_64( \
+        ( (uint64_t)alpha[21] << 32 ) | (uint64_t)alpha[20] ) ); \
+   sB = _mm256_xor_si256( sB, m256_const1_64( \
+        ( (uint64_t)alpha[23] << 32 ) | (uint64_t)alpha[22] ) ); \
+   sC = _mm256_xor_si256( sC, m256_const1_64( \
+        ( (uint64_t)alpha[25] << 32 ) | (uint64_t)alpha[24] ) ); \
+   sD = _mm256_xor_si256( sD, m256_const1_64( \
+        ( (uint64_t)alpha[27] << 32 ) | (uint64_t)alpha[26] ) ); \
+   sE = _mm256_xor_si256( sE, m256_const1_64( \
+        ( (uint64_t)alpha[29] << 32 ) | (uint64_t)alpha[28] ) ); \
+   sF = _mm256_xor_si256( sF, m256_const1_64( \
+        ( (uint64_t)alpha[31] << 32 ) | (uint64_t)alpha[30] ) ); \
 \
  SBOX( s0, s4, s8, sC ); \
  SBOX( s1, s5, s9, sD ); \
@@ -864,47 +832,22 @@ void hamsi_big_final( hamsi_4way_big_context *sc, __m256i *buf )
 void hamsi512_4way_init( hamsi_4way_big_context *sc )
 {
   sc->partial_len = 0;
-   sph_u32 lo, hi;
   sc->count_high = sc->count_low = 0;
-   for ( int i = 0; i < 8; i++ )
-   {
-      lo = 2*i;
-      hi = 2*i + 1;
-      sc->h[i] = _mm256_set_epi32( IV512[hi], IV512[lo], IV512[hi], IV512[lo],
-                                   IV512[hi], IV512[lo], IV512[hi], IV512[lo] );
-   }
+
+   sc->h[0] = m256_const1_64( 0x6c70617273746565 );
+   sc->h[1] = m256_const1_64( 0x656e62656b204172 );
+   sc->h[2] = m256_const1_64( 0x302c206272672031 );
+   sc->h[3] = m256_const1_64( 0x3434362c75732032 );
+   sc->h[4] = m256_const1_64( 0x3030312020422d33 );
+   sc->h[5] = m256_const1_64( 0x656e2d484c657576 );
+   sc->h[6] = m256_const1_64( 0x6c65652c65766572 );
+   sc->h[7] = m256_const1_64( 0x6769756d2042656c );
 }

 void hamsi512_4way( hamsi_4way_big_context *sc, const void *data, size_t len )
 {
   __m256i *vdata = (__m256i*)data;

-// It looks like the only way to get in here is if core was previously called
-// with a very small len
-// That's not likely even with 80 byte input so deprecate partial len
-/*
-   if ( sc->partial_len != 0 )
-   {
-      size_t mlen;
-
-      mlen = 8 - sc->partial_len;
-      if ( len < mlen )
-      {
-         memcpy_256( sc->partial + (sc->partial_len >> 3), data, len>>3 );
-         sc->partial_len += len;
-         return;
-      }
-      else
-      {
-         memcpy_256( sc->partial + (sc->partial_len >> 3), data, mlen>>3 );
-         len -= mlen;
-         vdata += mlen>>3;
-         hamsi_big( sc, sc->partial, 1 );
-         sc->partial_len = 0;
-      }
-   }
-*/
-
   hamsi_big( sc, vdata, len>>3 );
   vdata += ( (len& ~(size_t)7) >> 3 );
   len &= (size_t)7;
@@ -920,8 +863,9 @@ void hamsi512_4way_close( hamsi_4way_big_context *sc, void *dst )
   sph_enc32be( &ch, sc->count_high );
   sph_enc32be( &cl, sc->count_low + ( sc->partial_len << 3 ) );
   pad[0] =  _mm256_set_epi32( cl, ch, cl, ch, cl, ch, cl, ch );
-   sc->buf[0] = _mm256_set_epi32( 0UL, 0x80UL, 0UL, 0x80UL,
-                                  0UL, 0x80UL, 0UL, 0x80UL );
+   sc->buf[0] = m256_const1_64( 0x80 );
+//      sc->buf[0] = _mm256_set_epi32( 0UL, 0x80UL, 0UL, 0x80UL,
+//                                  0UL, 0x80UL, 0UL, 0x80UL );
   hamsi_big( sc, sc->buf, 1 );
   hamsi_big_final( sc, pad );

--- a/algo/hodl/hodl-gate.c
+++ b/algo/hodl/hodl-gate.c
@@ -15,11 +15,6 @@ pthread_barrier_t hodl_barrier;
 // need to be passed.
 unsigned char *hodl_scratchbuf = NULL;

-void hodl_set_target( struct work* work, double diff )
-{
-     diff_to_target(work->target, diff / 8388608.0 );
-}
-
 void hodl_le_build_stratum_request( char* req, struct work* work,
                                    struct stratum_ctx *sctx ) 
 {
@@ -170,7 +165,6 @@ bool register_hodl_algo( algo_gate_t* gate )
  gate->scanhash              = (void*)&hodl_scanhash;
  gate->get_new_work          = (void*)&hodl_get_new_work;
  gate->longpoll_rpc_call     = (void*)&hodl_longpoll_rpc_call;
-  gate->set_target            = (void*)&hodl_set_target;
  gate->build_stratum_request = (void*)&hodl_le_build_stratum_request;
  gate->malloc_txs_request    = (void*)&hodl_malloc_txs_request;
  gate->build_block_header    = (void*)&hodl_build_block_header;
@@ -179,6 +173,7 @@ bool register_hodl_algo( algo_gate_t* gate )
  gate->work_cmp_size         = 76;
  hodl_scratchbuf = (unsigned char*)malloc( 1 << 30 );
  allow_getwork = false;
+  opt_target_factor = 8388608.0;
  return ( hodl_scratchbuf != NULL );
 }

--- a/algo/jh/jh-hash-4way.c
+++ b/algo/jh/jh-hash-4way.c
@@ -94,7 +94,7 @@ extern "C"{

 #define Sb(x0, x1, x2, x3, c) \
 do { \
-   __m256i cc = _mm256_set_epi64x( c, c, c, c ); \
+   __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 ) ); \
--- a/algo/jh/jha-gate.c
+++ b/algo/jh/jha-gate.c
@@ -12,7 +12,7 @@ bool register_jha_algo( algo_gate_t* gate )
  gate->hash             = (void*)&jha_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target       = (void*)&scrypt_set_target;
+  opt_target_factor = 65536.0;
  return true;
 };

--- a/algo/keccak/keccak-gate.c
+++ b/algo/keccak/keccak-gate.c
@@ -1,18 +1,11 @@
 #include "keccak-gate.h"

-void keccak_set_target( struct work* work, double job_diff )
-{
-  work_set_target( work, job_diff / (128.0 * opt_diff_factor) );
-}
-
-int64_t keccak_get_max64() { return 0x7ffffLL; }

 bool register_keccak_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT;
  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
-  gate->set_target      = (void*)&keccak_set_target;
-  gate->get_max64       = (void*)&keccak_get_max64;
+  opt_target_factor = 128.0;
 #if defined (KECCAK_4WAY)
  gate->scanhash  = (void*)&scanhash_keccak_4way;
  gate->hash      = (void*)&keccakhash_4way;
@@ -23,17 +16,11 @@ bool register_keccak_algo( algo_gate_t* gate )
  return true;
 };

-void keccakc_set_target( struct work* work, double job_diff )
-{
-  work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool register_keccakc_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT;
  gate->gen_merkle_root = (void*)&sha256d_gen_merkle_root;
-  gate->set_target      = (void*)&keccakc_set_target;
-  gate->get_max64       = (void*)&keccak_get_max64;
+  opt_target_factor = 256.0;
 #if defined (KECCAK_4WAY)
  gate->scanhash  = (void*)&scanhash_keccak_4way;
  gate->hash      = (void*)&keccakhash_4way;
--- a/algo/luffa/luffa-hash-2way.c
+++ b/algo/luffa/luffa-hash-2way.c
@@ -1,23 +1,3 @@
-/*
- * luffa_for_sse2.c
- * Version 2.0 (Sep 15th 2009)
- *
- * Copyright (C) 2008-2009 Hitachi, Ltd. All rights reserved.
- *
- * Hitachi, Ltd. is the owner of this software and hereby grant
- * the U.S. Government and any interested party the right to use
- * this software for the purposes of the SHA-3 evaluation process,
- * notwithstanding that this software is copyrighted.
- *
- * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
- * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
- * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
- * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
- * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
- * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
- * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
- */
-
 #include <string.h>
 #include <immintrin.h>
 #include "luffa-hash-2way.h"
@@ -26,31 +6,30 @@

 #include "simd-utils.h"

-#define MASK _mm256_set_epi32( 0UL, 0UL, 0UL, 0xffffffffUL, \
-                               0UL, 0UL, 0UL, 0xffffffffUL )
+#define cns(i)  m256_const1_128( ( (__m128i*)CNS_INIT)[i] )

 #define ADD_CONSTANT(a,b,c0,c1)\
    a = _mm256_xor_si256(a,c0);\
    b = _mm256_xor_si256(b,c1);\

-#define MULT2(a0,a1) \
+#define MULT2( a0, a1, mask ) \
 do { \
-  register __m256i b = _mm256_xor_si256( a0, \
-                   _mm256_shuffle_epi32( _mm256_and_si256(a1,MASK), 16 ) ); \
+  __m256i b = _mm256_xor_si256( a0, \
+                   _mm256_shuffle_epi32( _mm256_and_si256(a1,mask), 16 ) ); \
  a0 = _mm256_or_si256( _mm256_srli_si256(b,4), _mm256_slli_si256(a1,12) ); \
  a1 = _mm256_or_si256( _mm256_srli_si256(a1,4), _mm256_slli_si256(b,12) );  \
 } while(0)

 // confirm pointer arithmetic
 // ok but use array indexes
-#define STEP_PART(x,c,t)\
+#define STEP_PART(x,c0,c1,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));
+    ADD_CONSTANT(*x, *(x+4), c0, c1);

 #define SUBCRUMB(a0,a1,a2,a3,t)\
    t  = _mm256_load_si256(&a0);\
@@ -245,7 +224,7 @@ static const uint32 CNS_INIT[128] __attribute((aligned(32))) = {
    0x00000000,0x00000000,0x00000000,0xfc053c31
 };

-__m256i CNS[32];
+

 /***************************************************/
 /* Round function         */
@@ -258,6 +237,7 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
    __m256i msg0, msg1;
    __m256i tmp[2];
    __m256i x[8];
+    const __m256i MASK = m256_const2_64( 0, 0x00000000ffffffff );

    t0 = chainv[0];
    t1 = chainv[1];
@@ -271,7 +251,7 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
    t0 = _mm256_xor_si256( t0, chainv[8] );
    t1 = _mm256_xor_si256( t1, chainv[9] );

-    MULT2( t0, t1 );
+    MULT2( t0, t1, MASK );

    msg0 = _mm256_shuffle_epi32( msg[0], 27 );
    msg1 = _mm256_shuffle_epi32( msg[1], 27 );
@@ -290,66 +270,66 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
    t0 = chainv[0];
    t1 = chainv[1];

-    MULT2( chainv[0], chainv[1]);
+    MULT2( chainv[0], chainv[1], MASK );
    chainv[0] = _mm256_xor_si256( chainv[0], chainv[2] );
    chainv[1] = _mm256_xor_si256( chainv[1], chainv[3] );

-    MULT2( chainv[2], chainv[3]);
+    MULT2( chainv[2], chainv[3], MASK );
    chainv[2] = _mm256_xor_si256(chainv[2], chainv[4]);
    chainv[3] = _mm256_xor_si256(chainv[3], chainv[5]);

-    MULT2( chainv[4], chainv[5]);
+    MULT2( chainv[4], chainv[5], MASK );
    chainv[4] = _mm256_xor_si256(chainv[4], chainv[6]);
    chainv[5] = _mm256_xor_si256(chainv[5], chainv[7]);

-    MULT2( chainv[6], chainv[7]);
+    MULT2( chainv[6], chainv[7], MASK );
    chainv[6] = _mm256_xor_si256(chainv[6], chainv[8]);
    chainv[7] = _mm256_xor_si256(chainv[7], chainv[9]);

-    MULT2( chainv[8], chainv[9]);
+    MULT2( chainv[8], chainv[9], MASK );
    chainv[8] = _mm256_xor_si256( chainv[8], t0 );
    chainv[9] = _mm256_xor_si256( chainv[9], t1 );

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

-    MULT2( chainv[8], chainv[9]);
+    MULT2( chainv[8], chainv[9], MASK );
    chainv[8] = _mm256_xor_si256( chainv[8], chainv[6] );
    chainv[9] = _mm256_xor_si256( chainv[9], chainv[7] );

-    MULT2( chainv[6], chainv[7]);
+    MULT2( chainv[6], chainv[7], MASK );
    chainv[6] = _mm256_xor_si256( chainv[6], chainv[4] );
    chainv[7] = _mm256_xor_si256( chainv[7], chainv[5] );

-    MULT2( chainv[4], chainv[5]);
+    MULT2( chainv[4], chainv[5], MASK );
    chainv[4] = _mm256_xor_si256( chainv[4], chainv[2] );
    chainv[5] = _mm256_xor_si256( chainv[5], chainv[3] );

-    MULT2( chainv[2], chainv[3] );
+    MULT2( chainv[2], chainv[3], MASK );
    chainv[2] = _mm256_xor_si256( chainv[2], chainv[0] );
    chainv[3] = _mm256_xor_si256( chainv[3], chainv[1] );

-    MULT2( chainv[0], chainv[1] );
+    MULT2( chainv[0], chainv[1], MASK );
    chainv[0] = _mm256_xor_si256( _mm256_xor_si256( chainv[0], t0 ), msg0 );
    chainv[1] = _mm256_xor_si256( _mm256_xor_si256( chainv[1], t1 ), msg1 );

-    MULT2( msg0, msg1);
+    MULT2( msg0, msg1, MASK );
    chainv[2] = _mm256_xor_si256( chainv[2], msg0 );
    chainv[3] = _mm256_xor_si256( chainv[3], msg1 );

-    MULT2( msg0, msg1);
+    MULT2( msg0, msg1, MASK );
    chainv[4] = _mm256_xor_si256( chainv[4], msg0 );
    chainv[5] = _mm256_xor_si256( chainv[5], msg1 );

-    MULT2( msg0, msg1);
+    MULT2( msg0, msg1, MASK );
    chainv[6] = _mm256_xor_si256( chainv[6], msg0 );
    chainv[7] = _mm256_xor_si256( chainv[7], msg1 );

-    MULT2( msg0, msg1);
+    MULT2( msg0, msg1, MASK );
    chainv[8] = _mm256_xor_si256( chainv[8], msg0 );
    chainv[9] = _mm256_xor_si256( chainv[9], msg1 );

-    MULT2( msg0, msg1);
+    MULT2( msg0, msg1, MASK );

    chainv[3] = _mm256_or_si256( _mm256_slli_epi32( chainv[3],  1 ),
                                 _mm256_srli_epi32( chainv[3], 31 ) );
@@ -365,14 +345,14 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
                chainv[1],chainv[3],chainv[5],chainv[7],
                x[4], x[5], x[6], x[7] );

-    STEP_PART( &x[0], &CNS[ 0], &tmp[0] );
-    STEP_PART( &x[0], &CNS[ 2], &tmp[0] );
-    STEP_PART( &x[0], &CNS[ 4], &tmp[0] );
-    STEP_PART( &x[0], &CNS[ 6], &tmp[0] );
-    STEP_PART( &x[0], &CNS[ 8], &tmp[0] );
-    STEP_PART( &x[0], &CNS[10], &tmp[0] );
-    STEP_PART( &x[0], &CNS[12], &tmp[0] );
-    STEP_PART( &x[0], &CNS[14], &tmp[0] );
+    STEP_PART( &x[0], cns( 0), cns( 1), &tmp[0] );
+    STEP_PART( &x[0], cns( 2), cns( 3), &tmp[0] );
+    STEP_PART( &x[0], cns( 4), cns( 5), &tmp[0] );
+    STEP_PART( &x[0], cns( 6), cns( 7), &tmp[0] );
+    STEP_PART( &x[0], cns( 8), cns( 9), &tmp[0] );
+    STEP_PART( &x[0], cns(10), cns(11), &tmp[0] );
+    STEP_PART( &x[0], cns(12), cns(13), &tmp[0] );
+    STEP_PART( &x[0], cns(14), cns(15), &tmp[0] );

    MIXTON1024( x[0], x[1], x[2], x[3],
                chainv[0], chainv[2], chainv[4],chainv[6],
@@ -380,25 +360,24 @@ void rnd512_2way( luffa_2way_context *state, __m256i *msg )
                chainv[1],chainv[3],chainv[5],chainv[7]);

    /* Process last 256-bit block */
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[16], CNS[17],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(16), cns(17),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[18], CNS[19],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(18), cns(19),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[20], CNS[21],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(20), cns(21),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[22], CNS[23],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(22), cns(23),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[24], CNS[25],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(24), cns(25),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[26], CNS[27],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(26), cns(27),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[28], CNS[29],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(28), cns(29),
                tmp[0], tmp[1] );
-    STEP_PART2( chainv[8], chainv[9], t0, t1, CNS[30], CNS[31],
+    STEP_PART2( chainv[8], chainv[9], t0, t1, cns(30), cns(31),
                tmp[0], tmp[1] );
 }

-
 /***************************************************/
 /* Finalization function  */
 /* state: hash context    */
@@ -410,8 +389,9 @@ void finalization512_2way( luffa_2way_context *state, uint32 *b )
    __m256i* chainv = state->chainv;
    __m256i t[2];
    __m256i zero[2];
-    zero[0] = zero[1] = _mm256_setzero_si256();
-
+    zero[0] = zero[1] = m256_zero;
+    const __m256i shuff_bswap32 = m256_const2_64( 0x0c0d0e0f08090a0b,
+                                                  0x0405060700010203 );
    /*---- blank round with m=0 ----*/
    rnd512_2way( state, zero );

@@ -433,8 +413,10 @@ void finalization512_2way( luffa_2way_context *state, uint32 *b )
    _mm256_store_si256( (__m256i*)&hash[0], t[0] );
    _mm256_store_si256( (__m256i*)&hash[8], t[1] );

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

    rnd512_2way( state, zero );

@@ -455,26 +437,27 @@ void finalization512_2way( luffa_2way_context *state, uint32 *b )
    _mm256_store_si256( (__m256i*)&hash[0], t[0] );
    _mm256_store_si256( (__m256i*)&hash[8], t[1] );

-    casti_m256i( b, 2 ) = mm256_bswap_32( casti_m256i( hash, 0 ) );
-    casti_m256i( b, 3 ) = mm256_bswap_32( casti_m256i( hash, 1 ) );
+    casti_m256i( b, 2 ) = _mm256_shuffle_epi8( 
+                                  casti_m256i( hash, 0 ), shuff_bswap32 );
+    casti_m256i( b, 3 ) = _mm256_shuffle_epi8( 
+                                  casti_m256i( hash, 1 ), shuff_bswap32 );
 }

 int luffa_2way_init( luffa_2way_context *state, int hashbitlen )
 {
-    int i;
    state->hashbitlen = hashbitlen;
-
-    for ( i=0; i<32; i++ ) CNS[i] =
-          _mm256_set_epi32( CNS_INIT[ (i<<2) + 3 ], CNS_INIT[ (i<<2) +2 ],
-                            CNS_INIT[ (i<<2) + 1 ], CNS_INIT[ (i<<2)    ],
-                            CNS_INIT[ (i<<2) + 3 ], CNS_INIT[ (i<<2) +2 ],
-                            CNS_INIT[ (i<<2) + 1 ], CNS_INIT[ (i<<2)    ] );
-
-    for ( i=0; i<10; i++ ) state->chainv[i] =
-          _mm256_set_epi32( IV[ (i<<2) +3 ], IV[ (i<<2) +2 ],
-                            IV[ (i<<2) +1 ], IV[ (i<<2)    ],
-                            IV[ (i<<2) +3 ], IV[ (i<<2) +2 ],
-                            IV[ (i<<2) +1 ], IV[ (i<<2)    ] );
+    __m128i *iv = (__m128i*)IV;
+    
+    state->chainv[0] = m256_const1_128( iv[0] );
+    state->chainv[1] = m256_const1_128( iv[1] );
+    state->chainv[2] = m256_const1_128( iv[2] );
+    state->chainv[3] = m256_const1_128( iv[3] );
+    state->chainv[4] = m256_const1_128( iv[4] );
+    state->chainv[5] = m256_const1_128( iv[5] );
+    state->chainv[6] = m256_const1_128( iv[6] );
+    state->chainv[7] = m256_const1_128( iv[7] );
+    state->chainv[8] = m256_const1_128( iv[8] );
+    state->chainv[9] = m256_const1_128( iv[9] );

    ((__m256i*)state->buffer)[0] = m256_zero;
    ((__m256i*)state->buffer)[1] = m256_zero;
@@ -492,13 +475,15 @@ int luffa_2way_update( luffa_2way_context *state, const void *data,
    __m256i msg[2];
    int i;
    int blocks = (int)len >> 5;
+    const __m256i shuff_bswap32 = m256_const2_64( 0x0c0d0e0f08090a0b,
+                                                  0x0405060700010203 );
    state-> rembytes = (int)len & 0x1F;

    // full blocks
    for ( i = 0; i < blocks; i++, vdata+=2 )
    {
-       msg[0] = mm256_bswap_32( vdata[ 0] );
-       msg[1] = mm256_bswap_32( vdata[ 1 ] );
+       msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
+       msg[1] = _mm256_shuffle_epi8( vdata[ 1 ], shuff_bswap32 );
       rnd512_2way( state, msg );
    }

@@ -507,9 +492,8 @@ int luffa_2way_update( luffa_2way_context *state, const void *data,
    if ( state->rembytes  )
    {
      // remaining data bytes
-      buffer[0] = mm256_bswap_32( vdata[0] );
-      buffer[1] = _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0,
-                                   0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 );
+      buffer[0] = _mm256_shuffle_epi8( vdata[0], shuff_bswap32 );
+      buffer[1] = m256_const2_64( 0, 0x0000000080000000 );
    }
    return 0;
 }
@@ -525,8 +509,7 @@ int luffa_2way_close( luffa_2way_context *state, void *hashval )
      rnd512_2way( state, buffer );
    else
    {     // empty pad block, constant data
-      msg[0] = _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0,
-                                0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 );
+      msg[0] = m256_const2_64( 0, 0x0000000080000000 );
      msg[1] = m256_zero;
      rnd512_2way( state, msg );
    }
@@ -545,13 +528,16 @@ int luffa_2way_update_close( luffa_2way_context *state,
    __m256i msg[2];
    int i;
    const int blocks = (int)( inlen >> 5 );
+    const __m256i shuff_bswap32 = m256_const2_64( 0x0c0d0e0f08090a0b,
+                                                  0x0405060700010203 );
+
    state->rembytes = inlen & 0x1F;

    // full blocks
    for ( i = 0; i < blocks; i++, vdata+=2 )
    {
-       msg[0] = mm256_bswap_32( vdata[ 0 ] );
-       msg[1] = mm256_bswap_32( vdata[ 1 ] );
+       msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
+       msg[1] = _mm256_shuffle_epi8( vdata[ 1 ], shuff_bswap32 );
       rnd512_2way( state, msg );
    }

@@ -559,16 +545,14 @@ int luffa_2way_update_close( luffa_2way_context *state,
    if ( state->rembytes  )
    {
       // padding of partial block
-       msg[0] = mm256_bswap_32( vdata[0] );
-       msg[1] = _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0,
-                                 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 );
+       msg[0] = _mm256_shuffle_epi8( vdata[ 0 ], shuff_bswap32 );
+       msg[1] = m256_const2_64( 0, 0x0000000080000000 );
       rnd512_2way( state, msg );
    }
    else
    {
       // empty pad block
-       msg[0] = _mm256_set_epi8( 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0,
-                                 0,0,0,0, 0,0,0,0, 0,0,0,0, 0x80,0,0,0 );
+       msg[0] = m256_const2_64( 0, 0x0000000080000000 );
       msg[1] = m256_zero;
       rnd512_2way( state, msg );
    }
--- a/algo/luffa/luffa_for_sse2.c
+++ b/algo/luffa/luffa_for_sse2.c
@@ -541,7 +541,9 @@ static void finalization512( hashState_luffa *state, uint32 *b )
    uint32   hash[8] __attribute((aligned(64)));
    __m256i* chainv = (__m256i*)state->chainv;
    __m256i  t;
-    const __m128i zero = _mm_setzero_si128();
+    const __m128i zero = m128_zero;
+    const __m256i shuff_bswap32 = m256_const2_64( 0x0c0d0e0f08090a0b,
+                                                  0x0405060700010203 );

    rnd512( state, zero, zero );

@@ -555,7 +557,9 @@ static void finalization512( hashState_luffa *state, uint32 *b )

    _mm256_store_si256( (__m256i*)hash, t );

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

    rnd512( state, zero, zero );

@@ -568,7 +572,9 @@ static void finalization512( hashState_luffa *state, uint32 *b )

    _mm256_store_si256( (__m256i*)hash, t );

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

 #else
--- a/algo/lyra2/lyra2-gate.c
+++ b/algo/lyra2/lyra2-gate.c
@@ -71,7 +71,7 @@ bool register_lyra2rev3_algo( algo_gate_t* gate )
 #endif
  gate->optimizations = SSE2_OPT | SSE42_OPT | AVX2_OPT;
  gate->miner_thread_init = (void*)&lyra2rev3_thread_init;
-  gate->set_target        = (void*)&alt_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -105,7 +105,7 @@ bool register_lyra2rev2_algo( algo_gate_t* gate )
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | SSE42_OPT | AVX2_OPT;
  gate->miner_thread_init = (void*)&lyra2rev2_thread_init;
-  gate->set_target        = (void*)&alt_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -127,8 +127,7 @@ bool register_lyra2z_algo( algo_gate_t* gate )
  gate->hash       = (void*)&lyra2z_hash;
 #endif
  gate->optimizations = SSE42_OPT | AVX2_OPT;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
-  gate->set_target = (void*)&alt_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -147,15 +146,12 @@ bool register_lyra2h_algo( algo_gate_t* gate )
  gate->hash       = (void*)&lyra2h_hash;
 #endif
  gate->optimizations = SSE42_OPT | AVX2_OPT;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
-  gate->set_target = (void*)&alt_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

 /////////////////////////////////

-int64_t allium_get_max64_0xFFFFLL() { return 0xFFFFLL; }
-
 bool register_allium_algo( algo_gate_t* gate )
 {
 #if defined (ALLIUM_4WAY)
@@ -168,8 +164,7 @@ bool register_allium_algo( algo_gate_t* gate )
  gate->hash      = (void*)&allium_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | SSE42_OPT | AVX2_OPT;
-  gate->set_target        = (void*)&alt_set_target;
-  gate->get_max64         = (void*)&allium_get_max64_0xFFFFLL;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -182,6 +177,7 @@ int phi2_get_work_data_size() { return phi2_use_roots ? 144 : 128; }

 void phi2_decode_extra_data( struct work *work )
 {
+   phi2_use_roots = false;
   if ( work->data[0] & ( 1<<30 ) ) phi2_use_roots = true;
   else for ( int i = 20; i < 36; i++ )
   {
@@ -213,8 +209,7 @@ bool register_phi2_algo( algo_gate_t* gate )
   gate->get_work_data_size = (void*)&phi2_get_work_data_size;
   gate->decode_extra_data  = (void*)&phi2_decode_extra_data;
   gate->build_extraheader  = (void*)&phi2_build_extraheader;
-   gate->set_target         = (void*)&alt_set_target; 
-   gate->get_max64          = (void*)&get_max64_0xffffLL;
+   opt_target_factor = 256.0;
 #if defined(PHI2_4WAY)
   gate->scanhash           = (void*)&scanhash_phi2_4way;
 #else
--- a/algo/lyra2/lyra2re.c
+++ b/algo/lyra2/lyra2re.c
@@ -113,24 +113,13 @@ int scanhash_lyra2re( struct work *work, uint32_t max_nonce,
 	return 0;
 }

-int64_t lyra2re_get_max64 ()
-{
-  return 0xffffLL;
-}
-
-void lyra2re_set_target ( struct work* work, double job_diff )
-{
-   work_set_target(work, job_diff / (128.0 * opt_diff_factor) );
-}
-
 bool register_lyra2re_algo( algo_gate_t* gate )
 {
  init_lyra2re_ctx();
  gate->optimizations = SSE2_OPT | AES_OPT | SSE42_OPT | AVX2_OPT;
  gate->scanhash   = (void*)&scanhash_lyra2re;
  gate->hash       = (void*)&lyra2re_hash;
-  gate->get_max64  = (void*)&lyra2re_get_max64;
-  gate->set_target = (void*)&lyra2re_set_target;
+  opt_target_factor = 128.0;
  return true;
 };

--- a/algo/lyra2/lyra2rev3-4way.c
+++ b/algo/lyra2/lyra2rev3-4way.c
@@ -113,17 +113,18 @@ int scanhash_lyra2rev3_8way( struct work *work, const uint32_t max_nonce,
      lyra2rev3_8way_hash( hash, vdata );
      pdata[19] = n;

-      for ( int lane = 0; lane < 8; lane++ ) if ( hash7[lane] <= Htarg )
+      for ( int lane = 0; lane < 8; lane++ )
+      if ( unlikely( hash7[lane] <= Htarg ) )
      {
         extr_lane_8x32( lane_hash, hash, lane, 256 );
-         if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
+         if ( likely( fulltest( lane_hash, ptarget ) && !opt_benchmark ) )
         {
              pdata[19] = n + lane;
              submit_lane_solution( work, lane_hash, mythr, lane );
         }
      }
      n += 8;
-   } while ( (n < max_nonce-8) && !work_restart[thr_id].restart);
+   } while ( likely( (n < max_nonce-8) && !work_restart[thr_id].restart ) );
   *hashes_done = n - first_nonce + 1;
   return 0;
 }
--- a/algo/lyra2/lyra2z330.c
+++ b/algo/lyra2/lyra2z330.c
@@ -53,11 +53,6 @@ int scanhash_lyra2z330( struct work *work, uint32_t max_nonce,
   return 0;
 }

-void lyra2z330_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool lyra2z330_thread_init()
 {
   const int64_t ROW_LEN_INT64 = BLOCK_LEN_INT64 * 256; // nCols
@@ -75,8 +70,7 @@ bool register_lyra2z330_algo( algo_gate_t* gate )
  gate->miner_thread_init = (void*)&lyra2z330_thread_init;
  gate->scanhash   = (void*)&scanhash_lyra2z330;
  gate->hash       = (void*)&lyra2z330_hash;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
-  gate->set_target = (void*)&lyra2z330_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

--- a/algo/lyra2/sponge.c
+++ b/algo/lyra2/sponge.c
@@ -263,10 +263,9 @@ inline void absorbBlockBlake2Safe( uint64_t *State, const uint64_t *In,
 #if defined (__AVX2__)

  register __m256i state0, state1, state2, state3;
-  const __m256i zero = m256_zero;

-  state0 = zero;
-  state1 = zero;
+  state0 = 
+  state1 = m256_zero;
  state2 = m256_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL,
                          0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
  state3 = m256_const_64( 0x5be0cd19137e2179ULL, 0x1f83d9abfb41bd6bULL,
@@ -290,12 +289,11 @@ inline void absorbBlockBlake2Safe( uint64_t *State, const uint64_t *In,
 #elif defined (__SSE2__)

  __m128i state0, state1, state2, state3, state4, state5, state6, state7;
-  const __m128i zero = m128_zero;

-  state0 = zero;
-  state1 = zero;
-  state2 = zero;
-  state3 = zero;
+  state0 = 
+  state1 =
+  state2 =
+  state3 = m128_zero;
  state4 = m128_const_64( 0xbb67ae8584caa73bULL, 0x6a09e667f3bcc908ULL );
  state5 = m128_const_64( 0xa54ff53a5f1d36f1ULL, 0x3c6ef372fe94f82bULL );
  state6 = m128_const_64( 0x9b05688c2b3e6c1fULL, 0x510e527fade682d1ULL );
--- a/algo/m7m.c
+++ b/algo/m7m.c
@@ -296,8 +296,6 @@ int scanhash_m7m_hash( struct work* work, uint64_t max_nonce,

     pdata[19] = n;

-// can this be skipped after finding a share? Seems to work ok.
-//out:
     mpf_set_prec_raw(magifpi, prec0);
     mpf_set_prec_raw(magifpi0, prec0);
     mpf_set_prec_raw(mptmp, prec0);
@@ -323,9 +321,8 @@ bool register_m7m_algo( algo_gate_t *gate )
  gate->build_stratum_request = (void*)&std_be_build_stratum_request;
  gate->work_decode           = (void*)&std_be_work_decode;
  gate->submit_getwork_result = (void*)&std_be_submit_getwork_result;
-  gate->set_target            = (void*)&scrypt_set_target;
-  gate->get_max64             = (void*)&get_max64_0x1ffff;
  gate->set_work_data_endian  = (void*)&set_work_data_big_endian;
+  opt_target_factor = 65536.0;
  return true;
 }

--- a/algo/nist5/zr5.c
+++ b/algo/nist5/zr5.c
@@ -208,12 +208,6 @@ void zr5_get_new_work( struct work* work, struct work* g_work, int thr_id,
       ++(*nonceptr);
 }

-int64_t zr5_get_max64 ()
-{
-//  return 0x1ffffLL;
-  return 0x1fffffLL;
-}
-
 void zr5_display_pok( struct work* work )
 {
      if ( work->data[0] & 0x00008000 )
@@ -229,7 +223,6 @@ bool register_zr5_algo( algo_gate_t* gate )
    gate->get_new_work          = (void*)&zr5_get_new_work;
    gate->scanhash              = (void*)&scanhash_zr5;
    gate->hash                  = (void*)&zr5hash;
-    gate->get_max64             = (void*)&zr5_get_max64;
    gate->decode_extra_data     = (void*)&zr5_display_pok;
    gate->build_stratum_request = (void*)&std_be_build_stratum_request;
    gate->work_decode           = (void*)&std_be_work_decode;
--- a/algo/quark/hmq1725-gate.c
+++ b/algo/quark/hmq1725-gate.c
@@ -10,8 +10,8 @@ bool register_hmq1725_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_hmq1725;
  gate->hash      = (void*)&hmq1725hash;
 #endif
-  gate->set_target       = (void*)&scrypt_set_target;
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
+  opt_target_factor = 65536.0;
  return true;
 };

--- a/algo/quark/hmq1725.c
+++ b/algo/quark/hmq1725.c
@@ -409,14 +409,3 @@ int scanhash_hmq1725( struct work *work, uint32_t max_nonce,
 	pdata[19] = n;
 	return 0;
 }
-/*
-bool register_hmq1725_algo( algo_gate_t* gate )
-{
-  init_hmq1725_ctx();
-  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target       = (void*)&scrypt_set_target;
-  gate->scanhash         = (void*)&scanhash_hmq1725;
-  gate->hash             = (void*)&hmq1725hash;
-  return true;
-};
-*/
--- a/algo/ripemd/lbry-gate.c
+++ b/algo/ripemd/lbry-gate.c
@@ -41,6 +41,7 @@ void lbry_le_build_stratum_request( char *req, struct work *work,
   free(xnonce2str);
 }

+/*
 void lbry_build_block_header( struct work* g_work, uint32_t version,
                             uint32_t *prevhash, uint32_t *merkle_root,
                             uint32_t ntime, uint32_t nbits )
@@ -63,6 +64,7 @@ void lbry_build_block_header( struct work* g_work, uint32_t version,
   g_work->data[ LBRY_NBITS_INDEX ] = nbits;
   g_work->data[28] = 0x80000000;
 }
+*/

 void lbry_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
 {
@@ -92,13 +94,6 @@ void lbry_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
   g_work->data[28] = 0x80000000;
 }

-void lbry_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
-int64_t lbry_get_max64() { return 0x1ffffLL; }
-
 int lbry_get_work_data_size() { return LBRY_WORK_DATA_SIZE; }

 bool register_lbry_algo( algo_gate_t* gate )
@@ -115,15 +110,14 @@ bool register_lbry_algo( algo_gate_t* gate )
  gate->hash                  = (void*)&lbry_hash;
 #endif
  gate->calc_network_diff     = (void*)&lbry_calc_network_diff;
-  gate->get_max64             = (void*)&lbry_get_max64;
  gate->build_stratum_request = (void*)&lbry_le_build_stratum_request;
 //  gate->build_block_header    = (void*)&build_block_header;
  gate->build_extraheader     = (void*)&lbry_build_extraheader;
-  gate->set_target            = (void*)&lbry_set_target;
  gate->ntime_index           = LBRY_NTIME_INDEX;
  gate->nbits_index           = LBRY_NBITS_INDEX;
  gate->nonce_index           = LBRY_NONCE_INDEX;
  gate->get_work_data_size    = (void*)&lbry_get_work_data_size;
+  opt_target_factor = 256.0;
  return true;
 }

--- a/algo/scrypt/neoscrypt.c
+++ b/algo/scrypt/neoscrypt.c
@@ -1070,17 +1070,6 @@ int scanhash_neoscrypt( struct work *work,
    return 0;
 }

-int64_t get_neoscrypt_max64() { return 0x3ffff; }
-
-void neoscrypt_wait_for_diff( struct stratum_ctx *stratum )
-{
-   while ( !stratum->job.diff )
-   {
-//      applog(LOG_DEBUG, "Waiting for Stratum to set the job difficulty");
-      sleep(1);
-   }
-}
-
 int neoscrypt_get_work_data_size () { return 80; }

 bool register_neoscrypt_algo( algo_gate_t* gate )
@@ -1088,14 +1077,12 @@ bool register_neoscrypt_algo( algo_gate_t* gate )
  gate->optimizations         = SSE2_OPT;
  gate->scanhash              = (void*)&scanhash_neoscrypt;
  gate->hash                  = (void*)&neoscrypt;
-  gate->get_max64             = (void*)&get_neoscrypt_max64;
-  gate->set_target            = (void*)&scrypt_set_target;
-  gate->wait_for_diff         = (void*)&neoscrypt_wait_for_diff;
  gate->build_stratum_request = (void*)&std_be_build_stratum_request;
  gate->work_decode           = (void*)&std_be_work_decode;
  gate->submit_getwork_result = (void*)&std_be_submit_getwork_result;
  gate->set_work_data_endian  = (void*)&set_work_data_big_endian;
  gate->get_work_data_size    = (void*)&neoscrypt_get_work_data_size;
+  opt_target_factor = 65536.0;
  return true;
 };

--- a/algo/scrypt/pluck.c
+++ b/algo/scrypt/pluck.c
@@ -483,11 +483,6 @@ int scanhash_pluck( struct work *work, uint32_t max_nonce,
 	return 0;
 }

-int64_t pluck_get_max64 ()
-{
-  return 0x1ffLL;
-}
-
 bool pluck_miner_thread_init( int thr_id )
 { 
  scratchbuf = malloc( 128 * 1024 ); 
@@ -503,8 +498,7 @@ bool register_pluck_algo( algo_gate_t* gate )
  gate->miner_thread_init = (void*)&pluck_miner_thread_init;
  gate->scanhash         = (void*)&scanhash_pluck;
  gate->hash             = (void*)&pluck_hash;
-  gate->set_target       = (void*)&scrypt_set_target;
-  gate->get_max64        = (void*)&pluck_get_max64;
+  opt_target_factor = 65536.0;
  return true;
 };

--- a/algo/scrypt/scrypt.c
+++ b/algo/scrypt/scrypt.c
@@ -698,8 +698,8 @@ static void scrypt_1024_1_1_256_24way(const uint32_t *input,
 extern int scanhash_scrypt( struct work *work, uint32_t max_nonce,
                            uint64_t *hashes_done, struct thr_info *mythr )
 {
-        uint32_t *pdata = work->data;
-        uint32_t *ptarget = work->target;
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
 	uint32_t data[SCRYPT_MAX_WAYS * 20], hash[SCRYPT_MAX_WAYS * 8];
 	uint32_t midstate[8];
 	uint32_t n = pdata[19] - 1;
@@ -766,8 +766,6 @@ extern int scanhash_scrypt( struct work *work, uint32_t max_nonce,
 	return 0;
 }

-int64_t scrypt_get_max64() { return 0xfff; }
-
 bool scrypt_miner_thread_init( int thr_id )
 {
 scratchbuf = scrypt_buffer_alloc( scratchbuf_size );  
@@ -783,13 +781,16 @@ bool register_scrypt_algo( algo_gate_t* gate )
  gate->miner_thread_init =(void*)&scrypt_miner_thread_init;
  gate->scanhash         = (void*)&scanhash_scrypt;
 //  gate->hash             = (void*)&scrypt_1024_1_1_256_24way;
-  gate->set_target       = (void*)&scrypt_set_target;
-  gate->get_max64        = (void*)&scrypt_get_max64;
+  opt_target_factor = 65536.0;

-  if ( !opt_scrypt_n )
+  if ( !opt_param_n )
+  {
+     opt_param_n = 1024;
     scratchbuf_size = 1024;
+  }
  else
-     scratchbuf_size = opt_scrypt_n;
+     scratchbuf_size = opt_param_n;
+  applog(LOG_INFO,"Scrypt paramaters: N= %d, R= 1.", opt_param_n );
  return true;
 };

--- a/algo/scryptjane/scrypt-jane-chacha.h
+++ b/algo/scryptjane/scrypt-jane-chacha.h
@@ -55,6 +55,7 @@ typedef uint32_t scrypt_mix_word_t;
 	#include "scrypt-jane-romix-template.h"
 #endif

+
 /* cpu agnostic */
 #define SCRYPT_ROMIX_FN scrypt_ROMix_basic
 #define SCRYPT_MIX_FN chacha_core_basic
--- a/algo/scryptjane/scrypt-jane-romix-template.h
+++ b/algo/scryptjane/scrypt-jane-romix-template.h
@@ -1,9 +1,11 @@
 #if !defined(SCRYPT_CHOOSE_COMPILETIME) || !defined(SCRYPT_HAVE_ROMIX)

+/*
 #if defined(SCRYPT_CHOOSE_COMPILETIME)
 #undef SCRYPT_ROMIX_FN
 #define SCRYPT_ROMIX_FN scrypt_ROMix
 #endif
+*/

 #undef SCRYPT_HAVE_ROMIX
 #define SCRYPT_HAVE_ROMIX
--- a/algo/scryptjane/scrypt-jane.c
+++ b/algo/scryptjane/scrypt-jane.c
@@ -240,24 +240,23 @@ bool register_scryptjane_algo( algo_gate_t* gate )
 {
    gate->scanhash   = (void*)&scanhash_scryptjane;
    gate->hash       = (void*)&scryptjanehash;
-    gate->set_target = (void*)&scrypt_set_target;
-    gate->get_max64  = (void*)&get_max64_0x40LL;
+    opt_target_factor = 65536.0;

    // figure out if arg in N or Nfactor
-    if ( !opt_scrypt_n )
+    if ( !opt_param_n )
    {
      applog( LOG_ERR, "The N factor must be specified in the form algo:nf");
      return false;
    }
-    else if ( opt_scrypt_n < 32 )
+    else if ( opt_param_n < 32 )
    {
      // arg is Nfactor, calculate N
-      sj_N = 1 << ( opt_scrypt_n + 1 );
+      sj_N = 1 << ( opt_param_n + 1 );
    }
    else
    {
      // arg is N
-      sj_N = opt_scrypt_n;
+      sj_N = opt_param_n;
    }
    return true;
 }
--- a/algo/sha/sha256-hash-4way.c
+++ b/algo/sha/sha256-hash-4way.c
@@ -305,9 +305,11 @@ void sha256_4way_close( sha256_4way_context *sc, void *dst )
    low = low << 3;

    sc->buf[ pad >> 2 ] =
-                 mm128_bswap_32( _mm_set1_epi32( high ) );
+                 mm128_bswap_32( m128_const1_32( high ) );
+//                 mm128_bswap_32( _mm_set1_epi32( high ) );
    sc->buf[ ( pad+4 ) >> 2 ] =
-                 mm128_bswap_32( _mm_set1_epi32( low ) );
+                 mm128_bswap_32( m128_const1_32( low ) );
+//                 mm128_bswap_32( _mm_set1_epi32( low ) );
    sha256_4way_round( sc, sc->buf, sc->val );

    mm128_block_bswap_32( dst, sc->val );
@@ -538,9 +540,9 @@ void sha256_8way_close( sha256_8way_context *sc, void *dst )
    low = low << 3;

    sc->buf[ pad >> 2 ] =
-                 mm256_bswap_32( _mm256_set1_epi32( high ) );
+                 mm256_bswap_32( m256_const1_32( high ) );
    sc->buf[ ( pad+4 ) >> 2 ] =
-                 mm256_bswap_32( _mm256_set1_epi32( low ) );
+                 mm256_bswap_32( m256_const1_32( low ) );

    sha256_8way_round( sc, sc->buf, sc->val );

--- a/algo/sha/sha256t-gate.c
+++ b/algo/sha/sha256t-gate.c
@@ -15,7 +15,6 @@ bool register_sha256t_algo( algo_gate_t* gate )
    gate->scanhash   = (void*)&scanhash_sha256t;
    gate->hash       = (void*)&sha256t_hash;
 #endif
-    gate->get_max64  = (void*)&get_max64_0x3ffff;
    return true;
 }

@@ -34,7 +33,6 @@ bool register_sha256q_algo( algo_gate_t* gate )
    gate->scanhash   = (void*)&scanhash_sha256q;
    gate->hash       = (void*)&sha256q_hash;
 #endif
-    gate->get_max64  = (void*)&get_max64_0x3ffff;
    return true;

 }
--- a/algo/sha/sha512-hash-4way.c
+++ b/algo/sha/sha512-hash-4way.c
@@ -252,16 +252,6 @@ void sha512_4way_init( sha512_4way_context *sc )
 {
   sc->initialized = false;
   sc->count = 0;
-/*
-   sc->val[0] = _mm256_set1_epi64x( H512[0] );
-   sc->val[1] = _mm256_set1_epi64x( H512[1] );
-   sc->val[2] = _mm256_set1_epi64x( H512[2] );
-   sc->val[3] = _mm256_set1_epi64x( H512[3] );
-   sc->val[4] = _mm256_set1_epi64x( H512[4] );
-   sc->val[5] = _mm256_set1_epi64x( H512[5] );
-   sc->val[6] = _mm256_set1_epi64x( H512[6] );
-   sc->val[7] = _mm256_set1_epi64x( H512[7] );
-*/
 }

 void sha512_4way( sha512_4way_context *sc, const void *data, size_t len )
@@ -295,6 +285,8 @@ void sha512_4way_close( sha512_4way_context *sc, void *dst )
    unsigned ptr;
    const int buf_size = 128;
    const int pad = buf_size - 16;
+    const __m256i shuff_bswap64 = m256_const2_64( 0x08090a0b0c0d0e0f,
+                                                  0x0001020304050607 );

    ptr = (unsigned)sc->count & (buf_size - 1U);
    sc->buf[ ptr>>3 ] = m256_const1_64( 0x80 );
@@ -308,10 +300,10 @@ void sha512_4way_close( sha512_4way_context *sc, void *dst )
    else
         memset_zero_256( sc->buf + (ptr>>3), (pad - ptr) >> 3 );

-    sc->buf[ pad >> 3 ] =
-                 mm256_bswap_64( _mm256_set1_epi64x( sc->count >> 61 ) );
-    sc->buf[ ( pad+8 ) >> 3 ] = 
-                 mm256_bswap_64( _mm256_set1_epi64x( sc->count << 3 ) );
+    sc->buf[ pad >> 3 ] = _mm256_shuffle_epi8(
+                       _mm256_set1_epi64x( sc->count >> 61 ), shuff_bswap64 );
+    sc->buf[ ( pad+8 ) >> 3 ] = _mm256_shuffle_epi8( 
+                       _mm256_set1_epi64x( sc->count <<  3 ), shuff_bswap64 );
    sha512_4way_round( sc, sc->buf, sc->val );

    mm256_block_bswap_64( dst, sc->val );
--- a/algo/shavite/shavite-hash-2way.c
+++ b/algo/shavite/shavite-hash-2way.c
@@ -5,6 +5,7 @@

 #if defined(__AVX2__)

+
 static const uint32_t IV512[] =
 {
        0x72FCCDD8, 0x79CA4727, 0x128A077B, 0x40D55AEC,
@@ -13,6 +14,7 @@ static const uint32_t IV512[] =
        0xE275EADE, 0x502D9FCD, 0xB9357178, 0x022A4B9A
 };

+
 #define mm256_ror2x256hi_1x32( a, b ) \
   _mm256_blend_epi32( mm256_ror1x32_128( a ), \
                       mm256_ror1x32_128( b ), 0x88 )
@@ -232,18 +234,14 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )

 void shavite512_2way_init( shavite512_2way_context *ctx )
 {
-   casti_m256i( ctx->h, 0 ) =
-            _mm256_set_epi32( IV512[ 3], IV512[ 2], IV512[ 1], IV512[ 0],
-                              IV512[ 3], IV512[ 2], IV512[ 1], IV512[ 0] );  
-   casti_m256i( ctx->h, 1 ) =
-            _mm256_set_epi32( IV512[ 7], IV512[ 6], IV512[ 5], IV512[ 4],
-                              IV512[ 7], IV512[ 6], IV512[ 5], IV512[ 4] );
-   casti_m256i( ctx->h, 2 ) =
-            _mm256_set_epi32( IV512[11], IV512[10], IV512[ 9], IV512[ 8],
-                              IV512[11], IV512[10], IV512[ 9], IV512[ 8] );
-   casti_m256i( ctx->h, 3 ) =
-            _mm256_set_epi32( IV512[15], IV512[14], IV512[13], IV512[12],
-                              IV512[15], IV512[14], IV512[13], IV512[12] );
+    __m256i *h = (__m256i*)ctx->h;
+    __m128i *iv = (__m128i*)IV512;
+   
+   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] );
+
   ctx->ptr    = 0;
   ctx->count0 = 0;
   ctx->count1 = 0;
@@ -251,6 +249,7 @@ void shavite512_2way_init( shavite512_2way_context *ctx )
   ctx->count3 = 0;
 }

+// not tested, use update_close
 void shavite512_2way_update( shavite512_2way_context *ctx, const void *data,
                             size_t len )
 {
@@ -287,6 +286,7 @@ void shavite512_2way_update( shavite512_2way_context *ctx, const void *data,
   ctx->ptr = ptr;
 }

+// not tested
 void shavite512_2way_close( shavite512_2way_context *ctx, void *dst )
 {
    unsigned char *buf;
@@ -300,7 +300,7 @@ void shavite512_2way_close( shavite512_2way_context *ctx, void *dst )
    uint32_t vp = ctx->ptr>>5;

    // Terminating byte then zero pad
-    casti_m256i( buf, vp++ ) = _mm256_set_epi32( 0,0,0,0x80, 0,0,0,0x80 );
+    casti_m256i( buf, vp++ ) = m256_const2_64( 0, 0x0000000000000080 );

    // Zero pad full vectors up to count
    for ( ; vp < 6; vp++ )      
@@ -314,14 +314,12 @@ void shavite512_2way_close( shavite512_2way_context *ctx, void *dst )
    count.u32[2] = ctx->count2;
    count.u32[3] = ctx->count3;

-    casti_m256i( buf, 6 ) = _mm256_set_epi16( count.u16[0], 0,0,0,0,0,0,0,
-		                              count.u16[0], 0,0,0,0,0,0,0 );
-    casti_m256i( buf, 7 ) = _mm256_set_epi16(
-		    0x0200      , count.u16[7], count.u16[6], count.u16[5],
-		    count.u16[4], count.u16[3], count.u16[2], count.u16[1],
-                    0x0200      , count.u16[7], count.u16[6], count.u16[5],
-                    count.u16[4], count.u16[3], count.u16[2], count.u16[1] );
-
+    casti_m256i( buf, 6 ) = m256_const1_128(
+                  _mm_insert_epi16( m128_zero, count.u16[0], 7 ) ); 
+    casti_m256i( buf, 7 ) = m256_const1_128( _mm_set_epi16(
+                  0x0200,       count.u16[7], count.u16[6], count.u16[5],
+                  count.u16[4], count.u16[3], count.u16[2], count.u16[1] ) );
+                
    c512_2way( ctx, buf);

    casti_m256i( dst, 0 ) = casti_m256i( ctx->h, 0 );
@@ -382,23 +380,21 @@ void shavite512_2way_update_close( shavite512_2way_context *ctx, void *dst,

   if ( vp == 0 )    // empty buf, xevan.
   { 
-      casti_m256i( buf, 0 ) = _mm256_set_epi32( 0,0,0,0x80, 0,0,0,0x80 );
+      casti_m256i( buf, 0 ) = m256_const2_64( 0, 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + 1, 5 );
      ctx->count0 = ctx->count1 = ctx->count2 = ctx->count3 = 0;
   }
   else     // half full buf, everyone else.
   {
-      casti_m256i( buf, vp++ ) = _mm256_set_epi32( 0,0,0,0x80, 0,0,0,0x80 );
+    casti_m256i( buf, vp++ ) = m256_const2_64( 0, 0x0000000000000080 );
      memset_zero_256( (__m256i*)buf + vp, 6 - vp );
   }

-   casti_m256i( buf, 6 ) = _mm256_set_epi16( count.u16[0], 0,0,0,0,0,0,0,
-                                             count.u16[0], 0,0,0,0,0,0,0 );
-   casti_m256i( buf, 7 ) = _mm256_set_epi16(
-                   0x0200      , count.u16[7], count.u16[6], count.u16[5],
-                   count.u16[4], count.u16[3], count.u16[2], count.u16[1],
-                   0x0200      , count.u16[7], count.u16[6], count.u16[5],
-                   count.u16[4], count.u16[3], count.u16[2], count.u16[1] );
+    casti_m256i( buf, 6 ) = m256_const1_128(
+                  _mm_insert_epi16( m128_zero, count.u16[0], 7 ) ); 
+    casti_m256i( buf, 7 ) = m256_const1_128( _mm_set_epi16(
+                  0x0200,       count.u16[7], count.u16[6], count.u16[5],
+                  count.u16[4], count.u16[3], count.u16[2], count.u16[1] ) );

   c512_2way( ctx, buf);

--- a/algo/simd/simd-hash-2way.c
+++ b/algo/simd/simd-hash-2way.c
@@ -110,14 +110,26 @@ static const m256_v16 FFT256_Twiddle[] =

 // imported from vector.c

+#define REDUCE(x) \
+  _mm256_sub_epi16( _mm256_and_si256( x, m256_const1_64( \
+                         0x00ff00ff00ff00ff ) ), _mm256_srai_epi16( x, 8 ) )
+/*
 #define REDUCE(x) \
  _mm256_sub_epi16( _mm256_and_si256( x, _mm256_set1_epi16( 255 ) ), \
                                         _mm256_srai_epi16( x, 8 ) )
+*/

+#define EXTRA_REDUCE_S(x)\
+  _mm256_sub_epi16( x, _mm256_and_si256( \
+             m256_const1_64( 0x0101010101010101 ), \
+             _mm256_cmpgt_epi16( x, m256_const1_64( 0x0080008000800080 ) ) ) )
+
+/*
 #define EXTRA_REDUCE_S(x)\
  _mm256_sub_epi16( x, \
         _mm256_and_si256( _mm256_set1_epi16( 257 ), \
                           _mm256_cmpgt_epi16( x, _mm256_set1_epi16( 128 ) ) ) )
+*/

 #define REDUCE_FULL_S( x )  EXTRA_REDUCE_S( REDUCE (x ) )

--- a/algo/skein/skein-gate.c
+++ b/algo/skein/skein-gate.c
@@ -2,8 +2,6 @@
 #include "sph_skein.h"
 #include "skein-hash-4way.h"

-int64_t skein_get_max64() { return 0x7ffffLL; }
-
 bool register_skein_algo( algo_gate_t* gate )
 {
    gate->optimizations = AVX2_OPT | SHA_OPT;
@@ -14,7 +12,6 @@ bool register_skein_algo( algo_gate_t* gate )
    gate->scanhash  = (void*)&scanhash_skein;
    gate->hash      = (void*)&skeinhash;
 #endif
-    gate->get_max64 = (void*)&skein_get_max64;
    return true;
 };

--- a/algo/skein/skein2-gate.c
+++ b/algo/skein/skein2-gate.c
@@ -2,11 +2,6 @@
 #include <stdint.h>
 #include "sph_skein.h"

-int64_t skein2_get_max64 ()
-{
-  return 0x7ffffLL;
-}
-
 bool register_skein2_algo( algo_gate_t* gate )
 {
  gate->optimizations = AVX2_OPT;
@@ -17,7 +12,6 @@ bool register_skein2_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_skein2;
  gate->hash      = (void*)&skein2hash;
 #endif
-  gate->get_max64 = (void*)&skein2_get_max64;
  return true;
 };

--- a/algo/sm3/sm3-hash-4way.c
+++ b/algo/sm3/sm3-hash-4way.c
@@ -181,7 +181,7 @@ void sm3_4way_compress( __m128i *digest, __m128i *block )
   for( j =0; j < 16; j++ )
   {
      SS1 = mm128_rol_32( _mm_add_epi32( _mm_add_epi32( mm128_rol_32(A,12), E ),
-                                      mm128_rol_32( T, j ) ), 7 );
+                                      mm128_rol_var_32( T, j ) ), 7 );
      SS2 = _mm_xor_si128( SS1, mm128_rol_32( A, 12 ) );
      TT1 = _mm_add_epi32( _mm_add_epi32( _mm_add_epi32( FF0( A, B, C ), D ),
                                          SS2 ), W1[j] );
@@ -201,7 +201,7 @@ void sm3_4way_compress( __m128i *digest, __m128i *block )
   for( j =16; j < 64; j++ )
   {
      SS1 = mm128_rol_32( _mm_add_epi32( _mm_add_epi32( mm128_rol_32(A,12), E ),
-                                      mm128_rol_32( T, j&31 ) ), 7 );
+                                      mm128_rol_var_32( T, j&31 ) ), 7 );
      SS2 = _mm_xor_si128( SS1, mm128_rol_32( A, 12 ) );
      TT1 = _mm_add_epi32( _mm_add_epi32( _mm_add_epi32( FF1( A, B, C ), D ), 
                                          SS2 ), W1[j] );
--- a/algo/x11/c11-gate.c
+++ b/algo/x11/c11-gate.c
@@ -12,7 +12,6 @@ bool register_c11_algo( algo_gate_t* gate )
  gate->hash      = (void*)&c11_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x11/fresh.c
+++ b/algo/x11/fresh.c
@@ -120,19 +120,12 @@ int scanhash_fresh( struct work *work,
 	return 0;
 }

-void fresh_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
-
 bool register_fresh_algo( algo_gate_t* gate )
 {
    algo_not_tested();
    gate->scanhash   = (void*)&scanhash_fresh;
    gate->hash       = (void*)&freshhash;
-    gate->set_target = (void*)&fresh_set_target;
-    gate->get_max64  = (void*)&get_max64_0x3ffff;
+    opt_target_factor = 256.0;
    return true;
 };

--- a/algo/x11/timetravel-gate.c
+++ b/algo/x11/timetravel-gate.c
@@ -1,10 +1,5 @@
 #include "timetravel-gate.h"

-void tt8_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool register_timetravel_algo( algo_gate_t* gate )
 {
 #ifdef TIMETRAVEL_4WAY
@@ -16,9 +11,8 @@ bool register_timetravel_algo( algo_gate_t* gate )
  gate->scanhash   = (void*)&scanhash_timetravel;
  gate->hash       = (void*)&timetravel_hash;
 #endif
-  gate->set_target = (void*)&tt8_set_target;
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
+  opt_target_factor = 256.0;
  return true;
 };

--- a/algo/x11/timetravel10-gate.c
+++ b/algo/x11/timetravel10-gate.c
@@ -1,10 +1,5 @@
 #include "timetravel10-gate.h"

-void tt10_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool register_timetravel10_algo( algo_gate_t* gate )
 {
 #ifdef TIMETRAVEL10_4WAY
@@ -16,9 +11,8 @@ bool register_timetravel10_algo( algo_gate_t* gate )
  gate->scanhash   = (void*)&scanhash_timetravel10;
  gate->hash       = (void*)&timetravel10_hash;
 #endif
-  gate->set_target = (void*)&tt10_set_target;
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
+  opt_target_factor = 256.0;
  return true;
 };

--- a/algo/x11/tribus-gate.c
+++ b/algo/x11/tribus-gate.c
@@ -3,7 +3,6 @@
 bool register_tribus_algo( algo_gate_t* gate )
 {
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64     = (void*)&get_max64_0x1ffff;
 #if defined (TRIBUS_4WAY)
 //  init_tribus_4way_ctx();
  gate->scanhash      = (void*)&scanhash_tribus_4way;
--- a/algo/x11/x11-gate.c
+++ b/algo/x11/x11-gate.c
@@ -12,7 +12,6 @@ bool register_x11_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x11_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x11/x11gost-gate.c
+++ b/algo/x11/x11gost-gate.c
@@ -12,7 +12,6 @@ bool register_x11gost_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x11gost_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x12/x12-gate.c
+++ b/algo/x12/x12-gate.c
@@ -12,7 +12,6 @@ bool register_x12_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x12hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x13/drop.c
+++ b/algo/x13/drop.c
@@ -249,7 +249,6 @@ bool register_drop_algo( algo_gate_t* gate )
    gate->scanhash              = (void*)&scanhash_drop;
    gate->hash                  = (void*)&droplp_hash_pok;
    gate->get_new_work          = (void*)&drop_get_new_work;
-    gate->set_target            = (void*)&scrypt_set_target;
    gate->build_stratum_request = (void*)&std_be_build_stratum_request;
    gate->work_decode           = (void*)&std_be_work_decode;
    gate->submit_getwork_result = (void*)&std_be_submit_getwork_result;
@@ -257,6 +256,7 @@ bool register_drop_algo( algo_gate_t* gate )
    gate->decode_extra_data     = (void*)&drop_display_pok;
    gate->get_work_data_size    = (void*)&drop_get_work_data_size;
    gate->work_cmp_size         = 72;
+    opt_target_factor = 65536.0;
    return true;
 };

--- a/algo/x13/phi1612-gate.c
+++ b/algo/x13/phi1612-gate.c
@@ -12,7 +12,6 @@ bool register_phi1612_algo( algo_gate_t* gate )
  gate->hash      = (void*)&phi1612_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x13/x13-gate.c
+++ b/algo/x13/x13-gate.c
@@ -12,7 +12,6 @@ bool register_x13_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x13hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x13/x13bcd-4way.c
+++ b/algo/x13/x13bcd-4way.c
@@ -12,7 +12,6 @@
 #include "algo/skein/skein-hash-4way.h"
 #include "algo/jh/jh-hash-4way.h"
 #include "algo/keccak/keccak-hash-4way.h"
-//#include "algo/luffa/luffa-hash-2way.h"
 #include "algo/cubehash/cubehash_sse2.h"
 #include "algo/shavite/sph_shavite.h"
 #include "algo/simd/simd-hash-2way.h"
@@ -28,7 +27,6 @@ typedef struct {
    skein512_4way_context   skein;
    jh512_4way_context      jh;
    keccak512_4way_context  keccak;
-//    luffa_2way_context      luffa;
    cubehashParam           cube;
    sph_shavite512_context  shavite;
    simd_2way_context       simd;
@@ -49,7 +47,6 @@ void init_x13bcd_4way_ctx()
     skein512_4way_init( &x13bcd_4way_ctx.skein );
     jh512_4way_init( &x13bcd_4way_ctx.jh );
     keccak512_4way_init( &x13bcd_4way_ctx.keccak );
-//     luffa_2way_init( &x13bcd_4way_ctx.luffa, 512 );
     cubehashInit( &x13bcd_4way_ctx.cube, 512, 16, 32 );
     sph_shavite512_init( &x13bcd_4way_ctx.shavite );
     simd_2way_init( &x13bcd_4way_ctx.simd, 512 );
@@ -72,8 +69,6 @@ void x13bcd_4way_hash( void *state, const void *input )
     // Blake
     memcpy( &ctx.blake, &x13bcd_ctx_mid, sizeof(x13bcd_ctx_mid) );
     blake512_4way( &ctx.blake, input + (64<<2), 16 );
-
-//     blake512_4way( &ctx.blake, input, 80 );
     blake512_4way_close( &ctx.blake, vhash );

     // Bmw
@@ -127,17 +122,6 @@ void x13bcd_4way_hash( void *state, const void *input )
     sm3_4way_close( &ctx.sm3, sm3_vhash );
     dintrlv_4x32( hash0, hash1, hash2, hash3, sm3_vhash, 512 );

-/*     
-     // Luffa
-     intrlv_2x128( vhash, hash0, hash1, 512 );
-     luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
-     dintrlv_2x128( hash0, hash1, vhash, 512 );
-     intrlv_2x128( vhash, hash2, hash3, 512 );
-     luffa_2way_init( &ctx.luffa, 512 );
-     luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
-     dintrlv_2x128( hash2, hash3, vhash, 512 );
-*/
-     
     // Cubehash
     cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
     memcpy( &ctx.cube, &x13bcd_4way_ctx.cube, sizeof(cubehashParam) );
@@ -185,26 +169,6 @@ void x13bcd_4way_hash( void *state, const void *input )
     update_final_echo( &ctx.echo, (BitSequence *)hash3,
                       (const BitSequence *) hash3, 512 );

-/*
-     intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
-
-     // SM3 parallel 32 bit
-     uint32_t sm3_vhash[32*4] __attribute__ ((aligned (64)));
-     memset( sm3_vhash, 0, sizeof sm3_vhash );
-     uint32_t sm3_hash0[32] __attribute__ ((aligned (32)));
-     memset( sm3_hash0, 0, sizeof sm3_hash0 );
-     uint32_t sm3_hash1[32] __attribute__ ((aligned (32)));
-     memset( sm3_hash1, 0, sizeof sm3_hash1 );
-     uint32_t sm3_hash2[32] __attribute__ ((aligned (32)));
-     memset( sm3_hash2, 0, sizeof sm3_hash2 );
-     uint32_t sm3_hash3[32] __attribute__ ((aligned (32)));
-     memset( sm3_hash3, 0, sizeof sm3_hash3 );
-
-     sm3_4way( &ctx.sm3, vhash, 64 );
-     sm3_4way_close( &ctx.sm3, sm3_vhash );
-     dintrlv_4x32( hash0, hash1, hash2, hash3, sm3_vhash, 512 );
-*/
-
     // Hamsi parallel 4x32x2
     intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
     hamsi512_4way( &ctx.hamsi, vhash, 64 );
--- a/algo/x13/x13sm3-gate.c
+++ b/algo/x13/x13sm3-gate.c
@@ -12,7 +12,6 @@ bool register_x13sm3_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x13sm3_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

@@ -28,7 +27,6 @@ bool register_x13bcd_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x13bcd_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x14/axiom.c
+++ b/algo/x14/axiom.c
@@ -83,6 +83,5 @@ bool register_axiom_algo( algo_gate_t* gate )
 {
    gate->scanhash  = (void*)&scanhash_axiom;
    gate->hash      = (void*)&axiomhash;
-    gate->get_max64 = (void*)&get_max64_0x40LL;
    return true;
 }
--- a/algo/x14/polytimos-gate.c
+++ b/algo/x14/polytimos-gate.c
@@ -11,7 +11,6 @@ bool register_polytimos_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_polytimos;
  gate->hash      = (void*)&polytimos_hash;
 #endif
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x14/veltor-gate.c
+++ b/algo/x14/veltor-gate.c
@@ -12,7 +12,6 @@ bool register_veltor_algo( algo_gate_t* gate )
  gate->hash      = (void*)&veltor_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x14/x14-gate.c
+++ b/algo/x14/x14-gate.c
@@ -12,7 +12,6 @@ bool register_x14_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x14hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->get_max64 = (void*)&get_max64_0x3ffff;
  return true;
 };

--- a/algo/x16/x16r-4way.c
+++ b/algo/x16/x16r-4way.c
@@ -68,21 +68,7 @@ void x16r_4way_hash( void* output, const void* input )
   int size = 80;

   dintrlv_4x64( hash0, hash1, hash2, hash3, input, 640 );
-/* 
-   if ( s_ntime == UINT32_MAX )
-   {
-      const uint8_t* tmp = (uint8_t*) in0;
-      x16_r_s_getAlgoString( &tmp[4], hashOrder );
-   }
-*/
-   // Input data is both 64 bit interleaved (input)
-   // and deinterleaved in inp0-3.
-   // If First function uses 64 bit data it is not required to interleave inp
-   // first. It may use the inerleaved data dmost convenient, ie 4way 64 bit.
-   // All other functions assume data is deinterleaved in hash0-3
-   // All functions must exit with data deinterleaved in hash0-3.
-   // Alias in0-3 points to either inp0-3 or hash0-3 according to
-   // its hashOrder position. Size is also set accordingly.
+
   for ( int i = 0; i < 16; i++ )
   {
      const char elem = hashOrder[i];
@@ -289,34 +275,31 @@ int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
 {
   uint32_t hash[4*16] __attribute__ ((aligned (64)));
   uint32_t vdata[24*4] __attribute__ ((aligned (64)));
-   uint32_t endiandata[20] __attribute__((aligned(64)));
+   uint32_t bedata1[2] __attribute__((aligned(64)));
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
   const uint32_t Htarg = ptarget[7];
   const uint32_t first_nonce = pdata[19];
   uint32_t n = first_nonce;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
    __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
+   int thr_id = mythr->id;
   volatile uint8_t *restart = &(work_restart[thr_id].restart);

-   casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
-   casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
-   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
-
-   if ( s_ntime != endiandata[17] )
-   {
-      uint32_t ntime = swab32(pdata[17]);
-      x16_r_s_getAlgoString( (const uint8_t*) (&endiandata[1]), hashOrder );
-      s_ntime = ntime;
-      if ( opt_debug && !thr_id )
-              applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
-   }
-
   if ( opt_benchmark )
      ptarget[7] = 0x0cff;

-   uint64_t *edata = (uint64_t*)endiandata;
-   intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+
+   bedata1[0] = bswap_32( pdata[1] );
+   bedata1[1] = bswap_32( pdata[2] );
+   const uint32_t ntime = bswap_32( pdata[17] );
+   if ( s_ntime != ntime )
+   {
+      x16_r_s_getAlgoString( (const uint8_t*)bedata1, hashOrder );
+      s_ntime = ntime;
+      if ( opt_debug && !thr_id )
+              applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
+   }

   do
   {
@@ -326,14 +309,15 @@ int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
      x16r_4way_hash( hash, vdata );
      pdata[19] = n;

-      for ( int i = 0; i < 4; i++ )  if ( (hash+(i<<3))[7] <= Htarg )
-      if( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
+      for ( int i = 0; i < 4; i++ )
+      if ( unlikely( (hash+(i<<3))[7] <= Htarg ) )
+      if( likely( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark ) )
      {
         pdata[19] = n+i;
         submit_lane_solution( work, hash+(i<<3), mythr, i );
      }
      n += 4;
-   } while ( (  n < max_nonce ) && !(*restart) );
+   } while ( likely( ( n < max_nonce ) && !(*restart) ) );

   *hashes_done = n - first_nonce + 1;
   return 0;
--- a/algo/x16/x16r-gate.c
+++ b/algo/x16/x16r-gate.c
@@ -42,8 +42,23 @@ bool register_x16r_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x16r_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target = (void*)&alt_set_target;
  x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
+  opt_target_factor = 256.0;
+  return true;
+};
+
+bool register_x16rv2_algo( algo_gate_t* gate )
+{
+#if defined (X16R_4WAY)
+  gate->scanhash  = (void*)&scanhash_x16rv2_4way;
+  gate->hash      = (void*)&x16rv2_4way_hash;
+#else
+  gate->scanhash  = (void*)&scanhash_x16rv2;
+  gate->hash      = (void*)&x16rv2_hash;
+#endif
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
+  x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -57,8 +72,8 @@ bool register_x16s_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x16r_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target = (void*)&alt_set_target;
  x16_r_s_getAlgoString = (void*)&x16s_getAlgoString;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -189,7 +204,7 @@ bool register_x16rt_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x16rt_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target = (void*)&alt_set_target;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -203,8 +218,8 @@ bool register_x16rt_veil_algo( algo_gate_t* gate )
  gate->hash      = (void*)&x16rt_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target = (void*)&alt_set_target;
  gate->build_extraheader = (void*)&veil_build_extraheader;
+  opt_target_factor = 256.0;
  return true;
 };

@@ -212,19 +227,13 @@ bool register_x16rt_veil_algo( algo_gate_t* gate )
 //
 //    HEX

-
-void hex_set_target( struct work* work, double job_diff )
-{
-   work_set_target( work, job_diff / (128.0 * opt_diff_factor) );
-}
-
 bool register_hex_algo( algo_gate_t* gate )
 {
  gate->scanhash        = (void*)&scanhash_hex;
  gate->hash            = (void*)&hex_hash;
  gate->optimizations   = SSE2_OPT | AES_OPT | AVX2_OPT;
  gate->gen_merkle_root = (void*)&SHA256_gen_merkle_root;
-  gate->set_target      = (void*)&hex_set_target;
+  opt_target_factor = 128.0;
  return true;
 };

@@ -244,8 +253,8 @@ bool register_x21s_algo( algo_gate_t* gate )
  gate->miner_thread_init = (void*)&x21s_thread_init;
 #endif
  gate->optimizations     = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT;
-  gate->set_target        = (void*)&alt_set_target;
  x16_r_s_getAlgoString   = (void*)&x16s_getAlgoString;
+  opt_target_factor = 256.0;
  return true;
 };

--- a/algo/x16/x16r-gate.h
+++ b/algo/x16/x16r-gate.h
@@ -38,6 +38,7 @@ void x16rt_getAlgoString( const uint32_t *timeHash, char *output );
 void x16rt_getTimeHash( const uint32_t timeStamp, void* timeHash );

 bool register_x16r_algo( algo_gate_t* gate );
+bool register_x16rv2_algo( algo_gate_t* gate );
 bool register_x16s_algo( algo_gate_t* gate );
 bool register_x16rt_algo( algo_gate_t* gate );
 bool register_hex__algo( algo_gate_t* gate );
@@ -49,6 +50,10 @@ void x16r_4way_hash( void *state, const void *input );
 int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
                        uint64_t *hashes_done, struct thr_info *mythr );

+void x16rv2_4way_hash( void *state, const void *input );
+int scanhash_x16rv2_4way( struct work *work, uint32_t max_nonce,
+                        uint64_t *hashes_done, struct thr_info *mythr );
+
 void x16rt_4way_hash( void *state, const void *input );
 int scanhash_x16rt_4way( struct work *work, uint32_t max_nonce,
                        uint64_t *hashes_done, struct thr_info *mythr );
@@ -64,6 +69,10 @@ void x16r_hash( void *state, const void *input );
 int scanhash_x16r( struct work *work, uint32_t max_nonce,
                   uint64_t *hashes_done, struct thr_info *mythr );

+void x16rv2_hash( void *state, const void *input );
+int scanhash_x16rv2( struct work *work, uint32_t max_nonce,
+                   uint64_t *hashes_done, struct thr_info *mythr );
+
 void x16rt_hash( void *state, const void *input );
 int scanhash_x16rt( struct work *work, uint32_t max_nonce,
                   uint64_t *hashes_done, struct thr_info *mythr );
--- a/algo/x16/x16rt-4way.c
+++ b/algo/x16/x16rt-4way.c
@@ -24,7 +24,6 @@
 #include "algo/sha/sha-hash-4way.h"

 static __thread uint32_t s_ntime = UINT32_MAX;
-static __thread bool s_implemented = false;
 static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };

 union _x16rt_4way_context_overlay
@@ -64,26 +63,8 @@ void x16rt_4way_hash( void* output, const void* input )

   dintrlv_4x64( hash0, hash1, hash2, hash3, input, 640 );
 
-/*
-   void *in = (void*) input;
-   uint32_t *in32 = (uint32_t*) hash0;
-   uint32_t ntime = in32[17];
-   if ( s_ntime == UINT32_MAX )
-   {
-      uint32_t _ALIGN(64) timeHash[8];
-      x16rt_getTimeHash(ntime, &timeHash);
-      x16rt_getAlgoString(&timeHash[0], hashOrder);
-   }
-*/
-
   // Input data is both 64 bit interleaved (input)
-   // and deinterleaved in inp0-3.
-   // If First function uses 64 bit data it is not required to interleave inp
-   // first. It may use the inerleaved data dmost convenient, ie 4way 64 bit.
-   // All other functions assume data is deinterleaved in hash0-3
-   // All functions must exit with data deinterleaved in hash0-3.
-   // Alias in0-3 points to either inp0-3 or hash0-3 according to
-   // its hashOrder position. Size is also set accordingly.
+   // and deinterleaved in inp0-3. First function has no need re-interleave.
   for ( int i = 0; i < 16; i++ )
   {
      const char elem = hashOrder[i];
@@ -290,44 +271,31 @@ int scanhash_x16rt_4way( struct work *work, uint32_t max_nonce,
 {
   uint32_t hash[4*16] __attribute__ ((aligned (64)));
   uint32_t vdata[24*4] __attribute__ ((aligned (64)));
-   uint32_t endiandata[20] __attribute__((aligned(64)));
   uint32_t _ALIGN(64) timeHash[4*8];
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
   const uint32_t Htarg = ptarget[7];
   const uint32_t first_nonce = pdata[19];
   uint32_t n = first_nonce;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   int thr_id = mythr->id;  
    __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
   volatile uint8_t *restart = &(work_restart[thr_id].restart);

-   casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
-   casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
-   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
+   if ( opt_benchmark )
+      ptarget[7] = 0x0cff;

-   uint32_t ntime = swab32( pdata[17] );
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+   
+   uint32_t ntime = bswap_32( pdata[17] );
   if ( s_ntime != ntime )
   {
      x16rt_getTimeHash( ntime, &timeHash );
      x16rt_getAlgoString( &timeHash[0], hashOrder );
      s_ntime = ntime;
-      s_implemented = true;
      if ( opt_debug && !thr_id )
          applog( LOG_INFO, "hash order: %s time: (%08x) time hash: (%08x)",
                               hashOrder, ntime, timeHash );
   }
-   if ( !s_implemented )
-   {
-      applog( LOG_WARNING, "s not implemented");
-      sleep(1);
-      return 0;
-   }
-
-   if ( opt_benchmark )
-      ptarget[7] = 0x0cff;
-
-   uint64_t *edata = (uint64_t*)endiandata;
-   intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );

   do
   {
--- a/algo/x16/x16rv2-4way.c
+++ b/algo/x16/x16rv2-4way.c
@@ -0,0 +1,382 @@
+/**
+ * x16r algo implementation
+ *
+ * Implementation by tpruvot@github Jan 2018
+ * Optimized by JayDDee@github Jan 2018
+ */
+#include "x16r-gate.h"
+
+#if defined (X16R_4WAY)
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "algo/blake/blake-hash-4way.h"
+#include "algo/bmw/bmw-hash-4way.h"
+#include "algo/groestl/aes_ni/hash-groestl.h"
+#include "algo/groestl/aes_ni/hash-groestl.h"
+#include "algo/skein/skein-hash-4way.h"
+#include "algo/jh/jh-hash-4way.h"
+#include "algo/keccak/keccak-hash-4way.h"
+#include "algo/shavite/sph_shavite.h"
+#include "algo/luffa/luffa-hash-2way.h"
+#include "algo/cubehash/cubehash_sse2.h"
+#include "algo/simd/simd-hash-2way.h"
+#include "algo/echo/aes_ni/hash_api.h"
+#include "algo/hamsi/hamsi-hash-4way.h"
+#include "algo/fugue/sph_fugue.h"
+#include "algo/shabal/shabal-hash-4way.h"
+#include "algo/whirlpool/sph_whirlpool.h"
+#include "algo/sha/sha-hash-4way.h"
+#include "algo/tiger/sph_tiger.h"
+
+static __thread uint32_t s_ntime = UINT32_MAX;
+static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };
+
+union _x16rv2_4way_context_overlay
+{
+    blake512_4way_context   blake;
+    bmw512_4way_context     bmw;
+    hashState_echo          echo;
+    hashState_groestl       groestl;
+    skein512_4way_context   skein;
+    jh512_4way_context      jh;
+    keccak512_4way_context  keccak;
+    luffa_2way_context      luffa;
+    cubehashParam           cube;
+    sph_shavite512_context  shavite;
+    simd_2way_context       simd;
+    hamsi512_4way_context   hamsi;
+    sph_fugue512_context    fugue;
+    shabal512_4way_context  shabal;
+    sph_whirlpool_context   whirlpool;
+    sha512_4way_context     sha512;
+    sph_tiger_context       tiger;
+};
+typedef union _x16rv2_4way_context_overlay x16rv2_4way_context_overlay;
+
+// Pad the 24 bytes tiger hash to 64 bytes
+inline void padtiger512( uint32_t* hash )
+{
+  for ( int i = 6; i < 16; i++ ) hash[i] = 0;
+}
+
+void x16rv2_4way_hash( void* output, const void* input )
+{
+   uint32_t hash0[24] __attribute__ ((aligned (64)));
+   uint32_t hash1[24] __attribute__ ((aligned (64)));
+   uint32_t hash2[24] __attribute__ ((aligned (64)));
+   uint32_t hash3[24] __attribute__ ((aligned (64)));
+   uint32_t vhash[24*4] __attribute__ ((aligned (64)));
+   x16rv2_4way_context_overlay ctx;
+   void *in0 = (void*) hash0;
+   void *in1 = (void*) hash1;
+   void *in2 = (void*) hash2;
+   void *in3 = (void*) hash3;
+   int size = 80;
+
+   dintrlv_4x64( hash0, hash1, hash2, hash3, input, 640 );
+
+   for ( int i = 0; i < 16; i++ )
+   {
+      const char elem = hashOrder[i];
+      const uint8_t algo = elem >= 'A' ? elem - 'A' + 10 : elem - '0';
+
+      switch ( algo )
+      {
+         case BLAKE:
+            blake512_4way_init( &ctx.blake );
+            if ( i == 0 )
+               blake512_4way( &ctx.blake, input, size );
+            else
+            {
+               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               blake512_4way( &ctx.blake, vhash, size );
+            }
+            blake512_4way_close( &ctx.blake, vhash );
+            dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case BMW:
+            bmw512_4way_init( &ctx.bmw );
+            if ( i == 0 )
+               bmw512_4way( &ctx.bmw, input, size );
+            else
+            {
+               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               bmw512_4way( &ctx.bmw, vhash, size );
+            }
+            bmw512_4way_close( &ctx.bmw, vhash );
+            dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case GROESTL:
+               init_groestl( &ctx.groestl, 64 );
+               update_and_final_groestl( &ctx.groestl, (char*)hash0,
+                                                 (const char*)in0, size<<3 );
+               init_groestl( &ctx.groestl, 64 );
+               update_and_final_groestl( &ctx.groestl, (char*)hash1,
+                                                 (const char*)in1, size<<3 );
+               init_groestl( &ctx.groestl, 64 );
+               update_and_final_groestl( &ctx.groestl, (char*)hash2,
+                                                 (const char*)in2, size<<3 );
+               init_groestl( &ctx.groestl, 64 );
+               update_and_final_groestl( &ctx.groestl, (char*)hash3,
+                                                 (const char*)in3, size<<3 );
+         break;
+         case SKEIN:
+            skein512_4way_init( &ctx.skein );
+            if ( i == 0 )
+               skein512_4way( &ctx.skein, input, size );
+            else
+            {
+               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               skein512_4way( &ctx.skein, vhash, size );
+            }
+            skein512_4way_close( &ctx.skein, vhash );
+            dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case JH:
+            jh512_4way_init( &ctx.jh );
+            if ( i == 0 )
+               jh512_4way( &ctx.jh, input, size );
+            else
+            {
+               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               jh512_4way( &ctx.jh, vhash, size );
+            }
+            jh512_4way_close( &ctx.jh, vhash );
+            dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case KECCAK:
+             sph_tiger_init( &ctx.tiger );
+			    sph_tiger( &ctx.tiger, in0, size );
+			    sph_tiger_close( &ctx.tiger, hash0 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in1, size );
+             sph_tiger_close( &ctx.tiger, hash1 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in2, size );
+             sph_tiger_close( &ctx.tiger, hash2 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in3, size );
+             sph_tiger_close( &ctx.tiger, hash3 );
+
+             for ( int i = (24/4); i < (64/4); i++ )
+                hash0[i] = hash1[i] = hash2[i] = hash3[i] = 0;
+
+             intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
+             keccak512_4way_init( &ctx.keccak );
+             keccak512_4way( &ctx.keccak, vhash, 64 );
+             keccak512_4way_close( &ctx.keccak, vhash );
+             dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case LUFFA:
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in0, size );
+             sph_tiger_close( &ctx.tiger, hash0 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in1, size );
+             sph_tiger_close( &ctx.tiger, hash1 );
+
+             for ( int i = (24/4); i < (64/4); i++ )
+                hash0[i] = hash1[i] = 0;
+
+             intrlv_2x128( vhash, hash0, hash1, 512 );
+             luffa_2way_init( &ctx.luffa, 512 );
+             luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
+             dintrlv_2x128( hash0, hash1, vhash, 512 );
+
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in2, size );
+             sph_tiger_close( &ctx.tiger, hash2 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in3, size );
+             sph_tiger_close( &ctx.tiger, hash3 );
+
+             for ( int i = (24/4); i < (64/4); i++ )
+                hash2[i] = hash3[i] = 0;
+             
+             intrlv_2x128( vhash, hash2, hash3, 512 );
+             luffa_2way_init( &ctx.luffa, 512 );
+             luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
+             dintrlv_2x128( hash2, hash3, vhash, 512 );
+         break;
+         case CUBEHASH:
+             cubehashInit( &ctx.cube, 512, 16, 32 );
+             cubehashUpdateDigest( &ctx.cube, (byte*) hash0,
+                                   (const byte*)in0, size );
+             cubehashInit( &ctx.cube, 512, 16, 32 );
+             cubehashUpdateDigest( &ctx.cube, (byte*) hash1,
+                                   (const byte*)in1, size );
+             cubehashInit( &ctx.cube, 512, 16, 32 );
+             cubehashUpdateDigest( &ctx.cube, (byte*) hash2,
+                                   (const byte*)in2, size );
+             cubehashInit( &ctx.cube, 512, 16, 32 );
+             cubehashUpdateDigest( &ctx.cube, (byte*) hash3,
+                                   (const byte*)in3, size );
+         break;
+         case SHAVITE:
+             sph_shavite512_init( &ctx.shavite );
+             sph_shavite512( &ctx.shavite, in0, size );
+             sph_shavite512_close( &ctx.shavite, hash0 );
+             sph_shavite512_init( &ctx.shavite );
+             sph_shavite512( &ctx.shavite, in1, size );
+             sph_shavite512_close( &ctx.shavite, hash1 );
+             sph_shavite512_init( &ctx.shavite );
+             sph_shavite512( &ctx.shavite, in2, size );
+             sph_shavite512_close( &ctx.shavite, hash2 );
+             sph_shavite512_init( &ctx.shavite );
+             sph_shavite512( &ctx.shavite, in3, size );
+             sph_shavite512_close( &ctx.shavite, hash3 );
+         break;
+         case SIMD:
+             intrlv_2x128( vhash, in0, in1, size<<3 );
+             simd_2way_init( &ctx.simd, 512 );
+             simd_2way_update_close( &ctx.simd, vhash, vhash, size<<3 );
+             dintrlv_2x128( hash0, hash1, vhash, 512 );
+             intrlv_2x128( vhash, in2, in3, size<<3 );
+             simd_2way_init( &ctx.simd, 512 );
+             simd_2way_update_close( &ctx.simd, vhash, vhash, size<<3 );
+             dintrlv_2x128( hash2, hash3, vhash, 512 );
+         break;
+         case ECHO:
+             init_echo( &ctx.echo, 512 );
+             update_final_echo ( &ctx.echo, (BitSequence *)hash0,
+                                (const BitSequence*)in0, size<<3 );
+             init_echo( &ctx.echo, 512 );
+             update_final_echo ( &ctx.echo, (BitSequence *)hash1,
+                                (const BitSequence*)in1, size<<3 );
+             init_echo( &ctx.echo, 512 );
+             update_final_echo ( &ctx.echo, (BitSequence *)hash2,
+                                (const BitSequence*)in2, size<<3 );
+             init_echo( &ctx.echo, 512 );
+             update_final_echo ( &ctx.echo, (BitSequence *)hash3,
+                                (const BitSequence*)in3, size<<3 );
+         break;
+         case HAMSI:
+             intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+             hamsi512_4way_init( &ctx.hamsi );
+             hamsi512_4way( &ctx.hamsi, vhash, size );
+             hamsi512_4way_close( &ctx.hamsi, vhash );
+             dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case FUGUE:
+             sph_fugue512_init( &ctx.fugue );
+             sph_fugue512( &ctx.fugue, in0, size );
+             sph_fugue512_close( &ctx.fugue, hash0 );
+             sph_fugue512_init( &ctx.fugue );
+             sph_fugue512( &ctx.fugue, in1, size );
+             sph_fugue512_close( &ctx.fugue, hash1 );
+             sph_fugue512_init( &ctx.fugue );
+             sph_fugue512( &ctx.fugue, in2, size );
+             sph_fugue512_close( &ctx.fugue, hash2 );
+             sph_fugue512_init( &ctx.fugue );
+             sph_fugue512( &ctx.fugue, in3, size );
+             sph_fugue512_close( &ctx.fugue, hash3 );
+         break;
+         case SHABAL:
+             intrlv_4x32( vhash, in0, in1, in2, in3, size<<3 );
+             shabal512_4way_init( &ctx.shabal );
+             shabal512_4way( &ctx.shabal, vhash, size );
+             shabal512_4way_close( &ctx.shabal, vhash );
+             dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case WHIRLPOOL:
+             sph_whirlpool_init( &ctx.whirlpool );
+             sph_whirlpool( &ctx.whirlpool, in0, size );
+             sph_whirlpool_close( &ctx.whirlpool, hash0 );
+             sph_whirlpool_init( &ctx.whirlpool );
+             sph_whirlpool( &ctx.whirlpool, in1, size );
+             sph_whirlpool_close( &ctx.whirlpool, hash1 );
+             sph_whirlpool_init( &ctx.whirlpool );
+             sph_whirlpool( &ctx.whirlpool, in2, size );
+             sph_whirlpool_close( &ctx.whirlpool, hash2 );
+             sph_whirlpool_init( &ctx.whirlpool );
+             sph_whirlpool( &ctx.whirlpool, in3, size );
+             sph_whirlpool_close( &ctx.whirlpool, hash3 );
+         break;
+         case SHA_512:
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in0, size );
+             sph_tiger_close( &ctx.tiger, hash0 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in1, size );
+             sph_tiger_close( &ctx.tiger, hash1 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in2, size );
+             sph_tiger_close( &ctx.tiger, hash2 );
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in3, size );
+             sph_tiger_close( &ctx.tiger, hash3 );
+
+             for ( int i = (24/4); i < (64/4); i++ )
+                hash0[i] = hash1[i] = hash2[i] = hash3[i] = 0;
+ 
+             intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
+             sha512_4way_init( &ctx.sha512 );
+             sha512_4way( &ctx.sha512, vhash, 64 );
+             sha512_4way_close( &ctx.sha512, vhash );
+             dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+      }
+      size = 64;
+   }
+   memcpy( output,    hash0, 32 );
+   memcpy( output+32, hash1, 32 );
+   memcpy( output+64, hash2, 32 );
+   memcpy( output+96, hash3, 32 );
+}
+
+int scanhash_x16rv2_4way( struct work *work, uint32_t max_nonce,
+                        uint64_t *hashes_done, struct thr_info *mythr)
+{
+   uint32_t hash[4*16] __attribute__ ((aligned (64)));
+   uint32_t vdata[24*4] __attribute__ ((aligned (64)));
+   uint32_t bedata1[2] __attribute__((aligned(64)));
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
+   const uint32_t Htarg = ptarget[7];
+   const uint32_t first_nonce = pdata[19];
+   uint32_t n = first_nonce;
+   int thr_id = mythr->id; 
+    __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
+   volatile uint8_t *restart = &(work_restart[thr_id].restart);
+
+   if ( opt_benchmark )
+      ptarget[7] = 0x0fff;
+   
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+
+   bedata1[0] = bswap_32( pdata[1] );
+   bedata1[1] = bswap_32( pdata[2] );
+   const uint32_t ntime = bswap_32(pdata[17]);
+   if ( s_ntime != ntime )
+   {
+      x16_r_s_getAlgoString( (const uint8_t*)bedata1, hashOrder );
+      s_ntime = ntime;
+      if ( opt_debug && !thr_id )
+              applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
+   }
+
+   do
+   {
+      *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
+               _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
+
+      x16rv2_4way_hash( hash, vdata );
+      pdata[19] = n;
+
+      for ( int i = 0; i < 4; i++ )
+      if ( unlikely( (hash+(i<<3))[7] <= Htarg ) )
+      if( likely( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark ) )
+      {
+         pdata[19] = n+i;
+         submit_lane_solution( work, hash+(i<<3), mythr, i );
+      }
+      n += 4;
+   } while ( likely( (  n < max_nonce ) && !(*restart) ) );
+
+   *hashes_done = n - first_nonce + 1;
+   return 0;
+}
+
+#endif
--- a/algo/x16/x16rv2.c
+++ b/algo/x16/x16rv2.c
@@ -0,0 +1,247 @@
+/**
+ * x16r algo implementation
+ *
+ * Implementation by tpruvot@github Jan 2018
+ * Optimized by JayDDee@github Jan 2018
+ */
+#include "x16r-gate.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "algo/blake/sph_blake.h"
+#include "algo/bmw/sph_bmw.h"
+#include "algo/groestl/sph_groestl.h"
+#include "algo/jh/sph_jh.h"
+#include "algo/keccak/sph_keccak.h"
+#include "algo/skein/sph_skein.h"
+#include "algo/shavite/sph_shavite.h"
+#include "algo/luffa/luffa_for_sse2.h"
+#include "algo/cubehash/cubehash_sse2.h"
+#include "algo/simd/nist.h"
+#include "algo/echo/sph_echo.h"
+#include "algo/hamsi/sph_hamsi.h"
+#include "algo/fugue/sph_fugue.h"
+#include "algo/shabal/sph_shabal.h"
+#include "algo/whirlpool/sph_whirlpool.h"
+#include <openssl/sha.h>
+#include "algo/tiger/sph_tiger.h"
+#if defined(__AES__)
+  #include "algo/echo/aes_ni/hash_api.h"
+  #include "algo/groestl/aes_ni/hash-groestl.h"
+#endif
+
+static __thread uint32_t s_ntime = UINT32_MAX;
+static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };
+
+union _x16rv2_context_overlay
+{
+#if defined(__AES__)
+        hashState_echo          echo;
+        hashState_groestl       groestl;
+#else
+        sph_groestl512_context   groestl;
+        sph_echo512_context      echo;
+#endif
+        sph_blake512_context    blake;
+        sph_bmw512_context      bmw;
+        sph_skein512_context    skein;
+        sph_jh512_context       jh;
+        sph_keccak512_context   keccak;
+        hashState_luffa         luffa;
+        cubehashParam           cube;
+        sph_shavite512_context  shavite;
+        hashState_sd            simd;
+        sph_hamsi512_context    hamsi;
+        sph_fugue512_context    fugue;
+        sph_shabal512_context   shabal;
+        sph_whirlpool_context   whirlpool;
+        SHA512_CTX              sha512;
+        sph_tiger_context       tiger;
+};
+typedef union _x16rv2_context_overlay x16rv2_context_overlay;
+
+// Pad the 24 bytes tiger hash to 64 bytes
+inline void padtiger512(uint32_t* hash) {
+   for (int i = (24/4); i < (64/4); i++) hash[i] = 0;
+}
+
+void x16rv2_hash( void* output, const void* input )
+{
+   uint32_t _ALIGN(128) hash[16];
+   x16rv2_context_overlay ctx;
+   void *in = (void*) input;
+   int size = 80;
+/*
+   if ( s_ntime == UINT32_MAX )
+   {
+      const uint8_t* in8 = (uint8_t*) input;
+      x16_r_s_getAlgoString( &in8[4], hashOrder );
+   }
+*/
+   for ( int i = 0; i < 16; i++ )
+   {
+      const char elem = hashOrder[i];
+      const uint8_t algo = elem >= 'A' ? elem - 'A' + 10 : elem - '0';
+
+      switch ( algo )
+      {
+         case BLAKE:
+            sph_blake512_init( &ctx.blake );
+            sph_blake512( &ctx.blake, in, size );
+            sph_blake512_close( &ctx.blake, hash );
+         break;
+         case BMW:
+            sph_bmw512_init( &ctx.bmw );
+            sph_bmw512(&ctx.bmw, in, size);
+            sph_bmw512_close(&ctx.bmw, hash);
+         break;
+         case GROESTL:
+#if defined(__AES__)
+            init_groestl( &ctx.groestl, 64 );
+            update_and_final_groestl( &ctx.groestl, (char*)hash,
+                                      (const char*)in, size<<3 );
+#else
+            sph_groestl512_init( &ctx.groestl );
+            sph_groestl512( &ctx.groestl, in, size );
+            sph_groestl512_close(&ctx.groestl, hash);
+#endif
+         break;
+         case SKEIN:
+            sph_skein512_init( &ctx.skein );
+            sph_skein512( &ctx.skein, in, size );
+            sph_skein512_close( &ctx.skein, hash );
+         break;
+         case JH:
+            sph_jh512_init( &ctx.jh );
+            sph_jh512(&ctx.jh, in, size );
+            sph_jh512_close(&ctx.jh, hash );
+         break;
+         case KECCAK:
+            sph_tiger_init( &ctx.tiger );
+            sph_tiger( &ctx.tiger, in, size );
+            sph_tiger_close( &ctx.tiger, hash );
+            padtiger512( hash );
+            sph_keccak512_init( &ctx.keccak );
+            sph_keccak512( &ctx.keccak, hash, 64 );
+            sph_keccak512_close( &ctx.keccak, hash );
+         break;
+         case LUFFA:
+            sph_tiger_init( &ctx.tiger );
+            sph_tiger( &ctx.tiger, in, size );
+            sph_tiger_close( &ctx.tiger, hash );
+            padtiger512( hash );
+            init_luffa( &ctx.luffa, 512 );
+            update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
+                                    (const BitSequence*)hash, 64 );
+         break;
+         case CUBEHASH:
+            cubehashInit( &ctx.cube, 512, 16, 32 );
+            cubehashUpdateDigest( &ctx.cube, (byte*) hash,
+                                  (const byte*)in, size );
+         break;
+         case SHAVITE:
+            sph_shavite512_init( &ctx.shavite );
+            sph_shavite512( &ctx.shavite, in, size );
+            sph_shavite512_close( &ctx.shavite, hash );
+         break;
+         case SIMD:
+             init_sd( &ctx.simd, 512 );
+             update_final_sd( &ctx.simd, (BitSequence *)hash,
+                              (const BitSequence*)in, size<<3 );
+         break;
+         case ECHO:
+#if defined(__AES__)
+             init_echo( &ctx.echo, 512 );
+             update_final_echo ( &ctx.echo, (BitSequence *)hash,
+                                (const BitSequence*)in, size<<3 );
+#else
+             sph_echo512_init( &ctx.echo );
+             sph_echo512( &ctx.echo, in, size );
+             sph_echo512_close( &ctx.echo, hash );
+#endif
+         break;
+         case HAMSI:
+             sph_hamsi512_init( &ctx.hamsi );
+             sph_hamsi512( &ctx.hamsi, in, size );
+             sph_hamsi512_close( &ctx.hamsi, hash );
+         break;
+         case FUGUE:
+             sph_fugue512_init( &ctx.fugue );
+             sph_fugue512( &ctx.fugue, in, size );
+             sph_fugue512_close( &ctx.fugue, hash );
+         break;
+         case SHABAL:
+             sph_shabal512_init( &ctx.shabal );
+             sph_shabal512( &ctx.shabal, in, size );
+             sph_shabal512_close( &ctx.shabal, hash );
+         break;
+         case WHIRLPOOL:
+             sph_whirlpool_init( &ctx.whirlpool );
+             sph_whirlpool( &ctx.whirlpool, in, size );
+             sph_whirlpool_close( &ctx.whirlpool, hash );
+         break;
+         case SHA_512:
+             sph_tiger_init( &ctx.tiger );
+             sph_tiger( &ctx.tiger, in, size );
+             sph_tiger_close( &ctx.tiger, hash );
+             padtiger512( hash );
+             SHA512_Init( &ctx.sha512 );
+             SHA512_Update( &ctx.sha512, hash, 64 );
+             SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+         break;
+      }
+      in = (void*) hash;
+      size = 64;
+   }
+   memcpy(output, hash, 32);
+}
+
+int scanhash_x16rv2( struct work *work, uint32_t max_nonce,
+                   uint64_t *hashes_done, struct thr_info *mythr )
+{
+   uint32_t _ALIGN(128) hash32[8];
+   uint32_t _ALIGN(128) endiandata[20];
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
+   const uint32_t Htarg = ptarget[7];
+   const uint32_t first_nonce = pdata[19];
+   int thr_id = mythr->id;  // thr_id arg is deprecated
+   uint32_t nonce = first_nonce;
+   volatile uint8_t *restart = &(work_restart[thr_id].restart);
+
+   casti_m128i( endiandata, 0 ) = mm128_bswap_32( casti_m128i( pdata, 0 ) );
+   casti_m128i( endiandata, 1 ) = mm128_bswap_32( casti_m128i( pdata, 1 ) );
+   casti_m128i( endiandata, 2 ) = mm128_bswap_32( casti_m128i( pdata, 2 ) );
+   casti_m128i( endiandata, 3 ) = mm128_bswap_32( casti_m128i( pdata, 3 ) );
+   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
+
+   if ( s_ntime != pdata[17] )
+   {
+      uint32_t ntime = swab32(pdata[17]);
+      x16_r_s_getAlgoString( (const uint8_t*) (&endiandata[1]), hashOrder );
+      s_ntime = ntime;
+      if ( opt_debug && !thr_id )
+              applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
+   }
+
+   if ( opt_benchmark )
+      ptarget[7] = 0x0cff;
+
+   do
+   {
+      be32enc( &endiandata[19], nonce );
+      x16rv2_hash( hash32, endiandata );
+
+      if ( hash32[7] <= Htarg )
+      if (fulltest( hash32, ptarget ) && !opt_benchmark )
+      {
+         pdata[19] = nonce;
+         submit_solution( work, hash32, mythr );
+      }
+      nonce++;
+   } while ( nonce < max_nonce && !(*restart) );
+   pdata[19] = nonce;
+   *hashes_done = pdata[19] - first_nonce + 1;
+   return 0;
+}
--- a/algo/x16/x21s-4way.c
+++ b/algo/x16/x21s-4way.c
@@ -368,7 +368,7 @@ int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
 {
   uint32_t hash[4*16] __attribute__ ((aligned (64)));
   uint32_t vdata[24*4] __attribute__ ((aligned (64)));
-   uint32_t endiandata[20] __attribute__((aligned(64)));
+   uint32_t bedata1[2] __attribute__((aligned(64)));
   uint32_t *pdata = work->data;
   uint32_t *ptarget = work->target;
   const uint32_t Htarg = ptarget[7];
@@ -378,25 +378,22 @@ int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
    __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
   volatile uint8_t *restart = &(work_restart[thr_id].restart);

-   casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
-   casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
-   casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
+   if ( opt_benchmark )
+    ptarget[7] = 0x0cff;
+ 
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );

-   if ( s_ntime != endiandata[17] )
+   bedata1[0] = bswap_32( pdata[1] );
+   bedata1[1] = bswap_32( pdata[2] );
+   uint32_t ntime = bswap_32( pdata[17] );
+   if ( s_ntime != ntime )
   {
-      uint32_t ntime = swab32(pdata[17]);
-      x16_r_s_getAlgoString( (const uint8_t*) (&endiandata[1]), hashOrder );
+      x16_r_s_getAlgoString( (const uint8_t*)bedata1, hashOrder );
      s_ntime = ntime;
      if ( opt_debug && !thr_id )
              applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
   }

-   if ( opt_benchmark )
-      ptarget[7] = 0x0cff;
-
-   uint64_t *edata = (uint64_t*)endiandata;
-   intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
-
   do
   {
      *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
--- a/algo/x17/sonoa-4way.c
+++ b/algo/x17/sonoa-4way.c
@@ -803,52 +803,40 @@ void sonoa_4way_hash( void *state, const void *input )
     haval256_5_4way_close( &ctx.haval, state );
 }

-int scanhash_sonoa_4way( struct work *work, uint32_t max_nonce,
+int scanhash_sonoa_4way( struct work *work, const uint32_t max_nonce,
 	            uint64_t *hashes_done, struct thr_info *mythr )
 {
     uint32_t hash[4*16] __attribute__ ((aligned (64)));
     uint32_t vdata[24*4] __attribute__ ((aligned (64)));
     uint32_t lane_hash[8] __attribute__ ((aligned (32)));
-     uint32_t *hash7 = &(hash[7<<2]);
+     uint32_t *hash7 = &( hash[7<<2] );
     uint32_t *pdata = work->data;
-     uint32_t *ptarget = work->target;
-     uint32_t n = pdata[19];
+     const uint32_t *ptarget = work->target;
     const uint32_t first_nonce = pdata[19];
-     __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
     const uint32_t Htarg = ptarget[7];
-     int thr_id = mythr->id;
-     uint64_t htmax[] = {          0,        0xF,       0xFF,
-                               0xFFF,     0xFFFF, 0x10000000  };
-     uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
-                          0xFFFFF000, 0xFFFF0000,          0  };
+     uint32_t n = first_nonce;
+     __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
+     const int thr_id = mythr->id;

-     // Need big endian data
     mm256_bswap32_intrlv80_4x64( vdata, pdata );
-     for ( int m=0; m < 6; m++ ) if ( Htarg <= htmax[m] )
+     do
     {
-        uint32_t mask = masks[m];
-        do
+        *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
+                _mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ), *noncev );
+        sonoa_4way_hash( hash, vdata );
+
+        for ( int lane = 0; lane < 4; lane++ )
+        if ( unlikely( hash7[ lane ] <= Htarg ) )
        {
-           *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
-                             _mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ),
-                                                *noncev );
-           sonoa_4way_hash( hash, vdata );
-
-           for ( int lane = 0; lane < 4; lane++ )
-           if ( ( ( hash7[ lane ] & mask ) == 0 ) )
+           extr_lane_4x32( lane_hash, hash, lane, 256 );
+           if ( likely( fulltest( lane_hash, ptarget ) && !opt_benchmark ) )
           {
-              extr_lane_4x32( lane_hash, hash, lane, 256 );
-              if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
-              {
-                 pdata[19] = n + lane;
-                 submit_lane_solution( work, lane_hash, mythr, lane );
-              }
+              pdata[19] = n + lane;
+              submit_lane_solution( work, lane_hash, mythr, lane );
           }
-           n += 4;
-        } while ( ( n < max_nonce - 4 ) && !work_restart[thr_id].restart );
-        break;
-     }
-
+        }
+        n += 4;
+     } while ( ( n < max_nonce - 4 ) && !work_restart[thr_id].restart );
     *hashes_done = n - first_nonce + 1;
     return 0;
 }
--- a/algo/x17/sonoa-gate.c
+++ b/algo/x17/sonoa-gate.c
@@ -11,7 +11,6 @@ bool register_sonoa_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_sonoa;
  gate->hash      = (void*)&sonoa_hash;
 #endif
-  gate->get_max64     = (void*)&get_max64_0x1ffff;
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
  return true;
 };
--- a/algo/x17/x17-4way.c
+++ b/algo/x17/x17-4way.c
@@ -205,50 +205,40 @@ void x17_4way_hash( void *state, const void *input )
 int scanhash_x17_4way( struct work *work, uint32_t max_nonce,
                       uint64_t *hashes_done, struct thr_info *mythr )
 {
-     uint32_t hash[4*16] __attribute__ ((aligned (64)));
-     uint32_t vdata[24*4] __attribute__ ((aligned (64)));
-     uint32_t lane_hash[8] __attribute__ ((aligned (32)));
-     uint32_t *hash7 = &(hash[7<<2]);
-     uint32_t *pdata = work->data;
-     const uint32_t *ptarget = work->target;
-     const uint32_t first_nonce = pdata[19];
-     __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
-     uint32_t n = first_nonce;
-     const int thr_id = mythr->id;
-     const uint32_t Htarg = ptarget[7];
-     uint64_t htmax[] = {          0,        0xF,       0xFF,
-                               0xFFF,     0xFFFF, 0x10000000  };
-     uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
-                          0xFFFFF000, 0xFFFF0000,          0  };
+   uint32_t hash[4*16] __attribute__ ((aligned (64)));
+   uint32_t vdata[24*4] __attribute__ ((aligned (64)));
+   uint32_t lane_hash[8] __attribute__ ((aligned (32)));
+   uint32_t *hash7 = &(hash[7<<2]);
+   uint32_t *pdata = work->data;
+   const uint32_t *ptarget = work->target;
+   const uint32_t first_nonce = pdata[19];
+   __m256i  *noncev = (__m256i*)vdata + 9;   // aligned
+   uint32_t n = first_nonce;
+   const int thr_id = mythr->id;
+   const uint32_t Htarg = ptarget[7];

-     // Need big endian data
-     mm256_bswap32_intrlv80_4x64( vdata, pdata );
-     for ( int m = 0; m < 6; m++ ) if ( Htarg <= htmax[m] )
-     {
-        const uint32_t mask = masks[ m ];
-        do
-        {
-           *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
-	              _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
-           x17_4way_hash( hash, vdata );
+   mm256_bswap32_intrlv80_4x64( vdata, pdata );
+   do
+   {
+      *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
+              _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
+      x17_4way_hash( hash, vdata );

-           for ( int lane = 0; lane < 4; lane++ )
-           if ( ( hash7[ lane ] & mask ) == 0 )
-           {
-              extr_lane_4x32( lane_hash, hash, lane, 256 );
-              if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
-              {
-                 pdata[19] = n + lane;
-                 submit_lane_solution( work, lane_hash, mythr, lane );
-              }
-           }
-           n += 4;
-        } while ( ( n < max_nonce - 4 ) && !work_restart[thr_id].restart );
-        break;
-     }
+      for ( int lane = 0; lane < 4; lane++ )
+      if unlikely( ( hash7[ lane ] <= Htarg ) )
+      {
+         extr_lane_4x32( lane_hash, hash, lane, 256 );
+         if ( likely( fulltest( lane_hash, ptarget ) && !opt_benchmark ) )
+         {
+            pdata[19] = n + lane;
+            submit_lane_solution( work, lane_hash, mythr, lane );
+         }
+      }
+      n += 4;
+   } while ( likely( ( n < max_nonce - 4 ) && !work_restart[thr_id].restart ) );

-     *hashes_done = n - first_nonce + 1;
-     return 0;
+   *hashes_done = n - first_nonce + 1;
+   return 0;
 }

 #endif
--- a/algo/x17/xevan-gate.c
+++ b/algo/x17/xevan-gate.c
@@ -1,10 +1,5 @@
 #include "xevan-gate.h"

-void xevan_set_target( struct work* work, double job_diff )
-{
- work_set_target( work, job_diff / (256.0 * opt_diff_factor) );
-}
-
 bool register_xevan_algo( algo_gate_t* gate )
 {
 #if defined (XEVAN_4WAY)
@@ -17,8 +12,7 @@ bool register_xevan_algo( algo_gate_t* gate )
  gate->hash      = (void*)&xevan_hash;
 #endif
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
-  gate->set_target = (void*)&xevan_set_target;
-  gate->get_max64  = (void*)&get_max64_0xffffLL;
+  opt_target_factor = 256.0;
  return true;
 };

--- a/algo/x20/x20r-gate.c
+++ b/algo/x20/x20r-gate.c
@@ -26,9 +26,9 @@ bool register_x20r_algo( algo_gate_t* gate )
  gate->scanhash  = (void*)&scanhash_x20r;
  gate->hash      = (void*)&x20r_hash;
 #endif
-  gate->set_target = (void*)&alt_set_target;
  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
  x20_r_s_getAlgoString = (void*)&x20r_getAlgoString;
+  opt_target_factor = 256.;
  return true;
 };

--- a/algo/yescrypt/yescrypt.c
+++ b/algo/yescrypt/yescrypt.c
@@ -416,40 +416,48 @@ int scanhash_yescrypt( struct work *work, uint32_t max_nonce,
        return 0;
 }

-int64_t yescrypt_get_max64()
-{
-  return 0x1ffLL;
-}
-
-int64_t yescryptr16_get_max64()
-{
-  return 0xfffLL;
-}
-
 void yescrypt_gate_base(algo_gate_t *gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yescrypt;
   gate->hash       = (void*)&yescrypt_hash;
-   gate->set_target = (void*)&scrypt_set_target;
+   opt_target_factor = 65536.0;
 }

 bool register_yescrypt_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
-   gate->get_max64  = (void*)&yescrypt_get_max64;
-   yescrypt_client_key = NULL;
-   yescrypt_client_key_len = 0;
-   YESCRYPT_N = 2048;
-   YESCRYPT_R = 8;
+
+   if ( opt_param_n )  YESCRYPT_N = opt_param_n;
+   else                YESCRYPT_N = 2048;
+
+   if ( opt_param_r )  YESCRYPT_R = opt_param_r;
+   else                YESCRYPT_R = 8;
+ 
+   if ( opt_param_key ) 
+   {   
+     yescrypt_client_key = opt_param_key;
+     yescrypt_client_key_len = strlen( opt_param_key );
+   }
+   else
+   {   
+     yescrypt_client_key = NULL;
+     yescrypt_client_key_len = 0;
+   }
+
   YESCRYPT_P = 1;
+
+   applog( LOG_NOTICE,"Yescrypt parameters: N= %d, R= %d.", YESCRYPT_N,
+                                                            YESCRYPT_R );
+   if ( yescrypt_client_key )
+     applog( LOG_NOTICE,"Key= \"%s\"\n", yescrypt_client_key );
+
   return true;
 }

 bool register_yescryptr8_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
-   gate->get_max64  = (void*)&yescrypt_get_max64;
   yescrypt_client_key = "Client Key";
   yescrypt_client_key_len = 10;
   YESCRYPT_N = 2048;
@@ -461,7 +469,6 @@ bool register_yescryptr8_algo( algo_gate_t* gate )
 bool register_yescryptr16_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
-   gate->get_max64  = (void*)&yescryptr16_get_max64;
   yescrypt_client_key = "Client Key";
   yescrypt_client_key_len = 10;
   YESCRYPT_N = 4096;   
@@ -473,7 +480,6 @@ bool register_yescryptr16_algo( algo_gate_t* gate )
 bool register_yescryptr32_algo( algo_gate_t* gate )
 {
   yescrypt_gate_base( gate );
-   gate->get_max64  = (void*)&yescryptr16_get_max64;
   yescrypt_client_key = "WaviBanana";
   yescrypt_client_key_len = 10;
   YESCRYPT_N = 4096;
--- a/algo/yespower/crypto/blake2b-yp.c
+++ b/algo/yespower/crypto/blake2b-yp.c
@@ -0,0 +1,322 @@
+/*
+ * Copyright 2009 Colin Percival, 2014 savale
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *  notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *  notice, this list of conditions and the following disclaimer in the
+ *  documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * This file was originally written by Colin Percival as part of the Tarsnap
+ * online backup system.
+ */
+
+#include <stdlib.h>
+#include <stdint.h>
+#include <string.h>
+
+#include <algo/yespower/crypto/sph_types.h>
+#include <algo/yespower/utils/sysendian.h>
+#include "blake2b-yp.h"
+
+// Cyclic right rotation.
+#ifndef ROTR64
+#define ROTR64(x, y)  (((x) >> (y)) ^ ((x) << (64 - (y))))
+#endif
+
+// Little-endian byte access.
+#define B2B_GET64(p)                            \
+    (((uint64_t) ((uint8_t *) (p))[0]) ^        \
+    (((uint64_t) ((uint8_t *) (p))[1]) << 8) ^  \
+    (((uint64_t) ((uint8_t *) (p))[2]) << 16) ^ \
+    (((uint64_t) ((uint8_t *) (p))[3]) << 24) ^ \
+    (((uint64_t) ((uint8_t *) (p))[4]) << 32) ^ \
+    (((uint64_t) ((uint8_t *) (p))[5]) << 40) ^ \
+    (((uint64_t) ((uint8_t *) (p))[6]) << 48) ^ \
+    (((uint64_t) ((uint8_t *) (p))[7]) << 56))
+
+// G Mixing function.
+#define B2B_G(a, b, c, d, x, y) {   \
+    v[a] = v[a] + v[b] + x;      \
+    v[d] = ROTR64(v[d] ^ v[a], 32); \
+    v[c] = v[c] + v[d];          \
+    v[b] = ROTR64(v[b] ^ v[c], 24); \
+    v[a] = v[a] + v[b] + y;      \
+    v[d] = ROTR64(v[d] ^ v[a], 16); \
+    v[c] = v[c] + v[d];          \
+    v[b] = ROTR64(v[b] ^ v[c], 63); }
+
+// Initialization Vector.
+static const uint64_t blake2b_iv[8] = {
+    0x6A09E667F3BCC908, 0xBB67AE8584CAA73B,
+    0x3C6EF372FE94F82B, 0xA54FF53A5F1D36F1,
+    0x510E527FADE682D1, 0x9B05688C2B3E6C1F,
+    0x1F83D9ABFB41BD6B, 0x5BE0CD19137E2179
+};
+
+// Compression function. "last" flag indicates last block.
+static void blake2b_compress(blake2b_yp_ctx *ctx, int last)
+{
+    const uint8_t sigma[12][16] = {
+        { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 },
+        { 14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3 },
+        { 11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4 },
+        { 7, 9, 3, 1, 13, 12, 11, 14, 2, 6, 5, 10, 4, 0, 15, 8 },
+        { 9, 0, 5, 7, 2, 4, 10, 15, 14, 1, 11, 12, 6, 8, 3, 13 },
+        { 2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9 },
+        { 12, 5, 1, 15, 14, 13, 4, 10, 0, 7, 6, 3, 9, 2, 8, 11 },
+        { 13, 11, 7, 14, 12, 1, 3, 9, 5, 0, 15, 4, 8, 6, 2, 10 },
+        { 6, 15, 14, 9, 11, 3, 0, 8, 12, 2, 13, 7, 1, 4, 10, 5 },
+        { 10, 2, 8, 4, 7, 6, 1, 5, 15, 11, 9, 14, 3, 12, 13, 0 },
+        { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 },
+        { 14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3 }
+    };
+    int i;
+    uint64_t v[16], m[16];
+
+    // init work variables
+    for (i = 0; i < 8; i++) {
+        v[i] = ctx->h[i];
+        v[i + 8] = blake2b_iv[i];
+    }
+
+    v[12] ^= ctx->t[0]; // low 64 bits of offset
+    v[13] ^= ctx->t[1]; // high 64 bits
+
+    // last block flag set ?
+    if (last) { 
+        v[14] = ~v[14];
+    }
+
+    // get little-endian words
+    for (i = 0; i < 16; i++) {
+        m[i] = B2B_GET64(&ctx->b[8 * i]);
+    }
+
+    // twelve rounds
+    for (i = 0; i < 12; i++) {
+        B2B_G( 0, 4,  8, 12, m[sigma[i][ 0]], m[sigma[i][ 1]]);
+        B2B_G( 1, 5,  9, 13, m[sigma[i][ 2]], m[sigma[i][ 3]]);
+        B2B_G( 2, 6, 10, 14, m[sigma[i][ 4]], m[sigma[i][ 5]]);
+        B2B_G( 3, 7, 11, 15, m[sigma[i][ 6]], m[sigma[i][ 7]]);
+        B2B_G( 0, 5, 10, 15, m[sigma[i][ 8]], m[sigma[i][ 9]]);
+        B2B_G( 1, 6, 11, 12, m[sigma[i][10]], m[sigma[i][11]]);
+        B2B_G( 2, 7,  8, 13, m[sigma[i][12]], m[sigma[i][13]]);
+        B2B_G( 3, 4,  9, 14, m[sigma[i][14]], m[sigma[i][15]]);
+    }
+
+    for(i = 0; i < 8; ++i) {
+        ctx->h[i] ^= v[i] ^ v[i + 8];
+    }
+}
+
+// Initialize the hashing context "ctx" with optional key "key".
+// 1 <= outlen <= 64 gives the digest size in bytes.
+// Secret key (also <= 64 bytes) is optional (keylen = 0).
+int blake2b_yp_init(blake2b_yp_ctx *ctx, size_t outlen,
+    const void *key, size_t keylen) // (keylen=0: no key)
+{
+    size_t i;
+
+    // illegal parameters
+    if (outlen == 0 || outlen > 64 || keylen > 64) {
+        return -1;
+    }
+
+    // state, "param block"
+    for (i = 0; i < 8; i++) {
+        ctx->h[i] = blake2b_iv[i];
+    }
+
+    ctx->h[0] ^= 0x01010000 ^ (keylen << 8) ^ outlen;
+
+    ctx->t[0] = 0; // input count low word
+    ctx->t[1] = 0; // input count high word
+    ctx->c = 0; // pointer within buffer
+    ctx->outlen = outlen;
+
+    // zero input block
+    for (i = keylen; i < 128; i++) {
+        ctx->b[i] = 0;
+    }
+
+    if (keylen > 0) {
+        blake2b_yp_update(ctx, key, keylen);
+        ctx->c = 128; // at the end
+    }
+
+    return 0;
+}
+
+// Add "inlen" bytes from "in" into the hash.
+void blake2b_yp_update(blake2b_yp_ctx *ctx,
+    const void *in, size_t inlen) // data bytes
+{
+    size_t i;
+    for (i = 0; i < inlen; i++) {
+        if (ctx->c == 128) { // buffer full ?
+            ctx->t[0] += ctx->c; // add counters
+            if (ctx->t[0] < ctx->c) // carry overflow ?
+                ctx->t[1]++; // high word
+            blake2b_compress(ctx, 0); // compress (not last)
+            ctx->c = 0; // counter to zero
+        }
+        ctx->b[ctx->c++] = ((const uint8_t *) in)[i];
+    }
+}
+
+// Generate the message digest (size given in init).
+// Result placed in "out".
+void blake2b_yp_final(blake2b_yp_ctx *ctx, void *out)
+{
+    size_t i;
+
+    ctx->t[0] += ctx->c; // mark last block offset
+    // carry overflow
+    if (ctx->t[0] < ctx->c) {
+        ctx->t[1]++; // high word
+    }
+
+    // fill up with zeros
+    while (ctx->c < 128) {
+        ctx->b[ctx->c++] = 0;
+    }
+
+    blake2b_compress(ctx, 1); // final block flag = 1
+
+    // little endian convert and store
+    for (i = 0; i < ctx->outlen; i++) {
+        ((uint8_t *) out)[i] =
+            (ctx->h[i >> 3] >> (8 * (i & 7))) & 0xFF;
+    }
+}
+
+// inlen = number of bytes
+void blake2b_yp_hash(void *out, const void *in, size_t inlen) {
+    blake2b_yp_ctx ctx;
+    blake2b_yp_init(&ctx, 32, NULL, 0);
+    blake2b_yp_update(&ctx, in, inlen);
+    blake2b_yp_final(&ctx, out);
+}
+
+// // keylen = number of bytes
+void hmac_blake2b_yp_init(hmac_yp_ctx *hctx, const void *_key, size_t keylen) {
+    const uint8_t *key = _key;
+    uint8_t keyhash[32];
+    uint8_t pad[64];
+    uint64_t i;
+
+    if (keylen > 64) {
+        blake2b_yp_hash(keyhash, key, keylen);
+        key = keyhash;
+        keylen = 32;
+    }
+
+    blake2b_yp_init(&hctx->inner, 32, NULL, 0);
+    memset(pad, 0x36, 64);
+    for (i = 0; i < keylen; ++i) {
+        pad[i] ^= key[i];
+    }
+
+    blake2b_yp_update(&hctx->inner, pad, 64);
+    blake2b_yp_init(&hctx->outer, 32, NULL, 0);
+    memset(pad, 0x5c, 64);
+    for (i = 0; i < keylen; ++i) {
+        pad[i] ^= key[i];
+    }
+
+    blake2b_yp_update(&hctx->outer, pad, 64);
+    memset(keyhash, 0, 32);
+}
+
+// datalen = number of bits
+void hmac_blake2b_yp_update(hmac_yp_ctx *hctx, const void *data, size_t datalen) {
+    // update the inner state
+    blake2b_yp_update(&hctx->inner, data, datalen);
+}
+
+void hmac_blake2b_yp_final(hmac_yp_ctx *hctx, uint8_t *digest) {
+    uint8_t ihash[32];
+    blake2b_yp_final(&hctx->inner, ihash);
+    blake2b_yp_update(&hctx->outer, ihash, 32);
+    blake2b_yp_final(&hctx->outer, digest);
+    memset(ihash, 0, 32);
+}
+
+// // keylen = number of bytes; inlen = number of bytes
+void hmac_blake2b_yp_hash(void *out, const void *key, size_t keylen, const void *in, size_t inlen) {
+    hmac_yp_ctx hctx;
+    hmac_blake2b_yp_init(&hctx, key, keylen);
+    hmac_blake2b_yp_update(&hctx, in, inlen);
+    hmac_blake2b_yp_final(&hctx, out);
+}
+
+void pbkdf2_blake2b_yp(const uint8_t * passwd, size_t passwdlen, const uint8_t * salt,
+    size_t saltlen, uint64_t c, uint8_t * buf, size_t dkLen)
+{
+    hmac_yp_ctx PShctx, hctx;
+    size_t i;
+    uint8_t ivec[4];
+    uint8_t U[32];
+    uint8_t T[32];
+    uint64_t j;
+    int k;
+    size_t clen;
+
+    /* Compute HMAC state after processing P and S. */
+    hmac_blake2b_yp_init(&PShctx, passwd, passwdlen);
+    hmac_blake2b_yp_update(&PShctx, salt, saltlen);
+
+    /* Iterate through the blocks. */
+    for (i = 0; i * 32 < dkLen; i++) {
+        /* Generate INT(i + 1). */
+        be32enc(ivec, (uint32_t)(i + 1));
+
+        /* Compute U_1 = PRF(P, S || INT(i)). */
+        memcpy(&hctx, &PShctx, sizeof(hmac_yp_ctx));
+        hmac_blake2b_yp_update(&hctx, ivec, 4);
+        hmac_blake2b_yp_final(&hctx, U);
+
+        /* T_i = U_1 ... */
+        memcpy(T, U, 32);
+
+        for (j = 2; j <= c; j++) {
+            /* Compute U_j. */
+            hmac_blake2b_yp_init(&hctx, passwd, passwdlen);
+            hmac_blake2b_yp_update(&hctx, U, 32);
+            hmac_blake2b_yp_final(&hctx, U);
+
+            /* ... xor U_j ... */
+            for (k = 0; k < 32; k++) {
+                T[k] ^= U[k];
+            }
+        }
+
+        /* Copy as many bytes as necessary into buf. */
+        clen = dkLen - i * 32;
+        if (clen > 32) {
+            clen = 32;
+        }
+
+        memcpy(&buf[i * 32], T, clen);
+    }
+
+    /* Clean PShctx, since we never called _Final on it. */
+    memset(&PShctx, 0, sizeof(hmac_yp_ctx));
+}
--- a/algo/yespower/crypto/blake2b-yp.h
+++ b/algo/yespower/crypto/blake2b-yp.h
@@ -0,0 +1,42 @@
+#pragma once
+#ifndef __BLAKE2B_H__
+#define __BLAKE2B_H__
+
+#include <stddef.h>
+#include <stdint.h>
+
+#if defined(_MSC_VER) || defined(__x86_64__) || defined(__x86__)
+#define NATIVE_LITTLE_ENDIAN
+#endif
+
+// state context
+typedef struct {
+    uint8_t b[128]; // input buffer
+    uint64_t h[8];  // chained state
+    uint64_t t[2];  // total number of bytes
+    size_t c;       // pointer for b[]
+    size_t outlen;  // digest size
+} blake2b_yp_ctx;
+
+typedef struct {
+    blake2b_yp_ctx inner;
+    blake2b_yp_ctx outer;
+} hmac_yp_ctx;
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+int blake2b_yp_init(blake2b_yp_ctx *ctx, size_t outlen, const void *key, size_t keylen);
+void blake2b_yp_update(blake2b_yp_ctx *ctx, const void *in, size_t inlen);
+void blake2b_yp_final(blake2b_yp_ctx *ctx, void *out);
+void blake2b_yp_hash(void *out, const void *in, size_t inlen);
+void hmac_blake2b_yp_hash(void *out, const void *key, size_t keylen, const void *in, size_t inlen);
+void pbkdf2_blake2b_yp(const uint8_t * passwd, size_t passwdlen, const uint8_t * salt,
+    size_t saltlen, uint64_t c, uint8_t * buf, size_t dkLen);
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif
--- a/algo/yespower/crypto/sph_types.h
+++ b/algo/yespower/crypto/sph_types.h
--- a/algo/yespower/utils/insecure_memzero.h
+++ b/algo/yespower/utils/insecure_memzero.h
@@ -0,0 +1 @@
+#define insecure_memzero(buf, len) /* empty */
--- a/algo/yespower/utils/sysendian.h
+++ b/algo/yespower/utils/sysendian.h
@@ -0,0 +1,94 @@
+/*-
+ * Copyright 2007-2014 Colin Percival
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#ifndef _SYSENDIAN_H_
+#define _SYSENDIAN_H_
+
+#include <stdint.h>
+
+/* Avoid namespace collisions with BSD <sys/endian.h>. */
+#define be32dec libcperciva_be32dec
+#define be32enc libcperciva_be32enc
+#define be64enc libcperciva_be64enc
+#define le32dec libcperciva_le32dec
+#define le32enc libcperciva_le32enc
+
+static inline uint32_t
+be32dec(const void * pp)
+{
+    const uint8_t * p = (uint8_t const *)pp;
+
+    return ((uint32_t)(p[3]) + ((uint32_t)(p[2]) << 8) +
+        ((uint32_t)(p[1]) << 16) + ((uint32_t)(p[0]) << 24));
+}
+
+static inline void
+be32enc(void * pp, uint32_t x)
+{
+    uint8_t * p = (uint8_t *)pp;
+
+    p[3] = x & 0xff;
+    p[2] = (x >> 8) & 0xff;
+    p[1] = (x >> 16) & 0xff;
+    p[0] = (x >> 24) & 0xff;
+}
+
+static inline void
+be64enc(void * pp, uint64_t x)
+{
+    uint8_t * p = (uint8_t *)pp;
+
+    p[7] = x & 0xff;
+    p[6] = (x >> 8) & 0xff;
+    p[5] = (x >> 16) & 0xff;
+    p[4] = (x >> 24) & 0xff;
+    p[3] = (x >> 32) & 0xff;
+    p[2] = (x >> 40) & 0xff;
+    p[1] = (x >> 48) & 0xff;
+    p[0] = (x >> 56) & 0xff;
+}
+
+static inline uint32_t
+le32dec(const void * pp)
+{
+    const uint8_t * p = (uint8_t const *)pp;
+
+    return ((uint32_t)(p[0]) + ((uint32_t)(p[1]) << 8) +
+        ((uint32_t)(p[2]) << 16) + ((uint32_t)(p[3]) << 24));
+}
+
+static inline void
+le32enc(void * pp, uint32_t x)
+{
+    uint8_t * p = (uint8_t *)pp;
+
+    p[0] = x & 0xff;
+    p[1] = (x >> 8) & 0xff;
+    p[2] = (x >> 16) & 0xff;
+    p[3] = (x >> 24) & 0xff;
+}
+
+#endif /* !_SYSENDIAN_H_ */
--- a/algo/yespower/yespower-blake2b.c
+++ b/algo/yespower/yespower-blake2b.c
--- a/algo/yespower/yespower-gate.c
+++ b/algo/yespower/yespower-gate.c
@@ -70,23 +70,75 @@ int scanhash_yespower( struct work *work, uint32_t max_nonce,
        return 0;
 }

-int64_t yespower_get_max64()
+
+void yespower_b2b_hash( const char *input, char *output, uint32_t len )
 {
-  return 0xfffLL;
+  yespower_b2b_tls( input, len, &yespower_params, (yespower_binary_t*)output );
+}
+
+int scanhash_yespower_b2b( struct work *work, uint32_t max_nonce,
+                       uint64_t *hashes_done, struct thr_info *mythr )
+{
+        uint32_t _ALIGN(64) vhash[8];
+        uint32_t _ALIGN(64) endiandata[20];
+        uint32_t *pdata = work->data;
+        uint32_t *ptarget = work->target;
+
+        const uint32_t Htarg = ptarget[7];
+        const uint32_t first_nonce = pdata[19];
+        uint32_t n = first_nonce;
+        int thr_id = mythr->id;  // thr_id arg is deprecated
+
+        for (int k = 0; k < 19; k++)
+                be32enc(&endiandata[k], pdata[k]);
+        do {
+           be32enc(&endiandata[19], n);
+           yespower_b2b_hash((char*) endiandata, (char*) vhash, 80);
+           if ( vhash[7] < Htarg && fulltest( vhash, ptarget )
+              && !opt_benchmark )
+           {
+               pdata[19] = n;
+               submit_solution( work, vhash, mythr );
+            }
+            n++;
+        } while (n < max_nonce && !work_restart[thr_id].restart);
+
+        *hashes_done = n - first_nonce + 1;
+        pdata[19] = n;
+
+        return 0;
 }

 bool register_yespower_algo( algo_gate_t* gate )
 {
  yespower_params.version = YESPOWER_1_0;
-  yespower_params.N       = 2048;
-  yespower_params.r       = 32;
-  yespower_params.pers    = NULL;
-  yespower_params.perslen = 0;
+
+  if ( opt_param_n )  yespower_params.N = opt_param_n;
+  else                yespower_params.N = 2048;
+
+  if ( opt_param_r )  yespower_params.r = opt_param_r;
+  else                yespower_params.r = 32;
+
+  if ( opt_param_key )
+  {
+     yespower_params.pers = opt_param_key;
+     yespower_params.perslen = strlen( opt_param_key );
+  }
+  else
+  {
+     yespower_params.pers    = NULL;
+     yespower_params.perslen = 0;
+  }
+
+  applog( LOG_NOTICE,"Yespower parameters: N= %d, R= %d.", yespower_params.N,
+                                                           yespower_params.r );
+  if ( yespower_params.pers )
+     applog( LOG_NOTICE,"Key= \"%s\"\n", yespower_params.pers );
+
  gate->optimizations = SSE2_OPT;
-  gate->get_max64     = (void*)&yespower_get_max64;
  gate->scanhash      = (void*)&scanhash_yespower;
  gate->hash          = (void*)&yespower_hash;
-  gate->set_target    = (void*)&scrypt_set_target;
+  opt_target_factor = 65536.0;
  return true;
 };

@@ -98,35 +150,23 @@ bool register_yespowerr16_algo( algo_gate_t* gate )
  yespower_params.pers    = NULL;
  yespower_params.perslen = 0;
  gate->optimizations = SSE2_OPT;
-  gate->get_max64     = (void*)&yespower_get_max64;
  gate->scanhash      = (void*)&scanhash_yespower;
  gate->hash          = (void*)&yespower_hash;
-  gate->set_target    = (void*)&scrypt_set_target;
+  opt_target_factor = 65536.0;
  return true;
 };


-int64_t yescrypt_05_get_max64()
-{
-  return 0x1ffLL;
-}
-
-int64_t yescryptr16_05_get_max64()
-{
-  return 0xfffLL;
-}
-
 bool register_yescrypt_05_algo( algo_gate_t* gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yespower;
-   gate->set_target = (void*)&scrypt_set_target;
-   gate->get_max64  = (void*)&yescrypt_05_get_max64;
   yespower_params.version = YESPOWER_0_5;
   yespower_params.N       = 2048;
   yespower_params.r       = 8;
   yespower_params.pers    = NULL;
   yespower_params.perslen = 0;
+   opt_target_factor = 65536.0;
   return true;
 }

@@ -134,13 +174,12 @@ bool register_yescryptr8_05_algo( algo_gate_t* gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yespower;
-   gate->set_target = (void*)&scrypt_set_target;
-   gate->get_max64  = (void*)&yescrypt_05_get_max64;
   yespower_params.version = YESPOWER_0_5;
   yespower_params.N       = 2048;
   yespower_params.r       = 8;
   yespower_params.pers    = "Client Key";
   yespower_params.perslen = 10;
+   opt_target_factor = 65536.0;
   return true;
 }

@@ -148,13 +187,12 @@ bool register_yescryptr16_05_algo( algo_gate_t* gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yespower;
-   gate->set_target = (void*)&scrypt_set_target;
-   gate->get_max64  = (void*)&yescryptr16_05_get_max64;
   yespower_params.version = YESPOWER_0_5;
   yespower_params.N       = 4096;
   yespower_params.r       = 16;
   yespower_params.pers    = NULL;
   yespower_params.perslen = 0;
+   opt_target_factor = 65536.0;
   return true;
 }

@@ -162,13 +200,73 @@ bool register_yescryptr32_05_algo( algo_gate_t* gate )
 {
   gate->optimizations = SSE2_OPT | SHA_OPT;
   gate->scanhash   = (void*)&scanhash_yespower;
-   gate->set_target = (void*)&scrypt_set_target;
-   gate->get_max64  = (void*)&yescryptr16_05_get_max64;
   yespower_params.version = YESPOWER_0_5;
   yespower_params.N       = 4096;
   yespower_params.r       = 32;
   yespower_params.pers    = "WaviBanana";
   yespower_params.perslen = 10;
+   opt_target_factor = 65536.0;
   return true;
 }

+bool register_power2b_algo( algo_gate_t* gate )
+{
+  yespower_params.version = YESPOWER_1_0;
+
+  yespower_params.N = 2048;
+  yespower_params.r = 32;
+  yespower_params.pers = "Now I am become Death, the destroyer of worlds";
+  yespower_params.perslen = 46;
+
+  applog( LOG_NOTICE,"yespower-b2b parameters: N= %d, R= %d.", yespower_params.N,
+                                                           yespower_params.r );
+  applog( LOG_NOTICE,"Key= \"%s\"", yespower_params.pers );
+  applog( LOG_NOTICE,"Key length= %d\n", yespower_params.perslen );
+
+  gate->optimizations = SSE2_OPT;
+  gate->scanhash      = (void*)&scanhash_yespower_b2b;
+  gate->hash          = (void*)&yespower_b2b_hash;
+  opt_target_factor = 65536.0;
+  return true;
+};
+
+// Generic yespower + blake2b
+bool register_yespower_b2b_algo( algo_gate_t* gate )
+{
+  yespower_params.version = YESPOWER_1_0;
+
+  if ( !( opt_param_n && opt_param_r ) )
+  {
+     applog(LOG_ERR,"Yespower-b2b N & R parameters are required");
+     return false;
+  }
+
+  yespower_params.N = opt_param_n;
+  yespower_params.r = opt_param_r;
+
+  if ( opt_param_key )
+  {
+     yespower_params.pers = opt_param_key;
+     yespower_params.perslen = strlen( opt_param_key );
+  }
+  else
+  {
+     yespower_params.pers    = NULL;
+     yespower_params.perslen = 0;
+  }
+
+  applog( LOG_NOTICE,"Yespower-b2b parameters: N= %d, R= %d",
+                       yespower_params.N, yespower_params.r );
+  if ( yespower_params.pers )
+  {
+     applog( LOG_NOTICE,"Key= \"%s\"", yespower_params.pers );
+     applog( LOG_NOTICE,"Key length= %d\n", yespower_params.perslen );
+  }  
+
+  gate->optimizations = SSE2_OPT;
+  gate->scanhash      = (void*)&scanhash_yespower_b2b;
+  gate->hash          = (void*)&yespower_b2b_hash;
+  opt_target_factor = 65536.0;
+  return true;
+};
+
--- a/algo/yespower/yespower.h
+++ b/algo/yespower/yespower.h
@@ -111,6 +111,10 @@ extern int yespower(yespower_local_t *local,
    const uint8_t *src, size_t srclen,
    const yespower_params_t *params, yespower_binary_t *dst);

+extern int yespower_b2b(yespower_local_t *local,
+    const uint8_t *src, size_t srclen,
+    const yespower_params_t *params, yespower_binary_t *dst);
+
 /**
 * yespower_tls(src, srclen, params, dst):
 * Compute yespower(src[0 .. srclen - 1], N, r), to be checked for "< target".
@@ -123,6 +127,9 @@ extern int yespower(yespower_local_t *local,
 extern int yespower_tls(const uint8_t *src, size_t srclen,
    const yespower_params_t *params, yespower_binary_t *dst);

+extern int yespower_b2b_tls(const uint8_t *src, size_t srclen,
+    const yespower_params_t *params, yespower_binary_t *dst);
+
 #ifdef __cplusplus
 }
 #endif
--- a/api.c
+++ b/api.c
@@ -32,7 +32,7 @@
 #include <sys/types.h>

 #include "miner.h"
-
+#include "sysinfos.c"
 #ifndef WIN32
 # include <errno.h>
 # include <sys/socket.h>
@@ -105,7 +105,7 @@ extern double global_hashrate;
 #define USE_MONITORING
 extern float cpu_temp(int);
 extern uint32_t cpu_clock(int);
-extern int cpu_fanpercent(void);
+//extern int cpu_fanpercent(void);

 /***************************************************************/

--- a/20
+++ b/20
@@ -1,6 +1,6 @@
 #! /bin/sh
 # Guess values for system-dependent variables and create Makefiles.
-# Generated by GNU Autoconf 2.69 for cpuminer-opt 3.9.6.2.
+# Generated by GNU Autoconf 2.69 for cpuminer-opt 3.9.10.
 #
 #
 # Copyright (C) 1992-1996, 1998-2012 Free Software Foundation, Inc.
@@ -577,8 +577,8 @@ MAKEFLAGS=
 # Identity of this package.
 PACKAGE_NAME='cpuminer-opt'
 PACKAGE_TARNAME='cpuminer-opt'
-PACKAGE_VERSION='3.9.6.2'
-PACKAGE_STRING='cpuminer-opt 3.9.6.2'
+PACKAGE_VERSION='3.9.10'
+PACKAGE_STRING='cpuminer-opt 3.9.10'
 PACKAGE_BUGREPORT=''
 PACKAGE_URL=''

@@ -1332,7 +1332,7 @@ if test "$ac_init_help" = "long"; then
  # Omit some internal or obsolete options to make the list less imposing.
  # This message is too long to be a string in the A/UX 3.1 sh.
  cat <<_ACEOF
-\`configure' configures cpuminer-opt 3.9.6.2 to adapt to many kinds of systems.
+\`configure' configures cpuminer-opt 3.9.10 to adapt to many kinds of systems.

 Usage: $0 [OPTION]... [VAR=VALUE]...

@@ -1404,7 +1404,7 @@ fi

 if test -n "$ac_init_help"; then
  case $ac_init_help in
-     short | recursive ) echo "Configuration of cpuminer-opt 3.9.6.2:";;
+     short | recursive ) echo "Configuration of cpuminer-opt 3.9.10:";;
   esac
  cat <<\_ACEOF

@@ -1509,7 +1509,7 @@ fi
 test -n "$ac_init_help" && exit $ac_status
 if $ac_init_version; then
  cat <<\_ACEOF
-cpuminer-opt configure 3.9.6.2
+cpuminer-opt configure 3.9.10
 generated by GNU Autoconf 2.69

 Copyright (C) 2012 Free Software Foundation, Inc.
@@ -2012,7 +2012,7 @@ cat >config.log <<_ACEOF
 This file contains any messages produced by compilers while
 running configure, to aid debugging if configure makes a mistake.

-It was created by cpuminer-opt $as_me 3.9.6.2, which was
+It was created by cpuminer-opt $as_me 3.9.10, which was
 generated by GNU Autoconf 2.69.  Invocation command line was

  $ $0 $@
@@ -2993,7 +2993,7 @@ fi

 # Define the identity of the package.
 PACKAGE='cpuminer-opt'
- VERSION='3.9.6.2'
+ VERSION='3.9.10'


 cat >>confdefs.h <<_ACEOF
@@ -6690,7 +6690,7 @@ cat >>$CONFIG_STATUS <<\_ACEOF || ac_write_fail=1
 # report actual input values of CONFIG_FILES etc. instead of their
 # values after options handling.
 ac_log="
-This file was extended by cpuminer-opt $as_me 3.9.6.2, which was
+This file was extended by cpuminer-opt $as_me 3.9.10, which was
 generated by GNU Autoconf 2.69.  Invocation command line was

  CONFIG_FILES    = $CONFIG_FILES
@@ -6756,7 +6756,7 @@ _ACEOF
 cat >>$CONFIG_STATUS <<_ACEOF || ac_write_fail=1
 ac_cs_config="`$as_echo "$ac_configure_args" | sed 's/^ //; s/[\\""\`\$]/\\\\&/g'`"
 ac_cs_version="\\
-cpuminer-opt config.status 3.9.6.2
+cpuminer-opt config.status 3.9.10
 configured by $0, generated by GNU Autoconf 2.69,
  with options \\"\$ac_cs_config\\"

--- a/configure.ac
+++ b/configure.ac
@@ -1,4 +1,4 @@
-AC_INIT([cpuminer-opt], [3.9.6.2])
+AC_INIT([cpuminer-opt], [3.9.10])

 AC_PREREQ([2.59c])
 AC_CANONICAL_SYSTEM
--- a/cpu-miner.c
+++ b/cpu-miner.c
--- a/miner.h
+++ b/miner.h
@@ -310,6 +310,7 @@ struct thr_api {
 #define CL_WHT  "\x1B[01;37m" /* white */

 void   applog(int prio, const char *fmt, ...);
+void   applog2(int prio, const char *fmt, ...);
 void   restart_threads(void);
 extern json_t *json_rpc_call( CURL *curl, const char *url, const char *userpass,
                	const char *rpc_req, int *curl_err, int flags );
@@ -331,8 +332,27 @@ extern void diff_to_target(uint32_t *target, double diff);
 double hash_target_ratio( uint32_t* hash, uint32_t* target );
 void   work_set_target_ratio( struct work* work, uint32_t* hash );

+struct thr_info {
+        int id;
+        pthread_t pth;
+        pthread_attr_t attr;
+        struct thread_q *q;
+        struct cpu_info cpu;
+};
+
+//struct thr_info *thr_info;
+
+bool   submit_solution( struct work *work, void *hash,
+                        struct thr_info *thr );
+bool   submit_lane_solution( struct work *work, void *hash,
+                             struct thr_info *thr, int lane );
+
+
+//bool submit_work( struct thr_info *thr, const struct work *work_in );
+

 void   get_currentalgo( char* buf, int sz );
+/*
 bool   has_sha();
 bool   has_aes_ni();
 bool   has_avx1();
@@ -349,13 +369,14 @@ void   cpu_getmodelid(char *outbuf, size_t maxsz);
 void   cpu_brand_string( char* s );

 float cpu_temp( int core );
+*/

 struct work {
 	uint32_t data[48] __attribute__ ((aligned (64)));
 	uint32_t target[8];

 	double targetdiff;
-	double shareratio;
+//	double shareratio;
 	double sharediff;

 	int height;
@@ -420,7 +441,7 @@ struct stratum_ctx {
 	struct work work __attribute__ ((aligned (64)));
 	pthread_mutex_t work_lock;

-	int bloc_height;
+	int block_height;
 } __attribute__ ((aligned (64)));

 bool stratum_socket_full(struct stratum_ctx *sctx, int timeout);
@@ -471,6 +492,9 @@ void print_hash_tests(void);

 void scale_hash_for_display ( double* hashrate, char* units );

+void report_summary_log( bool force );
+
+/*
 struct thr_info {
        int id;
        pthread_t pth;
@@ -478,6 +502,7 @@ struct thr_info {
        struct thread_q *q;
        struct cpu_info cpu;
 };
+*/

 struct work_restart {
        volatile uint8_t restart;
@@ -549,6 +574,7 @@ enum algos {
 	     ALGO_PHI2,
        ALGO_PLUCK,       
        ALGO_POLYTIMOS,
+        ALGO_POWER2B,
        ALGO_QUARK,
        ALGO_QUBIT,       
        ALGO_SCRYPT,
@@ -578,6 +604,7 @@ enum algos {
        ALGO_X14,        
        ALGO_X15,       
        ALGO_X16R,
+        ALGO_X16RV2,
        ALGO_X16RT,
        ALGO_X16RT_VEIL,
        ALGO_X16S,
@@ -590,6 +617,7 @@ enum algos {
        ALGO_YESCRYPTR32,
        ALGO_YESPOWER,
        ALGO_YESPOWERR16,
+        ALGO_YESPOWER_B2B,
        ALGO_ZR5,
        ALGO_COUNT
 };
@@ -643,6 +671,7 @@ static const char* const algo_names[] = {
        "phi2",
 	     "pluck",
        "polytimos",
+        "power2b",
        "quark",
        "qubit",
        "scrypt",
@@ -672,6 +701,7 @@ static const char* const algo_names[] = {
        "x14",
        "x15",
        "x16r",
+        "x16rv2",
        "x16rt",
        "x16rt-veil",
        "x16s",
@@ -684,6 +714,7 @@ static const char* const algo_names[] = {
        "yescryptr32",
        "yespower",
        "yespowerr16",
+        "yespower-b2b",
        "zr5",
        "\0"
 };
@@ -695,7 +726,6 @@ extern bool opt_debug;
 extern bool opt_debug_diff;
 extern bool opt_benchmark;
 extern bool opt_protocol;
-extern bool opt_showdiff;
 extern bool opt_extranonce;
 extern bool opt_quiet;
 extern bool opt_redirect;
@@ -726,11 +756,15 @@ extern uint32_t opt_work_size;
 extern double *thr_hashrates;
 extern double global_hashrate;
 extern double stratum_diff;
+extern bool opt_reset_on_stale;
 extern double net_diff;
 extern double net_hashrate;
 extern int opt_pluck_n;
-extern int opt_scrypt_n;
+extern int opt_param_n;
+extern int opt_param_r;
+extern char* opt_param_key;
 extern double opt_diff_factor;
+extern double opt_target_factor;
 extern bool opt_randomize;
 extern bool allow_mininginfo;
 extern time_t g_work_time;
@@ -796,10 +830,11 @@ Options:\n\
                          neoscrypt     NeoScrypt(128, 2, 1)\n\
                          nist5         Nist5\n\
                          pentablake    5 x blake512\n\
-                          phi1612       phi, LUX coin (original algo)\n\
-                          phi2          LUX (new algo)\n\
+                          phi1612       phi\n\
+                          phi2\n\
 			                 pluck         Pluck:128 (Supcoin)\n\
                          polytimos\n\
+                          power2b       MicroBitcoin (MBC)\n\
                          quark         Quark\n\
                          qubit         Qubit\n\
                          scrypt        scrypt(1024, 1, 1) (default)\n\
@@ -829,10 +864,11 @@ Options:\n\
                          x13sm3        hsr (Hshare)\n\
                          x14           X14\n\
                          x15           X15\n\
-                          x16r          Ravencoin (RVN)\n\
+                          x16r\n\
+                          x16rv2        Ravencoin (RVN)\n\
                          x16rt         Gincoin (GIN)\n\
                          x16rt-veil    Veil (VEIL)\n\
-                          x16s          Pigeoncoin (PGN)\n\
+                          x16s\n\
                          x17\n\
                          x21s\n\
                          xevan         Bitsend (BSD)\n\
@@ -842,7 +878,11 @@ Options:\n\
                          yescryptr32   WAVI\n\
                          yespower      Cryply\n\
                          yespowerr16   Yenten (YTN)\n\
+                          yespower-b2b  generic yespower + blake2b\n\
                          zr5           Ziftr\n\
+  -N, --param-n         N parameter for scrypt based algos\n\
+  -R, --patam-r         R parameter for scrypt based algos\n\
+  -K, --param-key       Key parameter for algos that use it\n\
  -o, --url=URL         URL of mining server\n\
  -O, --userpass=U:P    username:password pair for mining server\n\
  -u, --user=USERNAME   username for mining server\n\
@@ -852,16 +892,16 @@ Options:\n\
  -t, --threads=N       number of miner threads (default: number of processors)\n\
  -r, --retries=N       number of times to retry if a network call fails\n\
                          (default: retry indefinitely)\n\
-  -R, --retry-pause=N   time to pause between retries, in seconds (default: 30)\n\
+      --retry-pause=N   time to pause between retries, in seconds (default: 30)\n\
      --time-limit=N    maximum time [s] to mine before exiting the program.\n\
  -T, --timeout=N       timeout for long poll and stratum (default: 300 seconds)\n\
  -s, --scantime=N      upper bound on time spent scanning current work when\n\
                          long polling is unavailable, in seconds (default: 5)\n\
      --randomize       Randomize scan range start to reduce duplicates\n\
-  -f, --diff-factor     Divide req. difficulty by this factor (std is 1.0)\n\
+      --reset-on-stale  Workaround reset stratum if too many stale shares\n\
+      -f, --diff-factor     Divide req. difficulty by this factor (std is 1.0)\n\
  -m, --diff-multiplier Multiply difficulty by this factor (std is 1.0)\n\
      --hash-meter      Display thread hash rates\n\
-      --hide-diff       Do not display changes in difficulty\n\
      --coinbase-addr=ADDR  payout address for solo mining\n\
      --coinbase-sig=TEXT  data to insert in the coinbase when possible\n\
      --no-longpoll     disable long polling support\n\
@@ -925,8 +965,8 @@ static struct option const options[] = {
        { "diff", 1, NULL, 'f' }, // deprecated (alias)
        { "diff-multiplier", 1, NULL, 'm' },
        { "hash-meter", 0, NULL, 1014 },
-        { "hide-diff", 0, NULL, 1013 },
        { "help", 0, NULL, 'h' },
+        { "key", 1, NULL, 'K' },
        { "no-gbt", 0, NULL, 1011 },
        { "no-getwork", 0, NULL, 1010 },
        { "no-longpoll", 0, NULL, 1003 },
@@ -936,14 +976,18 @@ static struct option const options[] = {
        { "max-temp", 1, NULL, 1060 },
        { "max-diff", 1, NULL, 1061 },
        { "max-rate", 1, NULL, 1062 },
+        { "param-key", 1, NULL, 'K' },
+        { "param-n", 1, NULL, 'N' },
+        { "param-r", 1, NULL, 'R' },
        { "pass", 1, NULL, 'p' },
        { "protocol", 0, NULL, 'P' },
        { "protocol-dump", 0, NULL, 'P' },
        { "proxy", 1, NULL, 'x' },
        { "quiet", 0, NULL, 'q' },
        { "retries", 1, NULL, 'r' },
-        { "retry-pause", 1, NULL, 'R' },
+        { "retry-pause", 1, NULL, 1025 },
        { "randomize", 0, NULL, 1024 },
+        { "reset-on-stale", 0, NULL, 1026 },
        { "scantime", 1, NULL, 's' },
 #ifdef HAVE_SYSLOG_H
        { "syslog", 0, NULL, 'S' },
--- a/simd-utils.h
+++ b/simd-utils.h
@@ -36,7 +36,7 @@
 //    MMX:     64 bit vectors  
 //    SSE2:   128 bit vectors  (64 bit CPUs only, such as Intel Core2.
 //    AVX2:   256 bit vectors  (Starting with Intel Haswell and AMD Ryzen)
-//    AVX512: 512 bit vectors  (still under development)
+//    AVX512: 512 bit vectors  (Starting with SkylakeX)
 //
 //    Most functions are avalaible at the stated levels but in rare cases
 //    a higher level feature may be required with no compatible alternative.
@@ -105,59 +105,36 @@
 //    Ex: mm256_ror1x64_128 rotates each 128 bit lane of a 256 bit vector
 //        right by 64 bits.
 //
-//   Some random thoughts about macros and inline functions, the pros and
-//   cons, when to use them, etc:
+// Vector constants
 //
-// Macros are very convenient and efficient for statement functions.
-// Macro args are passed by value and modifications are seen by the caller.
-// Macros should not generally call regular functions unless it is for a
-// special purpose such overloading a function name.
-// Statement function macros that return a value should not end in ";"
-// Statement function macros that return a value and don't modify input args
-// may be used in function arguments and expressions.
-// Macro args used in expressions should be protected ex: (x)+1
-// Macros force inlining, function inlining can be overridden by the compiler.
-// Inline functions are preferred when multiple statements or local variables
-// are needed.
-// The compiler can't do any syntax checking or type checking of args making
-// macros difficult to debug.
-// Although it is technically posssible to access the callers data without
-// they being passed as arguments it is good practice to always define
-// arguments even if they have the same name.
+// Vector constants are a big problem because they technically don't exist.
+// All vectors used as constants either reside in memory or must be genererated
+// at run time at significant cost. The cost of generating a constant
+// increases non-linearly with the number of vector elements. A 4 element
+// vector costs between 7 and 11 clocks to generate, an 8 element vector
+// is 15-25 clocks. There are also additional clock due to data dependency
+// stalls.
 //
-// General guidelines for inline functions:
+// Vector constants are often used as control indexes for permute, blend, etc,
+// where generating the index can be over 90% of the operation. This is
+// where the problem occurs. An instruction that only requires one to 3
+// clocks needs may times more just to build the index argument.
 //
-// Inline functions should not have loops, it defeats the purpose of inlining.
-// Inline functions should be short, the benefit is lost and the memory cost
-// increases if the function is referenced often.
-// Inline functions may call other functions, inlined or not. It is convenient
-// for wrapper functions whether or not the wrapped function is itself inlined.
-// Care should be taken when unrolling loops that contain calls to inlined
-// functions that may be large.
-// Large code blocks used only once may use function inlining to
-// improve high level code readability without the penalty of function
-// overhead.
+// There is very little a programmer can do to avoid the worst case scenarios.
+// Smaller integers can be merged to form 64 bit integers, and vectors with
+// repeated elements can be generated more efficiently but they have limited
+// benefit and limited application.
 //
-// A major restructuring is taking place shifting the focus from pointers
-// to registers. Previously pointer casting used memory to provide transparency
-// leaving it up to the compiler to manage everything and it does a very good
-// job. The focus has shifted to register arguments for more control
-// over the actual instructions assuming the data is in a register and the
-// the compiler just needs to manage the registers.
+// If a vector constant is to be used repeatedly it is better to define a local
+// variable to generate the constant only once.
 //
-// Rather than use pointers to provide type transparency
-// specific instructions are used to access specific data as specific types.
-// Previously pointers were cast and the compiler was left to find a way
-// to get the data from wherever it happened to be to the correct registers.
+// If a sequence of constants is to be used it can be more efficient to
+// use arithmetic with already existing constants to generate new ones.
 //
-// The utilities defined here make use features like register aliasing
-// to optimize operations. Many operations have specialized versions as
-// well as more generic versions. It is preferable to use a specialized
-// version whenever possible a sthey can take advantage of certain
-// optimizations not available to the generic version. Specically the generic
-// version usually has a second argument used is some extra calculations.
-// 
-///////////////////////////////////////////////////////
+// ex: const __m512i one = _mm512_const1_64( 1 );
+//     const __m512i two = _mm512_add_epi64( one, one );
+//     
+//////////////////////////////////////////////////////////////////////////

 #include <inttypes.h>
 #include <x86intrin.h>
@@ -165,9 +142,6 @@
 #include <stdlib.h>
 #include <stdbool.h>

-// Various types and overlays
-#include "simd-utils/simd-types.h"
-
 // 64 and 128 bit integers.
 #include "simd-utils/simd-int.h"

@@ -175,7 +149,6 @@

 // 64 bit vectors
 #include "simd-utils/simd-64.h"
-//#include "simd-utils/intrlv-mmx.h"

 #if defined(__SSE2__)

@@ -189,17 +162,19 @@

 #if defined(__AVX2__)

+// Utilities that require AVX2 are defined in simd-256.h.
+
 // Skylake-X has all these
-#if defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
+#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)

 // 512 bit vectors
 #include "simd-utils/simd-512.h"

-#endif  // MMX
-#endif  // SSE2
-#endif  // AVX
-#endif  // AVX2
 #endif  // AVX512
+#endif  // AVX2
+#endif  // AVX
+#endif  // SSE2
+#endif  // MMX

 #include "simd-utils/intrlv.h"

--- a/simd-utils/intrlv.h
+++ b/simd-utils/intrlv.h
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
Jay D Dee	a52c5eccf7	v3.9.10	2019-11-22 20:29:18 -05:00
Jay D Dee	86b889e1b0	v3.9.9.1	2019-10-24 14:11:26 -04:00
Jay D Dee	72330eb5a7	v3.9.9	2019-10-10 19:58:34 -04:00
Jay D Dee	789c8b70bc	v3.9.8.1	2019-10-01 14:17:36 -04:00
Jay D Dee	01550d94a2	v3.9.8	2019-09-26 22:37:26 -04:00
Jay D Dee	a042fb7612	v3.9.7	2019-08-03 10:39:54 -04:00
				`@@ -0,0 +1 @@`
				`#define insecure_memzero(buf, len) /* empty */`