v3.9.6

algo/x16/x16r-gate.c

@@ -62,3 +62,149 @@ bool register_x16s_algo( algo_gate_t* gate )
   return true;
 };
 
+////////////////
+//
+//   X16RT
+
+
+void x16rt_getTimeHash( const uint32_t timeStamp, void* timeHash )
+{
+    int32_t maskedTime = timeStamp & 0xffffff80;
+    sha256d( (unsigned char*)timeHash, (const unsigned char*)( &maskedTime ),
+             sizeof( maskedTime ) );
+}
+
+void x16rt_getAlgoString( const uint32_t *timeHash, char *output)
+{
+   char *sptr = output;
+   uint8_t* data = (uint8_t*)timeHash;
+
+   for (uint8_t j = 0; j < X16R_HASH_FUNC_COUNT; j++) {
+      uint8_t b = (15 - j) >> 1; // 16 ascii hex chars, reversed
+      uint8_t algoDigit = (j & 1) ? data[b] & 0xF : data[b] >> 4;
+
+      if (algoDigit >= 10)
+         sprintf(sptr, "%c", 'A' + (algoDigit - 10));
+      else
+         sprintf(sptr, "%u", (uint32_t) algoDigit);
+      sptr++;
+   }
+   *sptr = '\0';
+}
+
+void x16rt_build_extraheader( struct work* g_work, struct stratum_ctx* sctx )
+{
+   uchar merkle_tree[64] = { 0 };
+   size_t t;
+
+   algo_gate.gen_merkle_root( merkle_tree, sctx );
+   // Increment extranonce2
+   for ( t = 0; t < sctx->xnonce2_size && !( ++sctx->job.xnonce2[t] ); t++ );
+
+   // Assemble block header
+//   algo_gate.build_block_header( g_work, le32dec( sctx->job.version ),
+//          (uint32_t*) sctx->job.prevhash, (uint32_t*) merkle_tree,
+//          le32dec( sctx->job.ntime ), le32dec(sctx->job.nbits) );
+   int i;
+
+   memset( g_work->data, 0, sizeof(g_work->data) );
+   g_work->data[0] = le32dec( sctx->job.version );
+
+   if ( have_stratum )
+      for ( i = 0; i < 8; i++ )
+         g_work->data[ 1+i ] = le32dec( (uint32_t*)sctx->job.prevhash + i );
+   else
+      for (i = 0; i < 8; i++)
+         g_work->data[ 8-i ] = le32dec( (uint32_t*)sctx->job.prevhash + i );
+
+   g_work->data[ algo_gate.ntime_index ] = le32dec( sctx->job.ntime );
+   g_work->data[ algo_gate.nbits_index ] = le32dec( sctx->job.nbits );
+   g_work->data[20] = 0x80000000;
+   g_work->data[31] = 0x00000280;
+
+   for ( i = 0; i < 8; i++ )
+      g_work->merkleroothash[7 - i] = be32dec((uint32_t *)merkle_tree + i);
+   for ( i = 0; i < 8; i++ )
+      g_work->witmerkleroothash[7 - i] = be32dec((uint32_t *)merkle_tree + i);
+   for ( i = 0; i < 8; i++ )
+      g_work->denom10[i] =    le32dec((uint32_t *)sctx->job.denom10 + i);
+   for ( i = 0; i < 8; i++ )
+      g_work->denom100[i] =   le32dec((uint32_t *)sctx->job.denom100 + i);
+   for ( i = 0; i < 8; i++ )
+      g_work->denom1000[i] =  le32dec((uint32_t *)sctx->job.denom1000 + i);
+   for ( i = 0; i < 8; i++ )
+      g_work->denom10000[i] = le32dec((uint32_t *)sctx->job.denom10000 + i);
+
+   uint32_t pofnhash[8];
+   memset(pofnhash, 0x00, 32);
+
+   char denom10_str      [ 2 * sizeof( g_work->denom10 )           + 1 ];
+   char denom100_str     [ 2 * sizeof( g_work->denom100 )          + 1 ];
+   char denom1000_str    [ 2 * sizeof( g_work->denom1000 )         + 1 ];
+   char denom10000_str   [ 2 * sizeof( g_work->denom10000 )        + 1 ];
+   char merkleroot_str   [ 2 * sizeof( g_work->merkleroothash )    + 1 ];
+   char witmerkleroot_str[ 2 * sizeof( g_work->witmerkleroothash ) + 1 ];
+   char pofn_str         [ 2 * sizeof( pofnhash )                  + 1 ];
+
+   cbin2hex( denom10_str,       (char*) g_work->denom10,           32 );
+   cbin2hex( denom100_str,      (char*) g_work->denom100,          32 );
+   cbin2hex( denom1000_str,     (char*) g_work->denom1000,         32 );
+   cbin2hex( denom10000_str,    (char*) g_work->denom10000,        32 );
+   cbin2hex( merkleroot_str,    (char*) g_work->merkleroothash,    32 );
+   cbin2hex( witmerkleroot_str, (char*) g_work->witmerkleroothash, 32 );
+   cbin2hex( pofn_str,          (char*) pofnhash,                  32 );
+
+   if ( true )
+   {
+       char* data;
+       data = (char*)malloc( 2 + strlen( denom10_str ) * 4 + 16 * 4
+                             + strlen( merkleroot_str ) * 3 );
+       // Build the block header veildatahash in hex
+       sprintf( data, "%s%s%s%s%s%s%s%s%s%s%s%s",
+                       merkleroot_str, witmerkleroot_str, "04",
+                       "0a00000000000000", denom10_str,
+                       "6400000000000000", denom100_str,
+                       "e803000000000000", denom1000_str,
+                       "1027000000000000", denom10000_str, pofn_str );
+       // Covert the hex to binary
+       uint32_t test[100];
+       hex2bin( (unsigned char*)(&test), data, 257);
+       // Compute the sha256d of the binary
+       uint32_t _ALIGN(64) hash[8];
+       sha256d( (unsigned char*)hash, (unsigned char*)&(test), 257);
+       // assign the veildatahash in the blockheader
+       for ( i = 0; i < 8; i++ )
+           g_work->data[16 - i] = le32dec(hash + i);
+       free(data);
+    }
+}
+
+bool register_x16rt_algo( algo_gate_t* gate )
+{
+#if defined (X16R_4WAY)
+  gate->scanhash  = (void*)&scanhash_x16rt_4way;
+  gate->hash      = (void*)&x16rt_4way_hash;
+#else
+  gate->scanhash  = (void*)&scanhash_x16rt;
+  gate->hash      = (void*)&x16rt_hash;
+#endif
+  gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
+  gate->set_target = (void*)&alt_set_target;
+  return true;
+};
+
+bool register_x16rt_veil_algo( algo_gate_t* gate )
+{
+#if defined (X16R_4WAY)
+  gate->scanhash  = (void*)&scanhash_x16rt_4way;
+  gate->hash      = (void*)&x16rt_4way_hash;
+#else
+  gate->scanhash  = (void*)&scanhash_x16rt;
+  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*)&x16rt_build_extraheader;
+  return true;
+};
+

algo/x16/x16r-gate.h

@@ -4,6 +4,7 @@
 #include "algo-gate-api.h"
 #include "simd-utils.h"
 #include <stdint.h>
+#include <unistd.h>
 
 #if defined(__AVX2__) && defined(__AES__)
   #define X16R_4WAY
@@ -30,11 +31,15 @@ enum x16r_Algo {
 };
 
 void (*x16_r_s_getAlgoString) ( const uint8_t*, char* );
-void x16r_getAlgoString( const uint8_t* prevblock, char *output );
-void x16s_getAlgoString( const uint8_t* prevblock, char *output );
+void x16r_getAlgoString( const uint8_t *prevblock, char *output );
+void x16s_getAlgoString( const uint8_t *prevblock, char *output );
+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_x16s_algo( algo_gate_t* gate );
+bool register_x16rt_algo( algo_gate_t* gate );
 
 #if defined(X16R_4WAY)
 
@@ -42,11 +47,18 @@ 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 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 );
+
 #endif
 
 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 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 );
 #endif
 

algo/x16/x16rt-4way.c

@@ -0,0 +1,353 @@
+#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/sha2-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
+{
+    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;
+};
+typedef union _x16rt_4way_context_overlay x16rt_4way_context_overlay;
+
+void x16rt_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)));
+   x16rt_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 );
+ 
+/*
+   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.
+   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:
+            keccak512_4way_init( &ctx.keccak );
+            if ( i == 0 )
+               keccak512_4way( &ctx.keccak, input, size );
+            else
+            {
+               intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+               keccak512_4way( &ctx.keccak, vhash, size );
+            }
+            keccak512_4way_close( &ctx.keccak, vhash );
+            dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
+         break;
+         case LUFFA:
+            intrlv_2x128( vhash, in0, in1, size<<3 );
+            luffa_2way_init( &ctx.luffa, 512 );
+            luffa_2way_update_close( &ctx.luffa, vhash, vhash, size );
+            dintrlv_2x128( hash0, hash1, vhash, 512 );
+            intrlv_2x128( vhash, in2, in3, size<<3 );
+            luffa_2way_init( &ctx.luffa, 512 );
+            luffa_2way_update_close( &ctx.luffa, vhash, vhash, size);
+            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:
+             intrlv_4x64( vhash, in0, in1, in2, in3, size<<3 );
+             sha512_4way_init( &ctx.sha512 );
+             sha512_4way( &ctx.sha512, vhash, size );
+             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_x16rt_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 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
+    __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 ) );
+
+   uint32_t ntime = swab32( 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
+   {
+      *noncev = mm256_intrlv_blend_32( mm256_bswap_32(
+               _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
+
+      x16rt_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 )
+      {
+         pdata[19] = n+i;
+         submit_lane_solution( work, hash+(i<<3), mythr, i );
+      }
+      n += 4;
+   } while ( (  n < max_nonce ) && !(*restart) );
+
+   *hashes_done = n - first_nonce + 1;
+   return 0;
+}
+
+#endif

algo/x16/x16rt.c

@@ -0,0 +1,239 @@
+#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>
+#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 bool s_implemented = false;
+static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };
+
+union _x16rt_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;
+};
+typedef union _x16rt_context_overlay x16rt_context_overlay;
+
+void x16rt_hash( void* output, const void* input )
+{
+   uint32_t _ALIGN(128) hash[16];
+   x16rt_context_overlay ctx;
+   int size = 80;
+   void *in = (void*) input;
+
+/*
+   void *in = (void*) input;
+   uint32_t *in32 = (uint32_t*) in;
+   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);
+   }
+*/
+   
+   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_keccak512_init( &ctx.keccak );
+            sph_keccak512( &ctx.keccak, in, size );
+            sph_keccak512_close( &ctx.keccak, hash );
+         break;
+         case LUFFA:
+            init_luffa( &ctx.luffa, 512 );
+            update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
+                                    (const BitSequence*)in, size );
+         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:
+             SHA512_Init( &ctx.sha512 );
+             SHA512_Update( &ctx.sha512, in, size );
+             SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
+         break;
+      }
+      in = (void*) hash;
+      size = 64;
+   }
+   memcpy(output, hash, 32);
+}
+
+int scanhash_x16rt( 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 _ALIGN(64) timeHash[8];
+   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 ) );
+
+   uint32_t ntime = swab32( 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;
+
+   do
+   {
+      be32enc( &endiandata[19], nonce );
+      x16rt_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;
+}