popov/cpuminer-opt-gpu
1
0
Fork 0
mirror of https://github.com/JayDDee/cpuminer-opt.git synced 2025-09-17 23:44:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

v3.15.2

Browse Source
This commit is contained in:
Jay D Dee
2020-11-15 17:57:06 -05:00
parent 4fa8fcea8b
commit 45ecd0de14
42 changed files with 2490 additions and 445 deletions
Download Patch File Download Diff File Expand all files Collapse all files

193
algo/x17/sonoa-4way.c
View File

@@ -1124,7 +1124,13 @@ union _sonoa_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
#if defined(__VAES__)
groestl512_2way_context groestl;
echo512_2way_context echo;
#else
hashState_groestl groestl;
hashState_echo echo;
#endif
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
@@ -1132,7 +1138,6 @@ union _sonoa_4way_context_overlay
cube_2way_context cube;
shavite512_2way_context shavite;
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
hashState_fugue fugue;
shabal512_4way_context shabal;
@@ -1162,6 +1167,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1171,6 +1187,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1195,6 +1213,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1206,16 +1233,29 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
(const BitSequence *)hash2, 64 );
echo_full( &ctx.echo, (BitSequence *)hash3, 512,
(const BitSequence *)hash3, 64 );
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
if ( work_restart[thr_id].restart ) return 0;
// 2
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1225,6 +1265,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1249,6 +1291,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1263,6 +1314,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1274,6 +1327,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1283,6 +1347,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1307,6 +1373,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1321,6 +1396,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1340,6 +1417,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1349,6 +1437,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1373,6 +1463,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1387,6 +1486,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1410,6 +1511,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
hamsi512_4way_update( &ctx.hamsi, vhashB, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
echo_full( &ctx.echo, (BitSequence *)hash0, 512,
@@ -1424,6 +1534,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_2x128_512( vhashA, hash0, hash1 );
intrlv_2x128_512( vhashB, hash2, hash3 );
#endif
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
@@ -1443,6 +1555,20 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
shabal512_4way_update( &ctx.shabal, vhashB, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
#if defined(__VAES__)
// rintrlv_4x32_2x128( vhashA, vhashB, vhash, 512 );
dintrlv_4x32_512( hash0, hash1, hash2, hash3, vhash );
intrlv_2x128_512( vhashA, hash0, hash1 );
intrlv_2x128_512( vhashB, hash2, hash3 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x32_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1452,6 +1578,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1476,6 +1604,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1490,6 +1627,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1523,6 +1662,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1532,6 +1682,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1556,6 +1708,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1570,6 +1731,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -1616,6 +1779,17 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -1625,6 +1799,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -1649,6 +1825,15 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -1663,6 +1848,8 @@ int sonoa_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );

2
algo/x17/sonoa-gate.c
View File

@@ -12,7 +12,7 @@ bool register_sonoa_algo( algo_gate_t* gate )
init_sonoa_ctx();
gate->hash = (void*)&sonoa_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
return true;
};

31
algo/x17/x17-4way.c
View File

@@ -240,7 +240,13 @@ union _x17_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
#if defined(__VAES__)
groestl512_2way_context groestl;
echo512_2way_context echo;
#else
hashState_groestl groestl;
hashState_echo echo;
#endif
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
@@ -248,7 +254,6 @@ union _x17_4way_context_overlay
cube_2way_context cube;
shavite512_2way_context shavite;
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
hashState_fugue fugue;
shabal512_4way_context shabal;
@@ -275,6 +280,17 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, 64 );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -284,6 +300,8 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, 64 );
jh512_4way_init( &ctx.jh );
@@ -308,6 +326,15 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, 64 );
simd512_2way_full( &ctx.simd, vhashB, vhashB, 64 );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, 64 );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, 64 );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_2x128_512( hash0, hash1, vhashA );
dintrlv_2x128_512( hash2, hash3, vhashB );
@@ -322,6 +349,8 @@ int x17_4way_hash( void *state, const void *input, int thr_id )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );

2
algo/x17/x17-gate.c
View File

@@ -11,7 +11,7 @@ bool register_x17_algo( algo_gate_t* gate )
#else
gate->hash = (void*)&x17_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
return true;
};

62
algo/x17/xevan-4way.c
View File

@@ -405,15 +405,20 @@ union _xevan_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
#if defined(__VAES__)
groestl512_2way_context groestl;
echo_2way_context echo;
#else
hashState_groestl groestl;
hashState_echo echo;
#endif
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
luffa_2way_context luffa;
cube_2way_context cube;
shavite512_2way_context shavite;
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
hashState_fugue fugue;
shabal512_4way_context shabal;
@@ -442,7 +447,17 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, dataLen );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, dataLen );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, dataLen );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
#else
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, dataLen<<3 );
@@ -450,9 +465,10 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, dataLen<<3 );
groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, dataLen<<3 );
// Parallel 4way
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, dataLen );
jh512_4way_init( &ctx.jh );
@@ -477,6 +493,15 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, dataLen );
simd512_2way_full( &ctx.simd, vhashB, vhashB, dataLen );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, dataLen );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, dataLen );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
#else
dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );
@@ -489,9 +514,10 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
echo_full( &ctx.echo, (BitSequence *)hash3, 512,
(const BitSequence *)hash3, dataLen );
// Parallel
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, dataLen );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -542,6 +568,17 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
bmw512_4way_update( &ctx.bmw, vhash, dataLen );
bmw512_4way_close( &ctx.bmw, vhash );
#if defined(__VAES__)
rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
groestl512_2way_full( &ctx.groestl, vhashA, vhashA, dataLen );
groestl512_2way_full( &ctx.groestl, vhashB, vhashB, dataLen );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
#else
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, dataLen<<3 );
@@ -551,6 +588,8 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
#endif
skein512_4way_full( &ctx.skein, vhash, vhash, dataLen );
jh512_4way_init( &ctx.jh );
@@ -575,6 +614,15 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
simd512_2way_full( &ctx.simd, vhashA, vhashA, dataLen );
simd512_2way_full( &ctx.simd, vhashB, vhashB, dataLen );
#if defined(__VAES__)
echo_2way_full( &ctx.echo, vhashA, 512, vhashA, dataLen );
echo_2way_full( &ctx.echo, vhashB, 512, vhashB, dataLen );
rintrlv_2x128_4x64( vhash, vhashA, vhashB, dataLen<<3 );
#else
dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );
@@ -589,6 +637,8 @@ int xevan_4way_hash( void *output, const void *input, int thr_id )
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
#endif
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, dataLen );
hamsi512_4way_close( &ctx.hamsi, vhash );

2
algo/x17/xevan-gate.c
View File

@@ -12,7 +12,7 @@ bool register_xevan_algo( algo_gate_t* gate )
init_xevan_ctx();
gate->hash = (void*)&xevan_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT | VAES256_OPT;
opt_target_factor = 256.0;
return true;
};