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

33
algo/x16/x16r-4way.c
View File

@@ -619,11 +619,20 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
break;
case GROESTL:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
groestl512_full( &ctx.groestl, (char*)hash0, (char*)in0, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash1, (char*)in1, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash2, (char*)in2, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash3, (char*)in3, size<<3 );
break;
#endif
break;
case JH:
if ( i == 0 )
jh512_4way_update( &ctx.jh, input + (64<<2), 16 );
@@ -711,11 +720,20 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
}
break;
case SHAVITE:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
shavite512_full( &ctx.shavite, hash0, in0, size );
shavite512_full( &ctx.shavite, hash1, in1, size );
shavite512_full( &ctx.shavite, hash2, in2, size );
shavite512_full( &ctx.shavite, hash3, in3, size );
break;
#endif
break;
case SIMD:
intrlv_2x128( vhash, in0, in1, size<<3 );
simd512_2way_full( &ctx.simd, vhash, vhash, size );
@@ -725,6 +743,14 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
dintrlv_2x128_512( hash2, hash3, vhash );
break;
case ECHO:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
echo_full( &ctx.echo, (BitSequence *)hash0, 512,
(const BitSequence *)in0, size );
echo_full( &ctx.echo, (BitSequence *)hash1, 512,
@@ -733,7 +759,8 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
(const BitSequence *)in2, size );
echo_full( &ctx.echo, (BitSequence *)hash3, 512,
(const BitSequence *)in3, size );
break;
#endif
break;
case HAMSI:
if ( i == 0 )
hamsi512_4way_update( &ctx.hamsi, input + (64<<2), 16 );

24
algo/x16/x16r-gate.c
View File

@@ -61,7 +61,8 @@ bool register_x16r_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x16r;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
opt_target_factor = 256.0;
return true;
@@ -79,7 +80,8 @@ bool register_x16rv2_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x16rv2;
gate->hash = (void*)&x16rv2_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
opt_target_factor = 256.0;
return true;
@@ -97,7 +99,8 @@ bool register_x16s_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x16r;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
x16_r_s_getAlgoString = (void*)&x16s_getAlgoString;
opt_target_factor = 256.0;
return true;
@@ -230,7 +233,8 @@ bool register_x16rt_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x16rt;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
opt_target_factor = 256.0;
return true;
};
@@ -247,7 +251,8 @@ bool register_x16rt_veil_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x16rt;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
gate->build_extraheader = (void*)&veil_build_extraheader;
opt_target_factor = 256.0;
return true;
@@ -277,22 +282,17 @@ bool register_x21s_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x21s_8way;
gate->hash = (void*)&x21s_8way_hash;
gate->miner_thread_init = (void*)&x21s_8way_thread_init;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT
| VAES_OPT;
#elif defined (X16R_4WAY)
gate->scanhash = (void*)&scanhash_x21s_4way;
gate->hash = (void*)&x21s_4way_hash;
gate->miner_thread_init = (void*)&x21s_4way_thread_init;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
| AVX512_OPT | VAES_OPT;
#else
gate->scanhash = (void*)&scanhash_x21s;
gate->hash = (void*)&x21s_hash;
gate->miner_thread_init = (void*)&x21s_thread_init;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT
| AVX512_OPT | VAES_OPT;
#endif
// gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | SHA_OPT | AVX512_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT |
VAES_OPT | VAES256_OPT;
x16_r_s_getAlgoString = (void*)&x16s_getAlgoString;
opt_target_factor = 256.0;
return true;

11
algo/x16/x16r-gate.h
View File

@@ -41,6 +41,7 @@
#include "algo/sha/sha-hash-4way.h"
#if defined(__VAES__)
#include "algo/groestl/groestl512-hash-4way.h"
#include "algo/shavite/shavite-hash-2way.h"
#include "algo/shavite/shavite-hash-4way.h"
#include "algo/echo/echo-hash-4way.h"
#endif
@@ -145,15 +146,21 @@ union _x16r_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_echo echo;
#if defined(__VAES__)
groestl512_2way_context groestl;
shavite512_2way_context shavite;
echo_2way_context echo;
#else
hashState_groestl groestl;
shavite512_context shavite;
hashState_echo echo;
#endif
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
luffa_2way_context luffa;
hashState_luffa luffa1;
cubehashParam cube;
shavite512_context shavite;
simd_2way_context simd;
hamsi512_4way_context hamsi;
hashState_fugue fugue;

37
algo/x16/x16rv2-4way.c
View File

@@ -672,14 +672,20 @@ union _x16rv2_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_echo echo;
#if defined(__VAES__)
groestl512_2way_context groestl;
shavite512_2way_context shavite;
echo_2way_context echo;
#else
hashState_groestl groestl;
shavite512_context shavite;
hashState_echo echo;
#endif
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
luffa_2way_context luffa;
cubehashParam cube;
shavite512_context shavite;
simd_2way_context simd;
hamsi512_4way_context hamsi;
hashState_fugue fugue;
@@ -745,10 +751,19 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
break;
case GROESTL:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
groestl512_2way_full( &ctx.groestl, vhash, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
groestl512_full( &ctx.groestl, (char*)hash0, (char*)in0, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash1, (char*)in1, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash2, (char*)in2, size<<3 );
groestl512_full( &ctx.groestl, (char*)hash3, (char*)in3, size<<3 );
#endif
break;
case JH:
if ( i == 0 )
@@ -887,10 +902,19 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
}
break;
case SHAVITE:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
shavite512_2way_full( &ctx.shavite, vhash, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
shavite512_full( &ctx.shavite, hash0, in0, size );
shavite512_full( &ctx.shavite, hash1, in1, size );
shavite512_full( &ctx.shavite, hash2, in2, size );
shavite512_full( &ctx.shavite, hash3, in3, size );
#endif
break;
case SIMD:
intrlv_2x128( vhash, in0, in1, size<<3 );
@@ -901,6 +925,14 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
dintrlv_2x128_512( hash2, hash3, vhash );
break;
case ECHO:
#if defined(__VAES__)
intrlv_2x128( vhash, in0, in1, size<<3 );
echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
dintrlv_2x128_512( hash0, hash1, vhash );
intrlv_2x128( vhash, in2, in3, size<<3 );
echo_2way_full( &ctx.echo, vhash, 512, vhash, size );
dintrlv_2x128_512( hash2, hash3, vhash );
#else
echo_full( &ctx.echo, (BitSequence *)hash0, 512,
(const BitSequence *)in0, size );
echo_full( &ctx.echo, (BitSequence *)hash1, 512,
@@ -909,6 +941,7 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
(const BitSequence *)in2, size );
echo_full( &ctx.echo, (BitSequence *)hash3, 512,
(const BitSequence *)in3, size );
#endif
break;
case HAMSI:
if ( i == 0 )