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

v3.7.9

Browse Source
This commit is contained in:
Jay D Dee
2018-01-08 22:04:43 -05:00
parent 2d2e54f001
commit bee78eac76
58 changed files with 2817 additions and 499 deletions
Download Patch File Download Diff File Expand all files Collapse all files

23
algo/x11/c11-4way.c
View File

@@ -7,7 +7,7 @@
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/bmw/bmw-hash-4way.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/skein/skein-hash-4way.h"
#include "algo/jh/jh-hash-4way.h"
@@ -20,7 +20,7 @@
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
bmw512_4way_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
@@ -37,7 +37,7 @@ c11_4way_ctx_holder c11_4way_ctx;
void init_c11_4way_ctx()
{
blake512_4way_init( &c11_4way_ctx.blake );
sph_bmw512_init( &c11_4way_ctx.bmw );
bmw512_4way_init( &c11_4way_ctx.bmw );
init_groestl( &c11_4way_ctx.groestl, 64 );
skein512_4way_init( &c11_4way_ctx.skein );
jh512_4way_init( &c11_4way_ctx.jh );
@@ -63,22 +63,13 @@ void c11_4way_hash( void *state, const void *input )
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
// 2 Bmw
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 2 Bmw
sph_bmw512( &ctx.bmw, hash0, 64 );
sph_bmw512_close( &ctx.bmw, hash0 );
memcpy( &ctx.bmw, &c11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash1, 64 );
sph_bmw512_close( &ctx.bmw, hash1 );
memcpy( &ctx.bmw, &c11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash2, 64 );
sph_bmw512_close( &ctx.bmw, hash2 );
memcpy( &ctx.bmw, &c11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash3, 64 );
sph_bmw512_close( &ctx.bmw, hash3 );
// 3 Groestl
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
memcpy( &ctx.groestl, &c11_4way_ctx.groestl, sizeof(hashState_groestl) );

2
algo/x11/c11-gate.c
View File

@@ -11,7 +11,7 @@ bool register_c11_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_c11;
gate->hash = (void*)&c11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

137
algo/x11/fresh.c Normal file
View File

@@ -0,0 +1,137 @@
#include "algo-gate-api.h"
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/sph_simd.h"
#include "algo/echo/sph_echo.h"
//#define DEBUG_ALGO
extern void freshhash(void* output, const void* input, uint32_t len)
{
unsigned char hash[128]; // uint32_t hashA[16], hashB[16];
#define hashA hash
#define hashB hash+64
sph_shavite512_context ctx_shavite;
sph_simd512_context ctx_simd;
sph_echo512_context ctx_echo;
sph_shavite512_init(&ctx_shavite);
sph_shavite512(&ctx_shavite, input, len);
sph_shavite512_close(&ctx_shavite, hashA);
sph_simd512_init(&ctx_simd);
sph_simd512(&ctx_simd, hashA, 64);
sph_simd512_close(&ctx_simd, hashB);
sph_shavite512_init(&ctx_shavite);
sph_shavite512(&ctx_shavite, hashB, 64);
sph_shavite512_close(&ctx_shavite, hashA);
sph_simd512_init(&ctx_simd);
sph_simd512(&ctx_simd, hashA, 64);
sph_simd512_close(&ctx_simd, hashB);
sph_echo512_init(&ctx_echo);
sph_echo512(&ctx_echo, hashB, 64);
sph_echo512_close(&ctx_echo, hashA);
memcpy(output, hash, 32);
}
int scanhash_fresh(int thr_id, struct work *work,
uint32_t max_nonce, uint64_t *hashes_done)
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t len = 80;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
#ifdef _MSC_VER
uint32_t __declspec(align(32)) hash64[8];
#else
uint32_t hash64[8] __attribute__((aligned(32)));
#endif
uint32_t endiandata[32];
uint64_t htmax[] = {
0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000
};
uint32_t masks[] = {
0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0
};
// we need bigendian data...
for (int k = 0; k < 19; k++)
be32enc(&endiandata[k], pdata[k]);
#ifdef DEBUG_ALGO
if (Htarg != 0)
printf("[%d] Htarg=%X\n", thr_id, Htarg);
#endif
for (int m=0; m < 6; m++) {
if (Htarg <= htmax[m]) {
uint32_t mask = masks[m];
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
freshhash(hash64, endiandata, len);
#ifndef DEBUG_ALGO
if ((!(hash64[7] & mask)) && fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
#else
if (!(n % 0x1000) && !thr_id) printf(".");
if (!(hash64[7] & mask)) {
printf("[%d]",thr_id);
if (fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
}
#endif
} while (n < max_nonce && !work_restart[thr_id].restart);
// see blake.c if else to understand the loop on htmax => mask
break;
}
}
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
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;
return true;
};

168
algo/x11/tribus-4way.c Normal file
View File

@@ -0,0 +1,168 @@
#include "tribus-gate.h"
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#if defined(__AVX2__) && !defined(NO_AES_NI)
#include "algo/jh/jh-hash-4way.h"
#include "algo/keccak/keccak-hash-4way.h"
#include "algo/echo/aes_ni/hash_api.h"
//hashState_echo tribus_4way_ctx __attribute__ ((aligned (64)));
static __thread jh512_4way_context ctx_mid;
/*
void init_tribus_4way_ctx()
{
init_echo( &tribus_4way_ctx, 512 );
}
*/
void tribus_hash_4way(void *state, const void *input)
{
uint64_t hash0[8] __attribute__ ((aligned (64)));
uint64_t hash1[8] __attribute__ ((aligned (64)));
uint64_t hash2[8] __attribute__ ((aligned (64)));
uint64_t hash3[8] __attribute__ ((aligned (64)));
uint64_t vhash[8*4] __attribute__ ((aligned (64)));
jh512_4way_context ctx_jh;
keccak512_4way_context ctx_keccak;
hashState_echo ctx_echo;
memcpy( &ctx_jh, &ctx_mid, sizeof(ctx_mid) );
jh512_4way( &ctx_jh, input + (64<<2), 16 );
jh512_4way_close( &ctx_jh, vhash );
keccak512_4way_init( &ctx_keccak );
keccak512_4way( &ctx_keccak, vhash, 64 );
keccak512_4way_close( &ctx_keccak, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// hash echo serially
init_echo( &ctx_echo, 512 );
update_final_echo( &ctx_echo, (BitSequence *) hash0,
(const BitSequence *) hash0, 512 );
init_echo( &ctx_echo, 512 );
update_final_echo( &ctx_echo, (BitSequence *) hash1,
(const BitSequence *) hash1, 512 );
init_echo( &ctx_echo, 512 );
update_final_echo( &ctx_echo, (BitSequence *) hash2,
(const BitSequence *) hash2, 512 );
init_echo( &ctx_echo, 512 );
update_final_echo( &ctx_echo, (BitSequence *) hash3,
(const BitSequence *) hash3, 512 );
memcpy( state, hash0, 32 );
memcpy( state+32, hash1, 32 );
memcpy( state+64, hash2, 32 );
memcpy( state+96, hash3, 32 );
}
int scanhash_tribus_4way(int thr_id, struct work *work, uint32_t max_nonce, uint64_t *hashes_done)
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t vdata[20*4] __attribute__ ((aligned (64)));
uint32_t _ALIGN(128) endiandata[20];
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
uint32_t n = pdata[19];
uint32_t *nonces = work->nonces;
bool *found = work->nfound;
int num_found = 0;
uint32_t *noncep0 = vdata + 73; // 9*8 + 1
uint32_t *noncep1 = vdata + 75;
uint32_t *noncep2 = vdata + 77;
uint32_t *noncep3 = vdata + 79;
uint64_t htmax[] = { 0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000 };
uint32_t masks[] = { 0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0 };
// we need bigendian data...
for ( int i = 0; i < 20; i++ )
{
be32enc( &endiandata[i], pdata[i] );
}
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
// precalc midstate
// doing it one way then then interleaving would be faster but too
// complicated tto interleave context.
jh512_4way_init( &ctx_mid );
jh512_4way( &ctx_mid, vdata, 64 );
for ( int m = 0; m < 6; m++ )
{
if ( Htarg <= htmax[m] )
{
uint32_t mask = masks[m];
do {
found[0] = found[1] = found[2] = found[3] = false;
be32enc( noncep0, n );
be32enc( noncep1, n+1 );
be32enc( noncep2, n+2 );
be32enc( noncep3, n+3 );
tribus_hash_4way( hash, vdata );
pdata[19] = n;
if ( ( !(hash[7] & mask) )
&& fulltest( hash, ptarget ) )
{
found[0] = true;
num_found++;
nonces[0] = n;
work_set_target_ratio(work, hash);
}
if ( ( !((hash+8)[7] & mask) )
&& fulltest (hash+8, ptarget ) )
{
found[1] = true;
num_found++;
nonces[1] = n+1;
work_set_target_ratio(work, hash+8);
}
if ( ( !((hash+16)[7] & mask) )
&& fulltest( hash+16, ptarget ) )
{
found[2] = true;
num_found++;
nonces[2] = n+2;
work_set_target_ratio(work, hash+16);
}
if ( ( !((hash+24)[7] & mask) )
&& fulltest( hash+24, ptarget ) )
{
found[3] = true;
num_found++;
nonces[3] = n+3;
work_set_target_ratio(work, hash+24);
}
n += 4;
} while ( (num_found == 0) && ( n < max_nonce )
&& !work_restart[thr_id].restart);
break;
}
}
*hashes_done = n - first_nonce + 1;
return num_found;
}
#endif

18
algo/x11/tribus-gate.c Normal file
View File

@@ -0,0 +1,18 @@
#include "tribus-gate.h"
bool register_tribus_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x1ffff;
#if defined (TRIBUS_4WAY)
// init_tribus_4way_ctx();
gate->scanhash = (void*)&scanhash_tribus_4way;
gate->hash = (void*)&tribus_hash_4way;
#else
gate->miner_thread_init = (void*)&tribus_thread_init;
gate->scanhash = (void*)&scanhash_tribus;
gate->hash = (void*)&tribus_hash;
#endif
return true;
};

31
algo/x11/tribus-gate.h Normal file
View File

@@ -0,0 +1,31 @@
#ifndef TRIBUS_GATE_H__
#define TRIBUS_GATE_H__
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#define TRIBUS_4WAY
#endif
#if defined(TRIBUS_4WAY)
//void init_tribus_4way_ctx();
void tribus_hash_4way( void *state, const void *input );
int scanhash_tribus_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
#else
void tribus_hash( void *state, const void *input );
int scanhash_tribus( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
bool tribus_thread_init();
#endif
#endif

137
algo/x11/tribus.c Normal file
View File

@@ -0,0 +1,137 @@
#include "tribus-gate.h"
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/jh//sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#ifdef NO_AES_NI
#include "algo/echo/sph_echo.h"
#else
#include "algo/echo/aes_ni/hash_api.h"
#endif
typedef struct {
sph_jh512_context jh;
sph_keccak512_context keccak;
#ifdef NO_AES_NI
sph_echo512_context echo;
#else
hashState_echo echo;
#endif
} tribus_ctx_holder;
static __thread tribus_ctx_holder tribus_ctx;
bool tribus_thread_init()
{
sph_jh512_init( &tribus_ctx.jh );
sph_keccak512_init( &tribus_ctx.keccak );
#ifdef NO_AES_NI
sph_echo512_init( &tribus_ctx.echo );
#else
init_echo( &tribus_ctx.echo, 512 );
#endif
return true;
}
void tribus_hash(void *state, const void *input)
{
unsigned char hash[128] __attribute__ ((aligned (32)));
tribus_ctx_holder ctx;
memcpy( &ctx, &tribus_ctx, sizeof(tribus_ctx) );
sph_jh512( &ctx.jh, input+64, 16 );
sph_jh512_close( &ctx.jh, (void*) hash );
sph_keccak512( &ctx.keccak, (const void*) hash, 64 );
sph_keccak512_close( &ctx.keccak, (void*) hash );
#ifdef NO_AES_NI
sph_echo512( &ctx.echo, hash, 64 );
sph_echo512_close (&ctx.echo, hash );
#else
update_final_echo( &ctx.echo, (BitSequence *) hash,
(const BitSequence *) hash, 512 );
#endif
memcpy(state, hash, 32);
}
int scanhash_tribus(int thr_id, struct work *work, uint32_t max_nonce, uint64_t *hashes_done)
{
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 first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
uint32_t n = pdata[19] - 1;
uint64_t htmax[] = {
0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000
};
uint32_t masks[] = {
0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0
};
// we need bigendian data...
for (int i=0; i < 19; i++) {
be32enc(&endiandata[i], pdata[i]);
}
// precalc midstate
sph_jh512_init( &tribus_ctx.jh );
sph_jh512( &tribus_ctx.jh, endiandata, 64 );
#ifdef DEBUG_ALGO
printf("[%d] Htarg=%X\n", thr_id, Htarg);
#endif
for (int m=0; m < 6; m++) {
if (Htarg <= htmax[m]) {
uint32_t mask = masks[m];
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
tribus_hash(hash32, endiandata);
#ifndef DEBUG_ALGO
if ((!(hash32[7] & mask)) && fulltest(hash32, ptarget)) {
work_set_target_ratio(work, hash32);
*hashes_done = n - first_nonce + 1;
return 1;
}
#else
if (!(n % 0x1000) && !thr_id) printf(".");
if (!(hash32[7] & mask)) {
printf("[%d]",thr_id);
if (fulltest(hash32, ptarget)) {
work_set_target_ratio(work, hash32);
*hashes_done = n - first_nonce + 1;
return 1;
}
}
#endif
} while (n < max_nonce && !work_restart[thr_id].restart);
// see blake.c if else to understand the loop on htmax => mask
break;
}
}
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;
}

23
algo/x11/x11-4way.c
View File

@@ -7,7 +7,7 @@
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/bmw/bmw-hash-4way.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/skein/skein-hash-4way.h"
#include "algo/jh/jh-hash-4way.h"
@@ -20,7 +20,7 @@
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
bmw512_4way_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
@@ -37,7 +37,7 @@ x11_4way_ctx_holder x11_4way_ctx;
void init_x11_4way_ctx()
{
blake512_4way_init( &x11_4way_ctx.blake );
sph_bmw512_init( &x11_4way_ctx.bmw );
bmw512_4way_init( &x11_4way_ctx.bmw );
init_groestl( &x11_4way_ctx.groestl, 64 );
skein512_4way_init( &x11_4way_ctx.skein );
jh512_4way_init( &x11_4way_ctx.jh );
@@ -63,22 +63,13 @@ void x11_4way_hash( void *state, const void *input )
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
// 2 Bmw
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 2 Bmw
sph_bmw512( &ctx.bmw, hash0, 64 );
sph_bmw512_close( &ctx.bmw, hash0 );
memcpy( &ctx.bmw, &x11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash1, 64 );
sph_bmw512_close( &ctx.bmw, hash1 );
memcpy( &ctx.bmw, &x11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash2, 64 );
sph_bmw512_close( &ctx.bmw, hash2 );
memcpy( &ctx.bmw, &x11_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash3, 64 );
sph_bmw512_close( &ctx.bmw, hash3 );
// 3 Groestl
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
memcpy( &ctx.groestl, &x11_4way_ctx.groestl, sizeof(hashState_groestl) );

2
algo/x11/x11-gate.c
View File

@@ -11,7 +11,7 @@ bool register_x11_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x11;
gate->hash = (void*)&x11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

23
algo/x11/x11gost-4way.c
View File

@@ -7,7 +7,7 @@
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/bmw/bmw-hash-4way.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/skein/skein-hash-4way.h"
#include "algo/jh/jh-hash-4way.h"
@@ -21,7 +21,7 @@
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
bmw512_4way_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
@@ -39,7 +39,7 @@ x11gost_4way_ctx_holder x11gost_4way_ctx;
void init_x11gost_4way_ctx()
{
blake512_4way_init( &x11gost_4way_ctx.blake );
sph_bmw512_init( &x11gost_4way_ctx.bmw );
bmw512_4way_init( &x11gost_4way_ctx.bmw );
init_groestl( &x11gost_4way_ctx.groestl, 64 );
skein512_4way_init( &x11gost_4way_ctx.skein );
jh512_4way_init( &x11gost_4way_ctx.jh );
@@ -65,21 +65,12 @@ void x11gost_4way_hash( void *state, const void *input )
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_bmw512( &ctx.bmw, hash0, 64 );
sph_bmw512_close( &ctx.bmw, hash0 );
memcpy( &ctx.bmw, &x11gost_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash1, 64 );
sph_bmw512_close( &ctx.bmw, hash1 );
memcpy( &ctx.bmw, &x11gost_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash2, 64 );
sph_bmw512_close( &ctx.bmw, hash2 );
memcpy( &ctx.bmw, &x11gost_4way_ctx.bmw, sizeof(sph_bmw512_context) );
sph_bmw512( &ctx.bmw, hash3, 64 );
sph_bmw512_close( &ctx.bmw, hash3 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
memcpy( &ctx.groestl, &x11gost_4way_ctx.groestl,
sizeof(hashState_groestl) );
@@ -110,8 +101,8 @@ void x11gost_4way_hash( void *state, const void *input )
sph_gost512_close( &ctx.gost, hash0 );
memcpy( &ctx.gost, &x11gost_4way_ctx.gost, sizeof(sph_gost512_context) );
sph_gost512( &ctx.gost, hash1, 64 );
memcpy( &ctx.gost, &x11gost_4way_ctx.gost, sizeof(sph_gost512_context) );
sph_gost512_close( &ctx.gost, hash1 );
memcpy( &ctx.gost, &x11gost_4way_ctx.gost, sizeof(sph_gost512_context) );
sph_gost512( &ctx.gost, hash2, 64 );
sph_gost512_close( &ctx.gost, hash2 );
memcpy( &ctx.gost, &x11gost_4way_ctx.gost, sizeof(sph_gost512_context) );

2
algo/x11/x11gost-gate.c
View File

@@ -11,7 +11,7 @@ bool register_x11gost_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x11gost;
gate->hash = (void*)&x11gost_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};