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

v3.7.8

Browse Source
This commit is contained in:
Jay D Dee
2017-12-30 19:19:46 -05:00
parent 79164c24b5
commit 2d2e54f001
66 changed files with 4321 additions and 1475 deletions
Download Patch File Download Diff File Expand all files Collapse all files

261
algo/x11/c11-4way.c Normal file
View File

@@ -0,0 +1,261 @@
#include "cpuminer-config.h"
#include "c11-gate.h"
#if defined (__AVX2__) && defined (__AES__)
#include <string.h>
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.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/luffa/sse2/luffa_for_sse2.h"
#include "algo/cubehash/sse2/cubehash_sse2.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/sse2/nist.h"
#include "algo/echo/aes_ni/hash_api.h"
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
hashState_luffa luffa;
cubehashParam cube;
sph_shavite512_context shavite;
hashState_sd simd;
hashState_echo echo;
} c11_4way_ctx_holder;
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 );
init_groestl( &c11_4way_ctx.groestl, 64 );
skein512_4way_init( &c11_4way_ctx.skein );
jh512_4way_init( &c11_4way_ctx.jh );
keccak512_4way_init( &c11_4way_ctx.keccak );
init_luffa( &c11_4way_ctx.luffa, 512 );
cubehashInit( &c11_4way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &c11_4way_ctx.shavite );
init_sd( &c11_4way_ctx.simd, 512 );
init_echo( &c11_4way_ctx.echo, 512 );
}
void c11_4way_hash( 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)));
c11_4way_ctx_holder ctx;
memcpy( &ctx, &c11_4way_ctx, sizeof(c11_4way_ctx) );
// 1 Blake 4way
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, 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) );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
memcpy( &ctx.groestl, &c11_4way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
memcpy( &ctx.groestl, &c11_4way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// 4way
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 JH
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
// 5 Keccak
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// 6 Skein
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 7 Luffa
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash0,
(const BitSequence*)hash0, 64 );
memcpy( &ctx.luffa, &c11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash1,
(const BitSequence*)hash1, 64 );
memcpy( &ctx.luffa, &c11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash2,
(const BitSequence*)hash2, 64 );
memcpy( &ctx.luffa, &c11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash3,
(const BitSequence*)hash3, 64 );
// 8 Cubehash
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
memcpy( &ctx.cube, &c11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash1, (const byte*) hash1, 64 );
memcpy( &ctx.cube, &c11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash2, (const byte*) hash2, 64 );
memcpy( &ctx.cube, &c11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash3, (const byte*) hash3, 64 );
// 9 Shavite
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &c11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash1, 64 );
sph_shavite512_close( &ctx.shavite, hash1 );
memcpy( &ctx.shavite, &c11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash2, 64 );
sph_shavite512_close( &ctx.shavite, hash2 );
memcpy( &ctx.shavite, &c11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash3, 64 );
sph_shavite512_close( &ctx.shavite, hash3 );
// 10 Simd
update_final_sd( &ctx.simd, (BitSequence *)hash0,
(const BitSequence *)hash0, 512 );
memcpy( &ctx.simd, &c11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash1,
(const BitSequence *)hash1, 512 );
memcpy( &ctx.simd, &c11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash2,
(const BitSequence *)hash2, 512 );
memcpy( &ctx.simd, &c11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash3,
(const BitSequence *)hash3, 512 );
// 11 Echo
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &c11_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash1,
(const BitSequence *) hash1, 512 );
memcpy( &ctx.echo, &c11_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash2,
(const BitSequence *) hash2, 512 );
memcpy( &ctx.echo, &c11_4way_ctx.echo, sizeof(hashState_echo) );
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_c11_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[24*4] __attribute__ ((aligned (64)));
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
const uint32_t first_nonce = 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;
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
0xFFFFF000, 0xFFFF0000, 0 };
// big endian encode 0..18 uint32_t, 64 bits at a time
swab32_array( endiandata, pdata, 20 );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
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 );
c11_4way_hash( hash, vdata );
pdata[19] = n;
if ( ( hash[7] & mask ) == 0 && fulltest( hash, ptarget ) )
{
found[0] = true;
num_found++;
nonces[0] = n;
work_set_target_ratio( work, hash );
}
if ( ( (hash+8)[7] & mask ) == 0 && 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 ) == 0 && 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 ) == 0 && 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/c11-gate.c Normal file
View File

@@ -0,0 +1,18 @@
#include "c11-gate.h"
bool register_c11_algo( algo_gate_t* gate )
{
#if defined (C11_4WAY)
init_c11_4way_ctx();
gate->scanhash = (void*)&scanhash_c11_4way;
gate->hash = (void*)&c11_4way_hash;
#else
init_c11_ctx();
gate->scanhash = (void*)&scanhash_c11;
gate->hash = (void*)&c11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

32
algo/x11/c11-gate.h Normal file
View File

@@ -0,0 +1,32 @@
#ifndef C11_GATE_H__
#define C11_GATE_H__ 1
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#define C11_4WAY
#endif
bool register_c11_algo( algo_gate_t* gate );
#if defined(C11_4WAY)
void c11_4way_hash( void *state, const void *input );
int scanhash_c11_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_c11_4way_ctx();
#endif
void c11_hash( void *state, const void *input );
int scanhash_c11( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_c11_ctx();
#endif

16
algo/x11/c11.c
View File

@@ -1,4 +1,4 @@
#include "algo-gate-api.h"
#include "c11-gate.h"
#include <stdlib.h>
#include <stdint.h>
@@ -64,7 +64,7 @@ void init_c11_ctx()
#endif
}
void c11hash( void *output, const void *input )
void c11_hash( void *output, const void *input )
{
unsigned char hash[128] _ALIGN(64); // uint32_t hashA[16], hashB[16];
// uint32_t _ALIGN(64) hash[16];
@@ -157,7 +157,7 @@ int scanhash_c11( int thr_id, struct work *work, uint32_t max_nonce,
do
{
be32enc( &endiandata[19], nonce );
c11hash( hash, endiandata );
c11_hash( hash, endiandata );
if ( hash[7] <= Htarg && fulltest(hash, ptarget) )
{
pdata[19] = nonce;
@@ -171,13 +171,3 @@ int scanhash_c11( int thr_id, struct work *work, uint32_t max_nonce,
return 0;
}
bool register_c11_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
init_c11_ctx();
gate->scanhash = (void*)&scanhash_c11;
gate->hash = (void*)&c11hash;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

139
algo/x11/phi1612.c
View File

@@ -1,139 +0,0 @@
#include "algo-gate-api.h"
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/gost/sph_gost.h"
#include "algo/echo/sph_echo.h"
#include "algo/fugue//sph_fugue.h"
#include "algo/cubehash/sse2/cubehash_sse2.h"
#include "algo/skein/sse2/skein.c"
#include "algo/jh/sph_jh.h"
#ifndef NO_AES_NI
#include "algo/echo/aes_ni/hash_api.h"
#endif
typedef struct {
sph_skein512_context skein;
sph_jh512_context jh;
cubehashParam cube;
sph_fugue512_context fugue;
sph_gost512_context gost;
#ifdef NO_AES_NI
sph_echo512_context echo;
#else
hashState_echo echo;
#endif
} phi_ctx_holder;
phi_ctx_holder phi_ctx;
static __thread sph_skein512_context phi_skein_mid
__attribute__ ((aligned (64)));
void init_phi_ctx()
{
sph_skein512_init( &phi_ctx.skein );
sph_jh512_init( &phi_ctx.jh );
cubehashInit( &phi_ctx.cube, 512, 16, 32 );
sph_fugue512_init( &phi_ctx.fugue );
sph_gost512_init( &phi_ctx.gost );
#ifdef NO_AES_NI
sph_echo512_init( &phi_ctx.echo );
#else
init_echo( &phi_ctx.echo, 512 );
#endif
}
void phi_skein_midstate( const void* input )
{
memcpy( &phi_skein_mid, &phi_ctx.skein, sizeof phi_skein_mid );
sph_skein512( &phi_skein_mid, input, 64 );
}
void phi1612hash(void *output, const void *input)
{
unsigned char hash[128] __attribute__ ((aligned (64)));
phi_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &phi_ctx, sizeof(phi_ctx) );
memcpy( &ctx.skein, &phi_skein_mid, sizeof phi_skein_mid );
sph_skein512( &ctx.skein, input + 64, 16 );
sph_skein512_close( &ctx.skein, hash );
// sph_skein512( &ctx.skein, input, 80 );
// sph_skein512_close( &ctx.skein, (void*)hash );
sph_jh512( &ctx.jh, (const void*)hash, 64 );
sph_jh512_close( &ctx.jh, (void*)hash );
cubehashUpdateDigest( &ctx.cube, (byte*) hash, (const byte*)hash, 64 );
sph_fugue512( &ctx.fugue, (const void*)hash, 64 );
sph_fugue512_close( &ctx.fugue, (void*)hash );
sph_gost512( &ctx.gost, hash, 64 );
sph_gost512_close( &ctx.gost, 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(output, hash, 32);
}
int scanhash_phi1612( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
uint32_t _ALIGN(64) endiandata[20];
uint32_t nonce = first_nonce;
volatile uint8_t *restart = &(work_restart[thr_id].restart);
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x0cff;
for (int k = 0; k < 19; k++)
be32enc(&endiandata[k], pdata[k]);
phi_skein_midstate( endiandata );
const uint32_t Htarg = ptarget[7];
do {
uint32_t hash[8];
be32enc(&endiandata[19], nonce);
phi1612hash(hash, endiandata);
if (hash[7] <= Htarg && fulltest(hash, ptarget)) {
pdata[19] = nonce;
*hashes_done = pdata[19] - first_nonce;
return 1;
}
nonce++;
} while (nonce < max_nonce && !(*restart));
pdata[19] = nonce;
*hashes_done = pdata[19] - first_nonce + 1;
return 0;
}
bool register_phi1612_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
init_phi_ctx();
gate->scanhash = (void*)&scanhash_phi1612;
gate->hash = (void*)&phi1612hash;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
}

261
algo/x11/x11-4way.c Normal file
View File

@@ -0,0 +1,261 @@
#include "cpuminer-config.h"
#include "x11-gate.h"
#if defined (__AVX2__) && defined (__AES__)
#include <string.h>
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.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/luffa/sse2/luffa_for_sse2.h"
#include "algo/cubehash/sse2/cubehash_sse2.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/sse2/nist.h"
#include "algo/echo/aes_ni/hash_api.h"
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
hashState_luffa luffa;
cubehashParam cube;
sph_shavite512_context shavite;
hashState_sd simd;
hashState_echo echo;
} x11_4way_ctx_holder;
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 );
init_groestl( &x11_4way_ctx.groestl, 64 );
skein512_4way_init( &x11_4way_ctx.skein );
jh512_4way_init( &x11_4way_ctx.jh );
keccak512_4way_init( &x11_4way_ctx.keccak );
init_luffa( &x11_4way_ctx.luffa, 512 );
cubehashInit( &x11_4way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &x11_4way_ctx.shavite );
init_sd( &x11_4way_ctx.simd, 512 );
init_echo( &x11_4way_ctx.echo, 512 );
}
void x11_4way_hash( 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)));
x11_4way_ctx_holder ctx;
memcpy( &ctx, &x11_4way_ctx, sizeof(x11_4way_ctx) );
// 1 Blake 4way
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, 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) );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
memcpy( &ctx.groestl, &x11_4way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
memcpy( &ctx.groestl, &x11_4way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// 4way
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
// 5 JH
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
// 6 Keccak
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 7 Luffa
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash0,
(const BitSequence*)hash0, 64 );
memcpy( &ctx.luffa, &x11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash1,
(const BitSequence*)hash1, 64 );
memcpy( &ctx.luffa, &x11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash2,
(const BitSequence*)hash2, 64 );
memcpy( &ctx.luffa, &x11_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash3,
(const BitSequence*)hash3, 64 );
// 8 Cubehash
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
memcpy( &ctx.cube, &x11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash1, (const byte*) hash1, 64 );
memcpy( &ctx.cube, &x11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash2, (const byte*) hash2, 64 );
memcpy( &ctx.cube, &x11_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash3, (const byte*) hash3, 64 );
// 9 Shavite
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &x11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash1, 64 );
sph_shavite512_close( &ctx.shavite, hash1 );
memcpy( &ctx.shavite, &x11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash2, 64 );
sph_shavite512_close( &ctx.shavite, hash2 );
memcpy( &ctx.shavite, &x11_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash3, 64 );
sph_shavite512_close( &ctx.shavite, hash3 );
// 10 Simd
update_final_sd( &ctx.simd, (BitSequence *)hash0,
(const BitSequence *)hash0, 512 );
memcpy( &ctx.simd, &x11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash1,
(const BitSequence *)hash1, 512 );
memcpy( &ctx.simd, &x11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash2,
(const BitSequence *)hash2, 512 );
memcpy( &ctx.simd, &x11_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash3,
(const BitSequence *)hash3, 512 );
// 11 Echo
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &x11_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash1,
(const BitSequence *) hash1, 512 );
memcpy( &ctx.echo, &x11_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash2,
(const BitSequence *) hash2, 512 );
memcpy( &ctx.echo, &x11_4way_ctx.echo, sizeof(hashState_echo) );
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_x11_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[24*4] __attribute__ ((aligned (64)));
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
const uint32_t first_nonce = 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;
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
0xFFFFF000, 0xFFFF0000, 0 };
// big endian encode 0..18 uint32_t, 64 bits at a time
swab32_array( endiandata, pdata, 20 );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
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 );
x11_4way_hash( hash, vdata );
pdata[19] = n;
if ( ( hash[7] & mask ) == 0 && fulltest( hash, ptarget ) )
{
found[0] = true;
num_found++;
nonces[0] = n;
work_set_target_ratio( work, hash );
}
if ( ( (hash+8)[7] & mask ) == 0 && 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 ) == 0 && 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 ) == 0 && 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

4
algo/x11/x11-gate.c
View File

@@ -5,13 +5,13 @@ bool register_x11_algo( algo_gate_t* gate )
#if defined (X11_4WAY)
init_x11_4way_ctx();
gate->scanhash = (void*)&scanhash_x11_4way;
gate->hash = (void*)&x11_hash_4way;
gate->hash = (void*)&x11_4way_hash;
#else
init_x11_ctx();
gate->scanhash = (void*)&scanhash_x11;
gate->hash = (void*)&x11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

10
algo/x11/x11-gate.h
View File

@@ -4,19 +4,21 @@
#include "algo-gate-api.h"
#include <stdint.h>
//#if defined(HASH_4WAY) && !defined(NO_AES_NI)
// #define X11_4WAY
//#endif
#if defined(HASH_4WAY) && defined(__AES__)
#define X11_4WAY
#endif
bool register_x11_algo( algo_gate_t* gate );
#if defined(X11_4WAY)
void x11_hash_4way( void *state, const void *input );
void x11_4way_hash( void *state, const void *input );
int scanhash_x11_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_x11_4way_ctx();
#endif
void x11_hash( void *state, const void *input );

268
algo/x11/x11gost-4way.c Normal file
View File

@@ -0,0 +1,268 @@
#include "cpuminer-config.h"
#include "x11gost-gate.h"
#if defined (__AVX2__) && defined (__AES__)
#include <string.h>
#include <stdint.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/sph_bmw.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/gost/sph_gost.h"
#include "algo/luffa/sse2/luffa_for_sse2.h"
#include "algo/cubehash/sse2/cubehash_sse2.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/sse2/nist.h"
#include "algo/echo/aes_ni/hash_api.h"
typedef struct {
blake512_4way_context blake;
sph_bmw512_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
sph_gost512_context gost;
hashState_luffa luffa;
cubehashParam cube;
sph_shavite512_context shavite;
hashState_sd simd;
hashState_echo echo;
} x11gost_4way_ctx_holder;
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 );
init_groestl( &x11gost_4way_ctx.groestl, 64 );
skein512_4way_init( &x11gost_4way_ctx.skein );
jh512_4way_init( &x11gost_4way_ctx.jh );
keccak512_4way_init( &x11gost_4way_ctx.keccak );
sph_gost512_init( &x11gost_4way_ctx.gost );
init_luffa( &x11gost_4way_ctx.luffa, 512 );
cubehashInit( &x11gost_4way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &x11gost_4way_ctx.shavite );
init_sd( &x11gost_4way_ctx.simd, 512 );
init_echo( &x11gost_4way_ctx.echo, 512 );
}
void x11gost_4way_hash( 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)));
x11gost_4way_ctx_holder ctx;
memcpy( &ctx, &x11gost_4way_ctx, sizeof(x11gost_4way_ctx) );
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, 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) );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
memcpy( &ctx.groestl, &x11gost_4way_ctx.groestl,
sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
memcpy( &ctx.groestl, &x11gost_4way_ctx.groestl,
sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// 4way
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_gost512( &ctx.gost, hash0, 64 );
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 );
sph_gost512( &ctx.gost, hash2, 64 );
sph_gost512_close( &ctx.gost, hash2 );
memcpy( &ctx.gost, &x11gost_4way_ctx.gost, sizeof(sph_gost512_context) );
sph_gost512( &ctx.gost, hash3, 64 );
sph_gost512_close( &ctx.gost, hash3 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash0,
(const BitSequence*)hash0, 64 );
memcpy( &ctx.luffa, &x11gost_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash1,
(const BitSequence*)hash1, 64 );
memcpy( &ctx.luffa, &x11gost_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash2,
(const BitSequence*)hash2, 64 );
memcpy( &ctx.luffa, &x11gost_4way_ctx.luffa, sizeof(hashState_luffa) );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash3,
(const BitSequence*)hash3, 64 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
memcpy( &ctx.cube, &x11gost_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash1, (const byte*) hash1, 64 );
memcpy( &ctx.cube, &x11gost_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash2, (const byte*) hash2, 64 );
memcpy( &ctx.cube, &x11gost_4way_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*)hash3, (const byte*) hash3, 64 );
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &x11gost_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash1, 64 );
sph_shavite512_close( &ctx.shavite, hash1 );
memcpy( &ctx.shavite, &x11gost_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash2, 64 );
sph_shavite512_close( &ctx.shavite, hash2 );
memcpy( &ctx.shavite, &x11gost_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash3, 64 );
sph_shavite512_close( &ctx.shavite, hash3 );
update_final_sd( &ctx.simd, (BitSequence *)hash0,
(const BitSequence *)hash0, 512 );
memcpy( &ctx.simd, &x11gost_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash1,
(const BitSequence *)hash1, 512 );
memcpy( &ctx.simd, &x11gost_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash2,
(const BitSequence *)hash2, 512 );
memcpy( &ctx.simd, &x11gost_4way_ctx.simd, sizeof(hashState_sd) );
update_final_sd( &ctx.simd, (BitSequence *)hash3,
(const BitSequence *)hash3, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &x11gost_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash1,
(const BitSequence *) hash1, 512 );
memcpy( &ctx.echo, &x11gost_4way_ctx.echo, sizeof(hashState_echo) );
update_final_echo( &ctx.echo, (BitSequence *)hash2,
(const BitSequence *) hash2, 512 );
memcpy( &ctx.echo, &x11gost_4way_ctx.echo, sizeof(hashState_echo) );
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_x11gost_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[24*4] __attribute__ ((aligned (64)));
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
const uint32_t first_nonce = 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;
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
0xFFFFF000, 0xFFFF0000, 0 };
// big endian encode 0..18 uint32_t, 64 bits at a time
swab32_array( endiandata, pdata, 20 );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
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 );
x11gost_4way_hash( hash, vdata );
pdata[19] = n;
if ( ( hash[7] & mask ) == 0 && fulltest( hash, ptarget ) )
{
found[0] = true;
num_found++;
nonces[0] = n;
work_set_target_ratio( work, hash );
}
if ( ( (hash+8)[7] & mask ) == 0 && 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 ) == 0 && 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 ) == 0 && 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/x11gost-gate.c Normal file
View File

@@ -0,0 +1,18 @@
#include "x11gost-gate.h"
bool register_x11gost_algo( algo_gate_t* gate )
{
#if defined (X11GOST_4WAY)
init_x11gost_4way_ctx();
gate->scanhash = (void*)&scanhash_x11gost_4way;
gate->hash = (void*)&x11gost_4way_hash;
#else
init_x11gost_ctx();
gate->scanhash = (void*)&scanhash_x11gost;
gate->hash = (void*)&x11gost_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

32
algo/x11/x11gost-gate.h Normal file
View File

@@ -0,0 +1,32 @@
#ifndef X11GOST_GATE_H__
#define X11GOST_GATE_H__ 1
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#define X11GOST_4WAY
#endif
bool register_x11gost_algo( algo_gate_t* gate );
#if defined(X11GOST_4WAY)
void x11gost_4way_hash( void *state, const void *input );
int scanhash_x11gost_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_x11gost_4way_ctx();
#endif
void x11gost_hash( void *state, const void *input );
int scanhash_x11gost( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_x11gost_ctx();
#endif

47
algo/x11/x11gost.c
View File

@@ -1,4 +1,4 @@
#include "algo-gate-api.h"
#include "x11gost-gate.h"
#include <stdlib.h>
#include <stdint.h>
@@ -37,28 +37,28 @@ typedef struct {
hashState_echo echo;
hashState_groestl groestl;
#endif
} sib_ctx_holder;
} x11gost_ctx_holder;
sib_ctx_holder sib_ctx;
x11gost_ctx_holder x11gost_ctx;
void init_sib_ctx()
void init_x11gost_ctx()
{
sph_gost512_init(&sib_ctx.gost);
sph_shavite512_init(&sib_ctx.shavite);
init_luffa( &sib_ctx.luffa, 512 );
cubehashInit( &sib_ctx.cube, 512, 16, 32 );
init_sd( &sib_ctx.simd, 512 );
sph_gost512_init( &x11gost_ctx.gost );
sph_shavite512_init( &x11gost_ctx.shavite );
init_luffa( &x11gost_ctx.luffa, 512 );
cubehashInit( &x11gost_ctx.cube, 512, 16, 32 );
init_sd( &x11gost_ctx.simd, 512 );
#ifdef NO_AES_NI
sph_groestl512_init( &sib_ctx.groestl );
sph_echo512_init( &sib_ctx.echo );
sph_groestl512_init( &x11gost_ctx.groestl );
sph_echo512_init( &x11gost_ctx.echo );
#else
init_echo( &sib_ctx.echo, 512 );
init_groestl( &sib_ctx.groestl, 64 );
init_echo( &x11gost_ctx.echo, 512 );
init_groestl( &x11gost_ctx.groestl, 64 );
#endif
}
void sibhash(void *output, const void *input)
void x11gost_hash(void *output, const void *input)
{
unsigned char hash[128] __attribute__ ((aligned (64)));
#define hashA hash
@@ -69,8 +69,8 @@ void sibhash(void *output, const void *input)
sph_u64 hashctA;
sph_u64 hashctB;
sib_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &sib_ctx, sizeof(sib_ctx) );
x11gost_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &x11gost_ctx, sizeof(x11gost_ctx) );
DECL_BLK;
BLK_I;
@@ -135,8 +135,8 @@ void sibhash(void *output, const void *input)
memcpy(output, hashA, 32);
}
int scanhash_sib(int thr_id, struct work *work,
uint32_t max_nonce, uint64_t *hashes_done)
int scanhash_x11gost( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done)
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
@@ -156,7 +156,7 @@ int scanhash_sib(int thr_id, struct work *work,
do {
uint32_t hash[8];
be32enc(&endiandata[19], nonce);
sibhash(hash, endiandata);
x11gost_hash(hash, endiandata);
if (hash[7] <= Htarg && fulltest(hash, ptarget)) {
pdata[19] = nonce;
@@ -172,12 +172,3 @@ int scanhash_sib(int thr_id, struct work *work,
return 0;
}
bool register_sib_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
init_sib_ctx();
gate->scanhash = (void*)&scanhash_sib;
gate->hash = (void*)&sibhash;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
}