mirror of
https://github.com/JayDDee/cpuminer-opt.git
synced 2025-09-17 23:44:27 +00:00
v3.10.4
This commit is contained in:
Jay D Dee
@@ -1,7 +1,4 @@
|
||||
#include "x12-gate.h"
|
||||
|
||||
#if defined(X12_4WAY)
|
||||
|
||||
#include <stdlib.h>
|
||||
#include <stdint.h>
|
||||
#include <string.h>
|
||||
@@ -14,11 +11,223 @@
|
||||
#include "algo/keccak/keccak-hash-4way.h"
|
||||
#include "algo/luffa/luffa-hash-2way.h"
|
||||
#include "algo/cubehash/cubehash_sse2.h"
|
||||
#include "algo/cubehash/cube-hash-2way.h"
|
||||
#include "algo/shavite/sph_shavite.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"
|
||||
|
||||
#if defined(X12_8WAY)
|
||||
|
||||
|
||||
typedef struct {
|
||||
blake512_8way_context blake;
|
||||
bmw512_8way_context bmw;
|
||||
hashState_groestl groestl;
|
||||
skein512_8way_context skein;
|
||||
jh512_8way_context jh;
|
||||
keccak512_8way_context keccak;
|
||||
luffa_4way_context luffa;
|
||||
cube_4way_context cube;
|
||||
sph_shavite512_context shavite;
|
||||
simd_4way_context simd;
|
||||
hashState_echo echo;
|
||||
hamsi512_8way_context hamsi;
|
||||
} x12_8way_ctx_holder;
|
||||
|
||||
x12_8way_ctx_holder x12_8way_ctx __attribute__ ((aligned (64)));
|
||||
|
||||
void init_x12_8way_ctx()
|
||||
{
|
||||
blake512_8way_init( &x12_8way_ctx.blake );
|
||||
bmw512_8way_init( &x12_8way_ctx.bmw );
|
||||
init_groestl( &x12_8way_ctx.groestl, 64 );
|
||||
skein512_8way_init( &x12_8way_ctx.skein );
|
||||
jh512_8way_init( &x12_8way_ctx.jh );
|
||||
keccak512_8way_init( &x12_8way_ctx.keccak );
|
||||
luffa_4way_init( &x12_8way_ctx.luffa, 512 );
|
||||
cube_4way_init( &x12_8way_ctx.cube, 512, 16, 32 );
|
||||
sph_shavite512_init( &x12_8way_ctx.shavite );
|
||||
simd_4way_init( &x12_8way_ctx.simd, 512 );
|
||||
init_echo( &x12_8way_ctx.echo, 512 );
|
||||
hamsi512_8way_init( &x12_8way_ctx.hamsi );
|
||||
};
|
||||
|
||||
void x12_8way_hash( void *state, const void *input )
|
||||
{
|
||||
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
|
||||
uint64_t vhash0[4*8] __attribute__ ((aligned (64)));
|
||||
uint64_t vhash1[4*8] __attribute__ ((aligned (64)));
|
||||
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 hash4[8] __attribute__ ((aligned (64)));
|
||||
uint64_t hash5[8] __attribute__ ((aligned (64)));
|
||||
uint64_t hash6[8] __attribute__ ((aligned (64)));
|
||||
uint64_t hash7[8] __attribute__ ((aligned (64)));
|
||||
|
||||
x12_8way_ctx_holder ctx;
|
||||
memcpy( &ctx, &x12_8way_ctx, sizeof(x12_8way_ctx) );
|
||||
blake512_8way_update( &ctx.blake, input, 80 );
|
||||
blake512_8way_close( &ctx.blake, vhash );
|
||||
|
||||
bmw512_8way_update( &ctx.bmw, vhash, 64 );
|
||||
bmw512_8way_close( &ctx.bmw, vhash );
|
||||
|
||||
rintrlv_8x64_4x128( vhash0, vhash1, vhash, 512 );
|
||||
|
||||
luffa_4way_update_close( &ctx.luffa, vhash0, vhash0, 64 );
|
||||
luffa_4way_init( &ctx.luffa, 512 );
|
||||
luffa_4way_update_close( &ctx.luffa, vhash1, vhash1, 64 );
|
||||
|
||||
cube_4way_update_close( &ctx.cube, vhash0, vhash0, 64 );
|
||||
cube_4way_init( &ctx.cube, 512, 16, 32 );
|
||||
cube_4way_update_close( &ctx.cube, vhash1, vhash1, 64 );
|
||||
|
||||
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash0 );
|
||||
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash1 );
|
||||
|
||||
sph_shavite512( &ctx.shavite, hash0, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash0 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash1, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash1 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash2, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash2 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash3, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash3 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash4, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash4 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash5, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash5 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash6, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash6 );
|
||||
memcpy( &ctx.shavite, &x12_8way_ctx.shavite,
|
||||
sizeof(sph_shavite512_context) );
|
||||
sph_shavite512( &ctx.shavite, hash7, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash7 );
|
||||
|
||||
intrlv_4x128_512( vhash, hash0, hash1, hash2, hash3 );
|
||||
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
|
||||
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash );
|
||||
intrlv_4x128_512( vhash, hash4, hash5, hash6, hash7 );
|
||||
simd_4way_init( &ctx.simd, 512 );
|
||||
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
|
||||
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash );
|
||||
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash1,
|
||||
(const BitSequence *) hash1, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash2,
|
||||
(const BitSequence *) hash2, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash3,
|
||||
(const BitSequence *) hash3, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash4,
|
||||
(const BitSequence *) hash4, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash5,
|
||||
(const BitSequence *) hash5, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash6,
|
||||
(const BitSequence *) hash6, 512 );
|
||||
memcpy( &ctx.echo, &x12_8way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash7,
|
||||
(const BitSequence *) hash7, 512 );
|
||||
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash4, (char*)hash4, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash5, (char*)hash5, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash6, (char*)hash6, 512 );
|
||||
memcpy( &ctx.groestl, &x12_8way_ctx.groestl,
|
||||
sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );
|
||||
|
||||
intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
|
||||
hash7 );
|
||||
|
||||
skein512_8way_update( &ctx.skein, vhash, 64 );
|
||||
skein512_8way_close( &ctx.skein, vhash );
|
||||
|
||||
jh512_8way_update( &ctx.jh, vhash, 64 );
|
||||
jh512_8way_close( &ctx.jh, vhash );
|
||||
|
||||
keccak512_8way_update( &ctx.keccak, vhash, 64 );
|
||||
keccak512_8way_close( &ctx.keccak, vhash );
|
||||
|
||||
hamsi512_8way_update( &ctx.hamsi, vhash, 64 );
|
||||
hamsi512_8way_close( &ctx.hamsi, state );
|
||||
}
|
||||
|
||||
int scanhash_x12_8way( struct work *work, uint32_t max_nonce,
|
||||
uint64_t *hashes_done, struct thr_info *mythr )
|
||||
{
|
||||
uint32_t hash[16*8] __attribute__ ((aligned (128)));
|
||||
uint32_t vdata[24*8] __attribute__ ((aligned (64)));
|
||||
uint32_t lane_hash[8] __attribute__ ((aligned (64)));
|
||||
uint32_t *hash7 = &(hash[49]);
|
||||
uint32_t *pdata = work->data;
|
||||
uint32_t *ptarget = work->target;
|
||||
uint32_t n = pdata[19];
|
||||
const uint32_t first_nonce = pdata[19];
|
||||
const uint32_t Htarg = ptarget[7];
|
||||
__m512i *noncev = (__m512i*)vdata + 9; // aligned
|
||||
int thr_id = mythr->id;
|
||||
|
||||
mm512_bswap32_intrlv80_8x64( vdata, pdata );
|
||||
|
||||
do {
|
||||
*noncev = mm512_intrlv_blend_32( mm512_bswap_32(
|
||||
_mm512_set_epi32( n+7, 0, n+6, 0, n+5, 0, n+4, 0,
|
||||
n+3, 0, n+2, 0, n+1, 0, n , 0 ) ), *noncev );
|
||||
|
||||
x12_8way_hash( hash, vdata );
|
||||
|
||||
for ( int lane = 0; lane < 8; lane++ )
|
||||
if ( hash7[ lane<<1 ] < Htarg )
|
||||
{
|
||||
extr_lane_8x64( lane_hash, hash, lane, 256 );
|
||||
if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
|
||||
{
|
||||
pdata[19] = n + lane;
|
||||
submit_lane_solution( work, lane_hash, mythr, lane );
|
||||
}
|
||||
}
|
||||
n += 8;
|
||||
} while ( ( n < max_nonce-8 ) && !work_restart[thr_id].restart );
|
||||
*hashes_done = n - first_nonce;
|
||||
return 0;
|
||||
}
|
||||
|
||||
#elif defined(X12_4WAY)
|
||||
|
||||
typedef struct {
|
||||
blake512_4way_context blake;
|
||||
@@ -63,45 +272,13 @@ void x12_4way_hash( void *state, const void *input )
|
||||
x12_4way_ctx_holder ctx;
|
||||
memcpy( &ctx, &x12_4way_ctx, sizeof(x12_4way_ctx) );
|
||||
|
||||
// 1 Blake
|
||||
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
|
||||
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
|
||||
|
||||
// 3 Groestl
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
|
||||
|
||||
// Parallel 4way 64 bit
|
||||
intrlv_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
|
||||
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
|
||||
|
||||
// 7 Luffa
|
||||
intrlv_2x128( vhash, hash0, hash1, 512 );
|
||||
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
|
||||
dintrlv_2x128( hash0, hash1, vhash, 512 );
|
||||
@@ -110,7 +287,6 @@ void x12_4way_hash( void *state, const void *input )
|
||||
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
|
||||
dintrlv_2x128( hash2, hash3, vhash, 512 );
|
||||
|
||||
// 8 Cubehash
|
||||
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
|
||||
cubehashInit( &ctx.cube, 512, 16, 32 );
|
||||
cubehashUpdateDigest( &ctx.cube, (byte*)hash1, (const byte*) hash1, 64 );
|
||||
@@ -119,7 +295,6 @@ void x12_4way_hash( void *state, const void *input )
|
||||
cubehashInit( &ctx.cube, 512, 16, 32 );
|
||||
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, &x12_4way_ctx.shavite,
|
||||
@@ -135,7 +310,6 @@ void x12_4way_hash( void *state, const void *input )
|
||||
sph_shavite512( &ctx.shavite, hash3, 64 );
|
||||
sph_shavite512_close( &ctx.shavite, hash3 );
|
||||
|
||||
// 10 Simd
|
||||
intrlv_2x128( vhash, hash0, hash1, 512 );
|
||||
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
|
||||
dintrlv_2x128( hash0, hash1, vhash, 512 );
|
||||
@@ -144,21 +318,25 @@ void x12_4way_hash( void *state, const void *input )
|
||||
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
|
||||
dintrlv_2x128( hash2, hash3, vhash, 512 );
|
||||
|
||||
// 11 Echo
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash0,
|
||||
(const BitSequence *) hash0, 512 );
|
||||
memcpy( &ctx.echo, &x12_4way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash1,
|
||||
(const BitSequence *) hash1, 512 );
|
||||
memcpy( &ctx.echo, &x12_4way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash2,
|
||||
(const BitSequence *) hash2, 512 );
|
||||
memcpy( &ctx.echo, &x12_4way_ctx.echo, sizeof(hashState_echo) );
|
||||
update_final_echo( &ctx.echo, (BitSequence *)hash3,
|
||||
(const BitSequence *) hash3, 512 );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
|
||||
memcpy( &ctx.groestl, &x12_4way_ctx.groestl, sizeof(hashState_groestl) );
|
||||
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
|
||||
|
||||
// 12 Hamsi parallel 4way 32 bit
|
||||
// Parallel 4way 64 bit
|
||||
intrlv_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 );
|
||||
|
||||
hamsi512_4way( &ctx.hamsi, vhash, 64 );
|
||||
hamsi512_4way_close( &ctx.hamsi, vhash );
|
||||
|
||||
|
||||
Reference in New Issue
Block a user