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

v3.8.0

Browse Source
This commit is contained in:
Jay D Dee
2018-01-23 21:02:16 -05:00
parent a90d75b8f5
commit ad2275f74a
121 changed files with 4662 additions and 467 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
algo/x13/phi1612-4way.c
View File

@@ -1,6 +1,6 @@
#include "x13-gate.h"
#include "phi1612-gate.h"
#if defined(__AVX2__) && defined(__AES__)
#if defined(PHI1612_4WAY)
#include <stdlib.h>
#include <stdint.h>

2
algo/x13/phi1612-gate.c
View File

@@ -11,7 +11,7 @@ bool register_phi1612_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_phi1612;
gate->hash = (void*)&phi1612_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

2
algo/x13/phi1612-gate.h
View File

@@ -4,7 +4,7 @@
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#if defined(__AVX2__) && defined(__AES__)
#define PHI1612_4WAY
#endif

2
algo/x13/skunk-4way.c
View File

@@ -1,6 +1,6 @@
#include "skunk-gate.h"
#ifdef __AVX2__
#if defined(SKUNK_4WAY)
#include <stdlib.h>
#include <stdint.h>

2
algo/x13/skunk-gate.c
View File

@@ -2,7 +2,7 @@
bool register_skunk_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AVX_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AVX_OPT | AVX2_OPT;
#if defined (SKUNK_4WAY)
gate->miner_thread_init = (void*)&skunk_4way_thread_init;
gate->scanhash = (void*)&scanhash_skunk_4way;

2
algo/x13/skunk-gate.h
View File

@@ -4,7 +4,7 @@
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY)
#if defined(__AVX2__)
#define SKUNK_4WAY
#endif

31
algo/x13/x13-4way.c
View File

@@ -1,6 +1,6 @@
#include "x13-gate.h"
#if defined(__AVX2__) && defined(__AES__)
#if defined(X13_4WAY)
#include <stdlib.h>
#include <stdint.h>
@@ -17,7 +17,7 @@
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/sse2/nist.h"
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/hamsi/hamsi-hash-4way.h"
#include "algo/fugue/sph_fugue.h"
typedef struct {
@@ -32,7 +32,7 @@ typedef struct {
sph_shavite512_context shavite;
hashState_sd simd;
hashState_echo echo;
sph_hamsi512_context hamsi;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
} x13_4way_ctx_holder;
@@ -51,7 +51,7 @@ void init_x13_4way_ctx()
sph_shavite512_init( &x13_4way_ctx.shavite );
init_sd( &x13_4way_ctx.simd, 512 );
init_echo( &x13_4way_ctx.echo, 512 );
sph_hamsi512_init( &x13_4way_ctx.hamsi );
hamsi512_4way_init( &x13_4way_ctx.hamsi );
sph_fugue512_init( &x13_4way_ctx.fugue );
};
@@ -85,7 +85,7 @@ void x13_4way_hash( void *state, const void *input )
memcpy( &ctx.groestl, &x13_4way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// Parallel 4way
// Parallel 4way 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein
@@ -100,7 +100,7 @@ void x13_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// Serial to the end
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 7 Luffa
@@ -167,20 +167,13 @@ void x13_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
// 12 Hamsi
sph_hamsi512( &ctx.hamsi, hash0, 64 );
sph_hamsi512_close( &ctx.hamsi, hash0 );
memcpy( &ctx.hamsi, &x13_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, hash1, 64 );
sph_hamsi512_close( &ctx.hamsi, hash1 );
memcpy( &ctx.hamsi, &x13_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, hash2, 64 );
sph_hamsi512_close( &ctx.hamsi, hash2 );
memcpy( &ctx.hamsi, &x13_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, hash3, 64 );
sph_hamsi512_close( &ctx.hamsi, hash3 );
// 12 Hamsi parallel 4way 32 bit
mm_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
// 13 Fugue
// 13 Fugue serial
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
memcpy( &ctx.fugue, &x13_4way_ctx.fugue, sizeof(sph_fugue512_context) );

2
algo/x13/x13-gate.c
View File

@@ -11,7 +11,7 @@ bool register_x13_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x13;
gate->hash = (void*)&x13hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

2
algo/x13/x13-gate.h
View File

@@ -4,7 +4,7 @@
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#if defined(__AVX2__) && defined(__AES__)
#define X13_4WAY
#endif

34
algo/x13/x13sm3-4way.c
View File

@@ -1,6 +1,6 @@
#include "x13sm3-gate.h"
#if defined(__AVX2__) && defined(__AES__)
#if defined(X13SM3_4WAY)
#include <stdlib.h>
#include <stdint.h>
@@ -18,7 +18,7 @@
#include "algo/simd/sse2/nist.h"
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/sm3/sm3-hash-4way.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/hamsi/hamsi-hash-4way.h"
#include "algo/fugue/sph_fugue.h"
typedef struct {
@@ -34,7 +34,7 @@ typedef struct {
hashState_sd simd;
hashState_echo echo;
sm3_4way_ctx_t sm3;
sph_hamsi512_context hamsi;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
} x13sm3_4way_ctx_holder;
@@ -55,7 +55,7 @@ void init_x13sm3_4way_ctx()
init_sd( &x13sm3_4way_ctx.simd, 512 );
init_echo( &x13sm3_4way_ctx.echo, 512 );
sm3_4way_init( &x13sm3_4way_ctx.sm3 );
sph_hamsi512_init( &x13sm3_4way_ctx.hamsi );
hamsi512_4way_init( &x13sm3_4way_ctx.hamsi );
sph_fugue512_init( &x13sm3_4way_ctx.fugue );
};
@@ -174,7 +174,9 @@ void x13sm3_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
// SM3
mm_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
// SM3 parallel 32 bit
uint32_t sm3_vhash[32*4] __attribute__ ((aligned (64)));
memset( sm3_vhash, 0, sizeof sm3_vhash );
uint32_t sm3_hash0[32] __attribute__ ((aligned (32)));
@@ -186,26 +188,16 @@ void x13sm3_4way_hash( void *state, const void *input )
uint32_t sm3_hash3[32] __attribute__ ((aligned (32)));
memset( sm3_hash3, 0, sizeof sm3_hash3 );
mm_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
sm3_4way( &ctx.sm3, vhash, 64 );
sm3_4way_close( &ctx.sm3, sm3_vhash );
mm_deinterleave_4x32( sm3_hash0, sm3_hash1, sm3_hash2, sm3_hash3,
sm3_vhash, 1024 );
// Hamsi
sph_hamsi512( &ctx.hamsi, sm3_hash0, 64 );
sph_hamsi512_close( &ctx.hamsi, hash0 );
memcpy( &ctx.hamsi, &x13sm3_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, sm3_hash1, 64 );
sph_hamsi512_close( &ctx.hamsi, hash1 );
memcpy( &ctx.hamsi, &x13sm3_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, sm3_hash2, 64 );
sph_hamsi512_close( &ctx.hamsi, hash2 );
memcpy( &ctx.hamsi, &x13sm3_4way_ctx.hamsi, sizeof(sph_hamsi512_context) );
sph_hamsi512( &ctx.hamsi, sm3_hash3, 64 );
sph_hamsi512_close( &ctx.hamsi, hash3 );
// Hamsi parallel 32 bit
hamsi512_4way( &ctx.hamsi, sm3_vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
// Fugue
mm_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
// Fugue serial
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
memcpy( &ctx.fugue, &x13sm3_4way_ctx.fugue, sizeof(sph_fugue512_context) );

2
algo/x13/x13sm3-gate.c
View File

@@ -11,7 +11,7 @@ bool register_x13sm3_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_x13sm3;
gate->hash = (void*)&x13sm3_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | FOUR_WAY_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
gate->get_max64 = (void*)&get_max64_0x3ffff;
return true;
};

2
algo/x13/x13sm3-gate.h
View File

@@ -4,7 +4,7 @@
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(HASH_4WAY) && defined(__AES__)
#if defined(__AVX2__) && defined(__AES__)
#define X13SM3_4WAY
#endif