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

v3.15.0

Browse Source
This commit is contained in:
Jay D Dee
2020-10-02 10:48:37 -04:00
parent cdd587537e
commit c85fb3842b
44 changed files with 1351 additions and 747 deletions
Download Patch File Download Diff File Expand all files Collapse all files

54
algo/quark/hmq1725-4way.c
View File

@@ -16,7 +16,7 @@
#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"
#include "algo/fugue/fugue-aesni.h"
#include "algo/shabal/shabal-hash-4way.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/haval/haval-hash-4way.h"
@@ -40,7 +40,7 @@ union _hmq1725_8way_context_overlay
cube_4way_context cube;
simd_4way_context simd;
hamsi512_8way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_8way_context shabal;
sph_whirlpool_context whirlpool;
sha512_8way_context sha512;
@@ -363,14 +363,14 @@ extern void hmq1725_8way_hash(void *state, const void *input)
dintrlv_8x64_512( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, vhash );
sph_fugue512_full( &ctx.fugue, hash0, hash0, 64 );
sph_fugue512_full( &ctx.fugue, hash1, hash1, 64 );
sph_fugue512_full( &ctx.fugue, hash2, hash2, 64 );
sph_fugue512_full( &ctx.fugue, hash3, hash3, 64 );
sph_fugue512_full( &ctx.fugue, hash4, hash4, 64 );
sph_fugue512_full( &ctx.fugue, hash5, hash5, 64 );
sph_fugue512_full( &ctx.fugue, hash6, hash6, 64 );
sph_fugue512_full( &ctx.fugue, hash7, hash7, 64 );
fugue512_full( &ctx.fugue, hash0, hash0, 64 );
fugue512_full( &ctx.fugue, hash1, hash1, 64 );
fugue512_full( &ctx.fugue, hash2, hash2, 64 );
fugue512_full( &ctx.fugue, hash3, hash3, 64 );
fugue512_full( &ctx.fugue, hash4, hash4, 64 );
fugue512_full( &ctx.fugue, hash5, hash5, 64 );
fugue512_full( &ctx.fugue, hash6, hash6, 64 );
fugue512_full( &ctx.fugue, hash7, hash7, 64 );
intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7 );
@@ -459,21 +459,21 @@ extern void hmq1725_8way_hash(void *state, const void *input)
m512_zero );
if ( hash0[0] & mask )
sph_fugue512_full( &ctx.fugue, hash0, hash0, 64 );
fugue512_full( &ctx.fugue, hash0, hash0, 64 );
if ( hash1[0] & mask )
sph_fugue512_full( &ctx.fugue, hash1, hash1, 64 );
fugue512_full( &ctx.fugue, hash1, hash1, 64 );
if ( hash2[0] & mask )
sph_fugue512_full( &ctx.fugue, hash2, hash2, 64 );
fugue512_full( &ctx.fugue, hash2, hash2, 64 );
if ( hash3[0] & mask )
sph_fugue512_full( &ctx.fugue, hash3, hash3, 64 );
fugue512_full( &ctx.fugue, hash3, hash3, 64 );
if ( hash4[0] & mask )
sph_fugue512_full( &ctx.fugue, hash4, hash4, 64 );
fugue512_full( &ctx.fugue, hash4, hash4, 64 );
if ( hash5[0] & mask )
sph_fugue512_full( &ctx.fugue, hash5, hash5, 64 );
fugue512_full( &ctx.fugue, hash5, hash5, 64 );
if ( hash6[0] & mask )
sph_fugue512_full( &ctx.fugue, hash6, hash6, 64 );
fugue512_full( &ctx.fugue, hash6, hash6, 64 );
if ( hash7[0] & mask )
sph_fugue512_full( &ctx.fugue, hash7, hash7, 64 );
fugue512_full( &ctx.fugue, hash7, hash7, 64 );
intrlv_8x64_512( vhashA, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7 );
@@ -628,7 +628,7 @@ union _hmq1725_4way_context_overlay
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_4way_context shabal;
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
@@ -846,10 +846,10 @@ extern void hmq1725_4way_hash(void *state, const void *input)
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_full( &ctx.fugue, hash0, hash0, 64 );
sph_fugue512_full( &ctx.fugue, hash1, hash1, 64 );
sph_fugue512_full( &ctx.fugue, hash2, hash2, 64 );
sph_fugue512_full( &ctx.fugue, hash3, hash3, 64 );
fugue512_full( &ctx.fugue, hash0, hash0, 64 );
fugue512_full( &ctx.fugue, hash1, hash1, 64 );
fugue512_full( &ctx.fugue, hash2, hash2, 64 );
fugue512_full( &ctx.fugue, hash3, hash3, 64 );
// In this situation serial simd seems to be faster.
@@ -920,13 +920,13 @@ extern void hmq1725_4way_hash(void *state, const void *input)
h_mask = _mm256_movemask_epi8( vh_mask );
if ( hash0[0] & mask )
sph_fugue512_full( &ctx.fugue, hash0, hash0, 64 );
fugue512_full( &ctx.fugue, hash0, hash0, 64 );
if ( hash1[0] & mask )
sph_fugue512_full( &ctx.fugue, hash1, hash1, 64 );
fugue512_full( &ctx.fugue, hash1, hash1, 64 );
if ( hash2[0] & mask )
sph_fugue512_full( &ctx.fugue, hash2, hash2, 64 );
fugue512_full( &ctx.fugue, hash2, hash2, 64 );
if ( hash3[0] & mask )
sph_fugue512_full( &ctx.fugue, hash3, hash3, 64 );
fugue512_full( &ctx.fugue, hash3, hash3, 64 );
intrlv_4x64( vhashA, hash0, hash1, hash2, hash3, 512 );

20
algo/quark/hmq1725.c
View File

@@ -21,9 +21,11 @@
#if defined(__AES__)
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/fugue/fugue-aesni.h"
#else
#include "algo/groestl/sph_groestl.h"
#include "algo/echo/sph_echo.h"
#include "algo/fugue/sph_fugue.h"
#endif
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
@@ -40,7 +42,6 @@ typedef struct {
sph_shavite512_context shavite1, shavite2;
hashState_sd simd1, simd2;
sph_hamsi512_context hamsi1;
sph_fugue512_context fugue1, fugue2;
sph_shabal512_context shabal1;
sph_whirlpool_context whirlpool1, whirlpool2, whirlpool3, whirlpool4;
SHA512_CTX sha1, sha2;
@@ -48,9 +49,11 @@ typedef struct {
#if defined(__AES__)
hashState_echo echo1, echo2;
hashState_groestl groestl1, groestl2;
hashState_fugue fugue1, fugue2;
#else
sph_groestl512_context groestl1, groestl2;
sph_echo512_context echo1, echo2;
sph_fugue512_context fugue1, fugue2;
#endif
} hmq1725_ctx_holder;
@@ -88,8 +91,13 @@ void init_hmq1725_ctx()
sph_hamsi512_init(&hmq1725_ctx.hamsi1);
#if defined(__AES__)
fugue512_Init( &hmq1725_ctx.fugue1, 512 );
fugue512_Init( &hmq1725_ctx.fugue2, 512 );
#else
sph_fugue512_init(&hmq1725_ctx.fugue1);
sph_fugue512_init(&hmq1725_ctx.fugue2);
#endif
sph_shabal512_init(&hmq1725_ctx.shabal1);
@@ -235,8 +243,13 @@ extern void hmq1725hash(void *state, const void *input)
sph_hamsi512 (&h_ctx.hamsi1, hashA, 64); //3
sph_hamsi512_close(&h_ctx.hamsi1, hashB); //4
#if defined(__AES__)
fugue512_Update( &h_ctx.fugue1, hashB, 512 ); //2 ////
fugue512_Final( &h_ctx.fugue1, hashA ); //3
#else
sph_fugue512 (&h_ctx.fugue1, hashB, 64); //2 ////
sph_fugue512_close(&h_ctx.fugue1, hashA); //3
#endif
if ( hashA[0] & mask ) //4
{
@@ -262,8 +275,13 @@ extern void hmq1725hash(void *state, const void *input)
if ( hashB[0] & mask ) //7
{
#if defined(__AES__)
fugue512_Update( &h_ctx.fugue2, hashB, 512 ); //
fugue512_Final( &h_ctx.fugue2, hashA ); //8
#else
sph_fugue512 (&h_ctx.fugue2, hashB, 64); //
sph_fugue512_close(&h_ctx.fugue2, hashA); //8
#endif
}
else
{