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

v3.9.1

Browse Source
This commit is contained in:
Jay D Dee
2019-05-30 16:59:49 -04:00
parent eb3f57bfc7
commit 77c5ae80ab
82 changed files with 6906 additions and 3706 deletions
Download Patch File Download Diff File Expand all files Collapse all files

75
algo/x17/hmq1725.c
View File

@@ -16,10 +16,9 @@
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/sha/sph_sha2.h"
#include "algo/haval/sph-haval.h"
#include <openssl/sha.h>
#ifndef NO_AES_NI
#if defined(__AES__)
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/echo/aes_ni/hash_api.h"
#endif
@@ -42,18 +41,14 @@ typedef struct {
sph_fugue512_context fugue1, fugue2;
sph_shabal512_context shabal1;
sph_whirlpool_context whirlpool1, whirlpool2, whirlpool3, whirlpool4;
#ifndef USE_SPH_SHA
SHA512_CTX sha1, sha2;
#else
sph_sha512_context sha1, sha2;
#endif
sph_haval256_5_context haval1, haval2;
#ifdef NO_AES_NI
sph_groestl512_context groestl1, groestl2;
sph_echo512_context echo1, echo2;
#else
#if defined(__AES__)
hashState_echo echo1, echo2;
hashState_groestl groestl1, groestl2;
#else
sph_groestl512_context groestl1, groestl2;
sph_echo512_context echo1, echo2;
#endif
} hmq1725_ctx_holder;
@@ -101,26 +96,22 @@ void init_hmq1725_ctx()
sph_whirlpool_init(&hmq1725_ctx.whirlpool3);
sph_whirlpool_init(&hmq1725_ctx.whirlpool4);
#ifndef USE_SPH_SHA
SHA512_Init( &hmq1725_ctx.sha1 );
SHA512_Init( &hmq1725_ctx.sha2 );
#else
sph_sha512_init(&hmq1725_ctx.sha1);
sph_sha512_init(&hmq1725_ctx.sha2);
#endif
sph_haval256_5_init(&hmq1725_ctx.haval1);
sph_haval256_5_init(&hmq1725_ctx.haval2);
#ifdef NO_AES_NI
sph_groestl512_init( &hmq1725_ctx.groestl1 );
sph_groestl512_init( &hmq1725_ctx.groestl2 );
sph_echo512_init( &hmq1725_ctx.echo1 );
sph_echo512_init( &hmq1725_ctx.echo2 );
#else
#if defined(__AES__)
init_echo( &hmq1725_ctx.echo1, 512 );
init_echo( &hmq1725_ctx.echo2, 512 );
init_groestl( &hmq1725_ctx.groestl1, 64 );
init_groestl( &hmq1725_ctx.groestl2, 64 );
#else
sph_groestl512_init( &hmq1725_ctx.groestl1 );
sph_groestl512_init( &hmq1725_ctx.groestl2 );
sph_echo512_init( &hmq1725_ctx.echo1 );
sph_echo512_init( &hmq1725_ctx.echo2 );
#endif
}
@@ -151,12 +142,12 @@ extern void hmq1725hash(void *state, const void *input)
if ( hashB[0] & mask ) //1
{
#ifdef NO_AES_NI
#if defined(__AES__)
update_and_final_groestl( &h_ctx.groestl1, (char*)hashA,
(const char*)hashB, 512 );
#else
sph_groestl512 (&h_ctx.groestl1, hashB, 64); //1
sph_groestl512_close(&h_ctx.groestl1, hashA); //2
#else
update_and_final_groestl( &h_ctx.groestl1, (char*)hashA,
(const char*)hashB, 512 );
#endif
}
else
@@ -217,12 +208,12 @@ extern void hmq1725hash(void *state, const void *input)
memset(&hashB[8], 0, 32);
}
#ifdef NO_AES_NI
sph_echo512 (&h_ctx.echo1, hashB, 64); //5
sph_echo512_close(&h_ctx.echo1, hashA); //6
#else
#if defined(__AES__)
update_final_echo ( &h_ctx.echo1, (BitSequence *)hashA,
(const BitSequence *)hashB, 512 );
#else
sph_echo512 (&h_ctx.echo1, hashB, 64); //5
sph_echo512_close(&h_ctx.echo1, hashA); //6
#endif
sph_blake512 (&h_ctx.blake2, hashA, 64); //6
@@ -247,12 +238,12 @@ extern void hmq1725hash(void *state, const void *input)
if ( hashA[0] & mask ) //4
{
#ifdef NO_AES_NI
sph_echo512 (&h_ctx.echo2, hashA, 64); //
sph_echo512_close(&h_ctx.echo2, hashB); //5
#else
#if defined(__AES__)
update_final_echo ( &h_ctx.echo2, (BitSequence *)hashB,
(const BitSequence *)hashA, 512 );
#else
sph_echo512 (&h_ctx.echo2, hashA, 64); //
sph_echo512_close(&h_ctx.echo2, hashB); //5
#endif
}
else
@@ -274,30 +265,20 @@ extern void hmq1725hash(void *state, const void *input)
}
else
{
#ifndef USE_SPH_SHA
SHA512_Update( &h_ctx.sha1, hashB, 64 );
SHA512_Final( (unsigned char*) hashA, &h_ctx.sha1 );
#else
sph_sha512 (&h_ctx.sha1, hashB, 64); //7
sph_sha512_close(&h_ctx.sha1, hashA); //8
#endif
}
#ifdef NO_AES_NI
sph_groestl512 (&h_ctx.groestl2, hashA, 64); //3
sph_groestl512_close(&h_ctx.groestl2, hashB); //4
#else
#if defined(__AES__)
update_and_final_groestl( &h_ctx.groestl2, (char*)hashB,
(const char*)hashA, 512 );
#else
sph_groestl512 (&h_ctx.groestl2, hashA, 64); //3
sph_groestl512_close(&h_ctx.groestl2, hashB); //4
#endif
#ifndef USE_SPH_SHA
SHA512_Update( &h_ctx.sha2, hashB, 64 );
SHA512_Final( (unsigned char*) hashA, &h_ctx.sha2 );
#else
sph_sha512 (&h_ctx.sha2, hashB, 64); //2
sph_sha512_close(&h_ctx.sha2, hashA); //3
#endif
if ( hashA[0] & mask ) //4
{

872
algo/x17/sonoa-4way.c Normal file
View File

@@ -0,0 +1,872 @@
#include "sonoa-gate.h"
#if defined(SONOA_4WAY)
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/blake/blake-hash-4way.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"
#include "algo/keccak/keccak-hash-4way.h"
#include "algo/luffa/luffa-hash-2way.h"
#include "algo/cubehash/cube-hash-2way.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/shavite/shavite-hash-2way.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"
#include "algo/shabal/shabal-hash-4way.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/haval/haval-hash-4way.h"
#include "algo/sha/sha2-hash-4way.h"
typedef struct {
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
luffa_2way_context luffa;
cube_2way_context cube;
shavite512_2way_context shavite;
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
shabal512_4way_context shabal;
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
haval256_5_4way_context haval;
} sonoa_4way_ctx_holder;
sonoa_4way_ctx_holder sonoa_4way_ctx __attribute__ ((aligned (64)));
void init_sonoa_4way_ctx()
{
blake512_4way_init( &sonoa_4way_ctx.blake );
bmw512_4way_init( &sonoa_4way_ctx.bmw );
init_groestl( &sonoa_4way_ctx.groestl, 64 );
skein512_4way_init( &sonoa_4way_ctx.skein );
jh512_4way_init( &sonoa_4way_ctx.jh );
keccak512_4way_init( &sonoa_4way_ctx.keccak );
luffa_2way_init( &sonoa_4way_ctx.luffa, 512 );
cube_2way_init( &sonoa_4way_ctx.cube, 512, 16, 32 );
shavite512_2way_init( &sonoa_4way_ctx.shavite );
simd_2way_init( &sonoa_4way_ctx.simd, 512 );
init_echo( &sonoa_4way_ctx.echo, 512 );
hamsi512_4way_init( &sonoa_4way_ctx.hamsi );
sph_fugue512_init( &sonoa_4way_ctx.fugue );
shabal512_4way_init( &sonoa_4way_ctx.shabal );
sph_whirlpool_init( &sonoa_4way_ctx.whirlpool );
sha512_4way_init( &sonoa_4way_ctx.sha512 );
haval256_5_4way_init( &sonoa_4way_ctx.haval );
};
void sonoa_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)));
uint64_t vhashA[8*4] __attribute__ ((aligned (64)));
uint64_t vhashB[8*4] __attribute__ ((aligned (64)));
sonoa_4way_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &sonoa_4way_ctx, sizeof(sonoa_4way_ctx) );
// 1
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
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 );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 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 );
// 2
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
// 3
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
// 4
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm256_reinterleave_4x64( vhashB, vhash, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhashB, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
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 );
mm256_interleave_2x128( vhashA, hash0, hash1, 512 );
mm256_interleave_2x128( vhashB, hash2, hash3, 512 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
// 5
mm256_reinterleave_2x128_4x64( vhash, vhashA, vhashB, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_reinterleave_4x32( vhashB, vhash, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhashB, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash1, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash2, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
// 6
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash1, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash2, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash1, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash2, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
// 7
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way_init( &ctx.keccak );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
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 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash1, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash2, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
mm256_reinterleave_4x32( vhashB, vhash, 512 );
haval256_5_4way( &ctx.haval, vhashB, 64 );
haval256_5_4way_close( &ctx.haval, state );
}
int scanhash_sonoa_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t *hash7 = &(hash[7<<2]);
uint32_t lane_hash[8];
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;
int num_found = 0;
__m256i *noncev = (__m256i*)vdata + 9; // aligned
// uint32_t *noncep = vdata + 73; // 9*8 + 1
const uint32_t Htarg = ptarget[7];
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
uint32_t masks[] = { 0xFFFFFFFF, 0xFFFFFFF0, 0xFFFFFF00,
0xFFFFF000, 0xFFFF0000, 0 };
// Need big endian data
casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
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
{
*noncev = mm256_interleave_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ),
*noncev );
sonoa_4way_hash( hash, vdata );
for ( int lane = 0; lane < 4; lane++ )
if ( ( ( hash7[ lane ] & mask ) == 0 ) )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
{
pdata[19] = n + lane;
nonces[ num_found++ ] = n + lane;
work_set_target_ratio( work, lane_hash );
}
}
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/x17/sonoa-gate.c Normal file
View File

@@ -0,0 +1,18 @@
#include "sonoa-gate.h"
bool register_sonoa_algo( algo_gate_t* gate )
{
#if defined (SONOA_4WAY)
init_sonoa_4way_ctx();
gate->scanhash = (void*)&scanhash_sonoa_4way;
gate->hash = (void*)&sonoa_4way_hash;
#else
init_sonoa_ctx();
gate->scanhash = (void*)&scanhash_sonoa;
gate->hash = (void*)&sonoa_hash;
#endif
gate->get_max64 = (void*)&get_max64_0x1ffff;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
return true;
};

32
algo/x17/sonoa-gate.h Normal file
View File

@@ -0,0 +1,32 @@
#ifndef SONOA_GATE_H__
#define SONOA_GATE_H__ 1
#include "algo-gate-api.h"
#include <stdint.h>
#if defined(__AVX2__) && defined(__AES__)
#define SONOA_4WAY
#endif
bool register_sonoa_algo( algo_gate_t* gate );
#if defined(SONOA_4WAY)
void sonoa_4way_hash( void *state, const void *input );
int scanhash_sonoa_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_sonoa_4way_ctx();
#endif
void sonoa_hash( void *state, const void *input );
int scanhash_sonoa( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_sonoa_ctx();
#endif

648
algo/x17/sonoa.c Normal file
View File

@@ -0,0 +1,648 @@
#include "sonoa-gate.h"
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/haval/sph-haval.h"
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
#include "algo/blake/sse2/blake.c"
#include "algo/bmw/sse2/bmw.c"
#include "algo/keccak/sse2/keccak.c"
#include "algo/skein/sse2/skein.c"
#include "algo/jh/sse2/jh_sse2_opt64.h"
#include <openssl/sha.h>
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#else
#include "algo/groestl/sph_groestl.h"
#include "algo/echo/sph_echo.h"
#endif
typedef struct {
sph_blake512_context blake;
sph_bmw512_context bmw;
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
#endif
sph_jh512_context jh;
sph_keccak512_context keccak;
sph_skein512_context skein;
hashState_luffa luffa;
cubehashParam cubehash;
sph_shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;
sph_haval256_5_context haval;
} sonoa_ctx_holder;
sonoa_ctx_holder sonoa_ctx __attribute__ ((aligned (64)));
void init_sonoa_ctx()
{
sph_blake512_init( &sonoa_ctx.blake);
sph_bmw512_init( &sonoa_ctx.bmw);
#if defined(__AES__)
init_echo( &sonoa_ctx.echo, 512 );
init_groestl( &sonoa_ctx.groestl, 64 );
#else
sph_groestl512_init(&sonoa_ctx.groestl );
sph_echo512_init( &sonoa_ctx.echo );
#endif
sph_skein512_init( &sonoa_ctx.skein);
sph_jh512_init( &sonoa_ctx.jh);
sph_keccak512_init( &sonoa_ctx.keccak );
init_luffa( &sonoa_ctx.luffa, 512 );
cubehashInit( &sonoa_ctx.cubehash, 512, 16, 32 );
sph_shavite512_init( &sonoa_ctx.shavite );
init_sd( &sonoa_ctx.simd, 512 );
sph_hamsi512_init( &sonoa_ctx.hamsi );
sph_fugue512_init( &sonoa_ctx.fugue );
sph_shabal512_init( &sonoa_ctx.shabal );
sph_whirlpool_init( &sonoa_ctx.whirlpool );
SHA512_Init( &sonoa_ctx.sha512 );
sph_haval256_5_init(&sonoa_ctx.haval);
};
void sonoa_hash( void *state, const void *input )
{
uint8_t hash[128] __attribute__ ((aligned (64)));
sonoa_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &sonoa_ctx, sizeof(sonoa_ctx) );
sph_blake512(&ctx.blake, input, 80);
sph_blake512_close(&ctx.blake, hash);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hash);
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_fugue512_init( &ctx.fugue );
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hash);
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close(&ctx.shabal, hash);
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
sph_shabal512_init( &ctx.shabal );
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close(&ctx.shabal, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_fugue512_init( &ctx.fugue );
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hash);
sph_shabal512_init( &ctx.shabal );
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close(&ctx.shabal, hash);
sph_whirlpool(&ctx.whirlpool, hash, 64);
sph_whirlpool_close(&ctx.whirlpool, hash);
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_fugue512_init( &ctx.fugue );
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hash);
sph_shabal512_init( &ctx.shabal );
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close(&ctx.shabal, hash);
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool(&ctx.whirlpool, hash, 64);
sph_whirlpool_close(&ctx.whirlpool, hash);
SHA512_Update( &ctx.sha512, hash, 64 );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool(&ctx.whirlpool, hash, 64);
sph_whirlpool_close(&ctx.whirlpool, hash);
//
sph_bmw512_init( &ctx.bmw);
sph_bmw512(&ctx.bmw, hash, 64);
sph_bmw512_close(&ctx.bmw, hash);
#if defined(__AES__)
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512_init(&ctx.groestl );
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512_init( &ctx.skein);
sph_skein512(&ctx.skein, hash, 64);
sph_skein512_close(&ctx.skein, hash);
sph_jh512_init( &ctx.jh);
sph_jh512(&ctx.jh, hash, 64);
sph_jh512_close(&ctx.jh, hash);
sph_keccak512_init( &ctx.keccak );
sph_keccak512(&ctx.keccak, hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
cubehashInit( &ctx.cubehash, 512, 16, 32 );
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hash, 64 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512(&ctx.shavite, hash, 64);
sph_shavite512_close(&ctx.shavite, hash);
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#if defined(__AES__)
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence *)hash, 512 );
#else
sph_echo512_init( &ctx.echo );
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512(&ctx.hamsi, hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_fugue512_init( &ctx.fugue );
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hash);
sph_shabal512_init( &ctx.shabal );
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close(&ctx.shabal, hash);
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool(&ctx.whirlpool, hash, 64);
sph_whirlpool_close(&ctx.whirlpool, hash);
SHA512_Init( &ctx.sha512 );
SHA512_Update( &ctx.sha512, hash, 64 );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
sph_haval256_5(&ctx.haval,(const void*) hash, 64);
sph_haval256_5_close(&ctx.haval, hash);
memcpy(state, hash, 32);
}
int scanhash_sonoa( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
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;
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
uint64_t htmax[] =
{
0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000
};
uint32_t masks[] =
{
0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0
};
// we need bigendian data...
casti_m128i( endiandata, 0 ) = mm128_bswap_32( casti_m128i( pdata, 0 ) );
casti_m128i( endiandata, 1 ) = mm128_bswap_32( casti_m128i( pdata, 1 ) );
casti_m128i( endiandata, 2 ) = mm128_bswap_32( casti_m128i( pdata, 2 ) );
casti_m128i( endiandata, 3 ) = mm128_bswap_32( casti_m128i( pdata, 3 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
#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);
sonoa_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;
}

354
algo/x17/x16r-4way.c
View File

@@ -1,354 +0,0 @@
/**
* x16r algo implementation
*
* Implementation by tpruvot@github Jan 2018
* Optimized by JayDDee@github Jan 2018
*/
#include "x16r-gate.h"
#if defined (X16R_4WAY)
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/bmw-hash-4way.h"
#include "algo/groestl/aes_ni/hash-groestl.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/shavite/sph_shavite.h"
#include "algo/luffa/luffa-hash-2way.h"
#include "algo/cubehash/cubehash_sse2.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"
#include "algo/shabal/shabal-hash-4way.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/sha/sha2-hash-4way.h"
static __thread uint32_t s_ntime = UINT32_MAX;
static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };
typedef struct {
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_echo echo;
hashState_groestl groestl;
skein512_4way_context skein;
jh512_4way_context jh;
keccak512_4way_context keccak;
luffa_2way_context luffa;
cubehashParam cube;
sph_shavite512_context shavite;
simd_2way_context simd;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
shabal512_4way_context shabal;
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
} x16r_4way_ctx_holder;
x16r_4way_ctx_holder x16r_4way_ctx __attribute__ ((aligned (64)));
// Cube needs one full init so fast reinits can be done in the hash loop.
void init_x16r_4way_ctx()
{
cubehashInit( &x16r_4way_ctx.cube, 512, 16, 32 );
};
void x16r_4way_hash( void* output, const void* input )
{
uint32_t hash0[24] __attribute__ ((aligned (64)));
uint32_t hash1[24] __attribute__ ((aligned (64)));
uint32_t hash2[24] __attribute__ ((aligned (64)));
uint32_t hash3[24] __attribute__ ((aligned (64)));
uint32_t vhash[24*4] __attribute__ ((aligned (64)));
x16r_4way_ctx_holder ctx;
void *in0 = (void*) hash0;
void *in1 = (void*) hash1;
void *in2 = (void*) hash2;
void *in3 = (void*) hash3;
int size = 80;
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, input, 640 );
if ( s_ntime == UINT32_MAX )
{
const uint8_t* tmp = (uint8_t*) in0;
x16_r_s_getAlgoString( &tmp[4], hashOrder );
}
// Input data is both 64 bit interleaved (input)
// and deinterleaved in inp0-3.
// If First function uses 64 bit data it is not required to interleave inp
// first. It may use the inerleaved data dmost convenient, ie 4way 64 bit.
// All other functions assume data is deinterleaved in hash0-3
// All functions must exit with data deinterleaved in hash0-3.
// Alias in0-3 points to either inp0-3 or hash0-3 according to
// its hashOrder position. Size is also set accordingly.
for ( int i = 0; i < 16; i++ )
{
const char elem = hashOrder[i];
const uint8_t algo = elem >= 'A' ? elem - 'A' + 10 : elem - '0';
switch ( algo )
{
case BLAKE:
blake512_4way_init( &ctx.blake );
if ( i == 0 )
blake512_4way( &ctx.blake, input, size );
else
{
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
blake512_4way( &ctx.blake, vhash, size );
}
blake512_4way_close( &ctx.blake, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case BMW:
bmw512_4way_init( &ctx.bmw );
if ( i == 0 )
bmw512_4way( &ctx.bmw, input, size );
else
{
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
bmw512_4way( &ctx.bmw, vhash, size );
}
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case GROESTL:
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0,
(const char*)in0, size<<3 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1,
(const char*)in1, size<<3 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2,
(const char*)in2, size<<3 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3,
(const char*)in3, size<<3 );
break;
case SKEIN:
skein512_4way_init( &ctx.skein );
if ( i == 0 )
skein512_4way( &ctx.skein, input, size );
else
{
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
skein512_4way( &ctx.skein, vhash, size );
}
skein512_4way_close( &ctx.skein, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case JH:
jh512_4way_init( &ctx.jh );
if ( i == 0 )
jh512_4way( &ctx.jh, input, size );
else
{
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
jh512_4way( &ctx.jh, vhash, size );
}
jh512_4way_close( &ctx.jh, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case KECCAK:
keccak512_4way_init( &ctx.keccak );
if ( i == 0 )
keccak512_4way( &ctx.keccak, input, size );
else
{
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
keccak512_4way( &ctx.keccak, vhash, size );
}
keccak512_4way_close( &ctx.keccak, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case LUFFA:
mm256_interleave_2x128( vhash, in0, in1, size<<3 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, size );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, in2, in3, size<<3 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, size);
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
break;
case CUBEHASH:
cubehashReinit( &ctx.cube );
cubehashUpdateDigest( &ctx.cube, (byte*) hash0,
(const byte*)in0, size );
cubehashReinit( &ctx.cube );
cubehashUpdateDigest( &ctx.cube, (byte*) hash1,
(const byte*)in1, size );
cubehashReinit( &ctx.cube );
cubehashUpdateDigest( &ctx.cube, (byte*) hash2,
(const byte*)in2, size );
cubehashReinit( &ctx.cube );
cubehashUpdateDigest( &ctx.cube, (byte*) hash3,
(const byte*)in3, size );
break;
case SHAVITE:
sph_shavite512_init( &ctx.shavite );
sph_shavite512( &ctx.shavite, in0, size );
sph_shavite512_close( &ctx.shavite, hash0 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512( &ctx.shavite, in1, size );
sph_shavite512_close( &ctx.shavite, hash1 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512( &ctx.shavite, in2, size );
sph_shavite512_close( &ctx.shavite, hash2 );
sph_shavite512_init( &ctx.shavite );
sph_shavite512( &ctx.shavite, in3, size );
sph_shavite512_close( &ctx.shavite, hash3 );
break;
case SIMD:
mm256_interleave_2x128( vhash, in0, in1, size<<3 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, size<<3 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, in2, in3, size<<3 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, size<<3 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
break;
case ECHO:
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash0,
(const BitSequence*)in0, size<<3 );
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash1,
(const BitSequence*)in1, size<<3 );
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash2,
(const BitSequence*)in2, size<<3 );
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash3,
(const BitSequence*)in3, size<<3 );
break;
case HAMSI:
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, size );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case FUGUE:
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, in0, size );
sph_fugue512_close( &ctx.fugue, hash0 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, in1, size );
sph_fugue512_close( &ctx.fugue, hash1 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, in2, size );
sph_fugue512_close( &ctx.fugue, hash2 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, in3, size );
sph_fugue512_close( &ctx.fugue, hash3 );
break;
case SHABAL:
mm128_interleave_4x32( vhash, in0, in1, in2, in3, size<<3 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, size );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
break;
case WHIRLPOOL:
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, in0, size );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, in1, size );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, in2, size );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, in3, size );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
break;
case SHA_512:
mm256_interleave_4x64( vhash, in0, in1, in2, in3, size<<3 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, size );
sha512_4way_close( &ctx.sha512, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
}
size = 64;
}
memcpy( output, hash0, 32 );
memcpy( output+32, hash1, 32 );
memcpy( output+64, hash2, 32 );
memcpy( output+96, hash3, 32 );
}
int scanhash_x16r_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
{
uint32_t hash[4*16] __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;
const uint32_t Htarg = ptarget[7];
const uint32_t first_nonce = pdata[19];
uint32_t n = first_nonce;
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
volatile uint8_t *restart = &(work_restart[thr_id].restart);
for ( int k=0; k < 19; k++ )
be32enc( &endiandata[k], pdata[k] );
// if ( s_ntime != pdata[17] )
if ( s_ntime != endiandata[17] )
{
uint32_t ntime = swab32(pdata[17]);
x16_r_s_getAlgoString( (const uint8_t*) (&endiandata[1]), hashOrder );
s_ntime = ntime;
if ( opt_debug && !thr_id )
applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
}
if ( opt_benchmark )
ptarget[7] = 0x0cff;
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
do
{
be32enc( noncep, n );
be32enc( noncep+2, n+1 );
be32enc( noncep+4, n+2 );
be32enc( noncep+6, n+3 );
x16r_4way_hash( hash, vdata );
pdata[19] = n;
for ( int i = 0; i < 4; i++ )
if ( (hash+(i<<3))[7] <= Htarg && fulltest( hash+(i<<3), ptarget ) )
{
pdata[19] = n+i;
nonces[ num_found++ ] = n+i;
work_set_target_ratio( work, hash+(i<<3) );
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce ) && !(*restart) );
*hashes_done = n - first_nonce + 1;
return num_found;
}
#endif

68
algo/x17/x16r-gate.c
View File

@@ -1,68 +0,0 @@
#include "x16r-gate.h"
void x16r_getAlgoString( const uint8_t* prevblock, char *output )
{
char *sptr = output;
for ( int j = 0; j < X16R_HASH_FUNC_COUNT; j++ )
{
uint8_t b = (15 - j) >> 1; // 16 first ascii hex chars (lsb in uint256)
uint8_t algoDigit = (j & 1) ? prevblock[b] & 0xF : prevblock[b] >> 4;
if (algoDigit >= 10)
sprintf(sptr, "%c", 'A' + (algoDigit - 10));
else
sprintf(sptr, "%u", (uint32_t) algoDigit);
sptr++;
}
*sptr = '\0';
}
void x16s_getAlgoString( const uint8_t* prevblock, char *output )
{
strcpy( output, "0123456789ABCDEF" );
for ( int i = 0; i < 16; i++ )
{
uint8_t b = (15 - i) >> 1; // 16 ascii hex chars, reversed
uint8_t algoDigit = (i & 1) ? prevblock[b] & 0xF : prevblock[b] >> 4;
int offset = algoDigit;
// insert the nth character at the front
char oldVal = output[offset];
for( int j = offset; j-- > 0; )
output[j+1] = output[j];
output[0] = oldVal;
}
}
bool register_x16r_algo( algo_gate_t* gate )
{
#if defined (X16R_4WAY)
init_x16r_4way_ctx();
gate->scanhash = (void*)&scanhash_x16r_4way;
gate->hash = (void*)&x16r_4way_hash;
#else
init_x16r_ctx();
gate->scanhash = (void*)&scanhash_x16r;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
gate->set_target = (void*)&alt_set_target;
x16_r_s_getAlgoString = (void*)&x16r_getAlgoString;
return true;
};
bool register_x16s_algo( algo_gate_t* gate )
{
#if defined (X16R_4WAY)
init_x16r_4way_ctx();
gate->scanhash = (void*)&scanhash_x16r_4way;
gate->hash = (void*)&x16r_4way_hash;
#else
init_x16r_ctx();
gate->scanhash = (void*)&scanhash_x16r;
gate->hash = (void*)&x16r_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT;
gate->set_target = (void*)&alt_set_target;
x16_r_s_getAlgoString = (void*)&x16s_getAlgoString;
return true;
};

58
algo/x17/x16r-gate.h
View File

@@ -1,58 +0,0 @@
#ifndef X16R_GATE_H__
#define X16R_GATE_H__ 1
#include "algo-gate-api.h"
#include "avxdefs.h"
#include <stdint.h>
#if defined(__AVX2__) && defined(__AES__)
#define X16R_4WAY
#endif
enum x16r_Algo {
BLAKE = 0,
BMW,
GROESTL,
JH,
KECCAK,
SKEIN,
LUFFA,
CUBEHASH,
SHAVITE,
SIMD,
ECHO,
HAMSI,
FUGUE,
SHABAL,
WHIRLPOOL,
SHA_512,
X16R_HASH_FUNC_COUNT
};
void (*x16_r_s_getAlgoString) ( const uint8_t*, char* );
void x16r_getAlgoString( const uint8_t* prevblock, char *output );
void x16s_getAlgoString( const uint8_t* prevblock, char *output );
bool register_x16r_algo( algo_gate_t* gate );
bool register_x16s_algo( algo_gate_t* gate );
#if defined(X16R_4WAY)
void x16r_4way_hash( void *state, const void *input );
int scanhash_x16r_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_x16r_4way_ctx();
#endif
void x16r_hash( void *state, const void *input );
int scanhash_x16r( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
void init_x16r_ctx();
#endif

236
algo/x17/x16r.c
View File

@@ -1,236 +0,0 @@
/**
* x16r algo implementation
*
* Implementation by tpruvot@github Jan 2018
* Optimized by JayDDee@github Jan 2018
*/
#include "x16r-gate.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
#include "algo/echo/sph_echo.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include <openssl/sha.h>
#ifndef NO_AES_NI
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#endif
static __thread uint32_t s_ntime = UINT32_MAX;
static __thread char hashOrder[X16R_HASH_FUNC_COUNT + 1] = { 0 };
typedef struct {
#ifdef NO_AES_NI
sph_groestl512_context groestl;
sph_echo512_context echo;
#else
hashState_echo echo;
hashState_groestl groestl;
#endif
sph_blake512_context blake;
sph_bmw512_context bmw;
sph_skein512_context skein;
sph_jh512_context jh;
sph_keccak512_context keccak;
hashState_luffa luffa;
cubehashParam cube;
sph_shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;
} x16r_ctx_holder;
x16r_ctx_holder x16r_ctx __attribute__ ((aligned (64)));
void init_x16r_ctx()
{
cubehashInit( &x16r_ctx.cube, 512, 16, 32 );
};
void x16r_hash( void* output, const void* input )
{
uint32_t _ALIGN(128) hash[16];
x16r_ctx_holder ctx;
void *in = (void*) input;
int size = 80;
if ( s_ntime == UINT32_MAX )
{
const uint8_t* in8 = (uint8_t*) input;
x16_r_s_getAlgoString( &in8[4], hashOrder );
}
for ( int i = 0; i < 16; i++ )
{
const char elem = hashOrder[i];
const uint8_t algo = elem >= 'A' ? elem - 'A' + 10 : elem - '0';
switch ( algo )
{
case BLAKE:
sph_blake512_init( &ctx.blake );
sph_blake512( &ctx.blake, in, size );
sph_blake512_close( &ctx.blake, hash );
break;
case BMW:
sph_bmw512_init( &ctx.bmw );
sph_bmw512(&ctx.bmw, in, size);
sph_bmw512_close(&ctx.bmw, hash);
break;
case GROESTL:
#ifdef NO_AES_NI
sph_groestl512_init( &ctx.groestl );
sph_groestl512( &ctx.groestl, in, size );
sph_groestl512_close(&ctx.groestl, hash);
#else
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)in, size<<3 );
#endif
break;
case SKEIN:
sph_skein512_init( &ctx.skein );
sph_skein512( &ctx.skein, in, size );
sph_skein512_close( &ctx.skein, hash );
break;
case JH:
sph_jh512_init( &ctx.jh );
sph_jh512(&ctx.jh, in, size );
sph_jh512_close(&ctx.jh, hash );
break;
case KECCAK:
sph_keccak512_init( &ctx.keccak );
sph_keccak512( &ctx.keccak, in, size );
sph_keccak512_close( &ctx.keccak, hash );
break;
case LUFFA:
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)in, size );
break;
case CUBEHASH:
memcpy( &ctx.cube, &x16r_ctx.cube, sizeof(cubehashParam) );
cubehashUpdateDigest( &ctx.cube, (byte*) hash,
(const byte*)in, size );
break;
case SHAVITE:
sph_shavite512_init( &ctx.shavite );
sph_shavite512( &ctx.shavite, in, size );
sph_shavite512_close( &ctx.shavite, hash );
break;
case SIMD:
init_sd( &ctx.simd, 512 );
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence*)in, size<<3 );
break;
case ECHO:
#ifdef NO_AES_NI
sph_echo512_init( &ctx.echo );
sph_echo512( &ctx.echo, in, size );
sph_echo512_close( &ctx.echo, hash );
#else
init_echo( &ctx.echo, 512 );
update_final_echo ( &ctx.echo, (BitSequence *)hash,
(const BitSequence*)in, size<<3 );
#endif
break;
case HAMSI:
sph_hamsi512_init( &ctx.hamsi );
sph_hamsi512( &ctx.hamsi, in, size );
sph_hamsi512_close( &ctx.hamsi, hash );
break;
case FUGUE:
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, in, size );
sph_fugue512_close( &ctx.fugue, hash );
break;
case SHABAL:
sph_shabal512_init( &ctx.shabal );
sph_shabal512( &ctx.shabal, in, size );
sph_shabal512_close( &ctx.shabal, hash );
break;
case WHIRLPOOL:
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, in, size );
sph_whirlpool_close( &ctx.whirlpool, hash );
break;
case SHA_512:
SHA512_Init( &ctx.sha512 );
SHA512_Update( &ctx.sha512, in, size );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
break;
}
in = (void*) hash;
size = 64;
}
memcpy(output, hash, 32);
}
int scanhash_x16r( 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 Htarg = ptarget[7];
const uint32_t first_nonce = pdata[19];
uint32_t nonce = first_nonce;
volatile uint8_t *restart = &(work_restart[thr_id].restart);
for ( int k=0; k < 19; k++ )
be32enc( &endiandata[k], pdata[k] );
// This code is suspicious. s_ntime is saved after byteswapping pdata[17]
// but is tested vs unswapped pdata[17]. This should result in calling
// getAlgoString every pass, but that doesn't seem to be the case.
// It appears to be working correctly as is.
if ( s_ntime != pdata[17] )
{
uint32_t ntime = swab32(pdata[17]);
x16_r_s_getAlgoString( (const uint8_t*) (&endiandata[1]), hashOrder );
s_ntime = ntime;
if ( opt_debug && !thr_id )
applog( LOG_DEBUG, "hash order %s (%08x)", hashOrder, ntime );
}
if ( opt_benchmark )
ptarget[7] = 0x0cff;
do
{
be32enc( &endiandata[19], nonce );
x16r_hash( hash32, endiandata );
if ( hash32[7] <= Htarg && fulltest( hash32, ptarget ) )
{
work_set_target_ratio( work, hash32 );
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;
}

197
algo/x17/x17-4way.c
View File

@@ -15,6 +15,7 @@
#include "algo/luffa/luffa-hash-2way.h"
#include "algo/cubehash/cube-hash-2way.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/shavite/shavite-hash-2way.h"
#include "algo/simd/simd-hash-2way.h"
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/hamsi/hamsi-hash-4way.h"
@@ -24,7 +25,9 @@
#include "algo/haval/haval-hash-4way.h"
#include "algo/sha/sha2-hash-4way.h"
typedef struct {
//typedef struct {
union _x17_4way_context_overlay
{
blake512_4way_context blake;
bmw512_4way_context bmw;
hashState_groestl groestl;
@@ -33,7 +36,7 @@ typedef struct {
keccak512_4way_context keccak;
luffa_2way_context luffa;
cube_2way_context cube;
sph_shavite512_context shavite;
shavite512_2way_context shavite;
simd_2way_context simd;
hashState_echo echo;
hamsi512_4way_context hamsi;
@@ -42,8 +45,10 @@ typedef struct {
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
haval256_5_4way_context haval;
} x17_4way_ctx_holder;
};
typedef union _x17_4way_context_overlay x17_4way_context_overlay;
/*
x17_4way_ctx_holder x17_4way_ctx __attribute__ ((aligned (64)));
void init_x17_4way_ctx()
@@ -56,16 +61,17 @@ void init_x17_4way_ctx()
keccak512_4way_init( &x17_4way_ctx.keccak );
luffa_2way_init( &x17_4way_ctx.luffa, 512 );
cube_2way_init( &x17_4way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &x17_4way_ctx.shavite );
shavite512_2way_init( &x17_4way_ctx.shavite );
simd_2way_init( &x17_4way_ctx.simd, 512 );
init_echo( &x17_4way_ctx.echo, 512 );
hamsi512_4way_init( &x17_4way_ctx.hamsi );
sph_fugue512_init( &x17_4way_ctx.fugue );
shabal512_4way_init( &x17_4way_ctx.shabal );
sph_whirlpool_init( &x17_4way_ctx.whirlpool );
sha512_4way_init( &x17_4way_ctx.sha512 );
haval256_5_4way_init( &x17_4way_ctx.haval );
};
*/
void x17_4way_hash( void *state, const void *input )
{
uint64_t hash0[8] __attribute__ ((aligned (64)));
@@ -73,155 +79,159 @@ void x17_4way_hash( void *state, const void *input )
uint64_t hash2[8] __attribute__ ((aligned (64)));
uint64_t hash3[8] __attribute__ ((aligned (64)));
uint64_t vhash[8*4] __attribute__ ((aligned (64)));
uint64_t vhashA[8*4] __attribute__ ((aligned (64)));
uint64_t vhashB[8*4] __attribute__ ((aligned (64)));
x17_4way_ctx_holder ctx;
memcpy( &ctx, &x17_4way_ctx, sizeof(x17_4way_ctx) );
x17_4way_context_overlay ctx;
// memcpy( &ctx, &x17_4way_ctx, sizeof(x17_4way_ctx) );
// 1 Blake 4 way 64 bit
// 1 Blake parallel 4 way 64 bit
blake512_4way_init( &ctx.blake );
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
// 2 Bmw
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
// Serialize
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 3 Groestl
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
memcpy( &ctx.groestl, &x17_4way_ctx.groestl, sizeof(hashState_groestl) );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
memcpy( &ctx.groestl, &x17_4way_ctx.groestl, sizeof(hashState_groestl) );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash2, (char*)hash2, 512 );
memcpy( &ctx.groestl, &x17_4way_ctx.groestl, sizeof(hashState_groestl) );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// Parallel 4way
// Parallellize
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein
// 4 Skein parallel 4 way 64 bit
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
// 5 JH
jh512_4way_init( &ctx.jh );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
// 6 Keccak
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 );
// 7 Luffa parallel 2 way 128 bit
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
// 7 Luffa parallel 2 way
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
mm256_interleave_2x128( vhashB, hash2, hash3, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
// 8 Cubehash parallel 2 way
cube_2way_update_close( &ctx.cube, vhash, vhash, 64 );
cube_2way_reinit( &ctx.cube );
// 8 Cubehash
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_2way_init( &ctx.cube, 512, 16, 32 );
cube_2way_update_close( &ctx.cube, vhashB, vhashB, 64 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
// 9 Shavite
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
// 10 Simd
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
// 9 Shavite serial
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &x17_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash1, 64 );
sph_shavite512_close( &ctx.shavite, hash1 );
memcpy( &ctx.shavite, &x17_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash2, 64 );
sph_shavite512_close( &ctx.shavite, hash2 );
memcpy( &ctx.shavite, &x17_4way_ctx.shavite,
sizeof(sph_shavite512_context) );
sph_shavite512( &ctx.shavite, hash3, 64 );
sph_shavite512_close( &ctx.shavite, hash3 );
// 10 Simd parallel 2 way 128 bit
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, hash2, hash3, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
// 11 Echo serial
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &x17_4way_ctx.echo, sizeof(hashState_echo) );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash1,
(const BitSequence *) hash1, 512 );
memcpy( &ctx.echo, &x17_4way_ctx.echo, sizeof(hashState_echo) );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash2,
(const BitSequence *) hash2, 512 );
memcpy( &ctx.echo, &x17_4way_ctx.echo, sizeof(hashState_echo) );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
// 12 Hamsi parallel 4 way 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 13 Fugue serial
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
sph_fugue512_close( &ctx.fugue, hash0 );
memcpy( &ctx.fugue, &x17_4way_ctx.fugue, sizeof(sph_fugue512_context) );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash1, 64 );
sph_fugue512_close( &ctx.fugue, hash1 );
memcpy( &ctx.fugue, &x17_4way_ctx.fugue, sizeof(sph_fugue512_context) );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash2, 64 );
sph_fugue512_close( &ctx.fugue, hash2 );
memcpy( &ctx.fugue, &x17_4way_ctx.fugue, sizeof(sph_fugue512_context) );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
// 14 Shabal, parallel 4 way 32 bit SSE
// 14 Shabal, parallel 4 way 32 bit
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
// 15 Whirlpool
// 15 Whirlpool serial
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash0 );
memcpy( &ctx.whirlpool, &x17_4way_ctx.whirlpool,
sizeof(sph_whirlpool_context) );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash1, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash1 );
memcpy( &ctx.whirlpool, &x17_4way_ctx.whirlpool,
sizeof(sph_whirlpool_context) );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash2, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash2 );
memcpy( &ctx.whirlpool, &x17_4way_ctx.whirlpool,
sizeof(sph_whirlpool_context) );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
// 16 SHA512 parallel 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
// 17 Haval parallel 32 bit
mm256_reinterleave_4x32( vhashB, vhash, 512 );
haval256_5_4way_init( &ctx.haval );
haval256_5_4way( &ctx.haval, vhashB, 64 );
haval256_5_4way_close( &ctx.haval, state );
}
int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t *hash7 = &(hash[7<<2]);
@@ -234,49 +244,48 @@ int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
const uint32_t first_nonce = pdata[19];
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
__m256i *noncev = (__m256i*)vdata + 9; // aligned
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
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 );
// Need big endian data
casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
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
{
be32enc( noncep, n );
be32enc( noncep+2, n+1 );
be32enc( noncep+4, n+2 );
be32enc( noncep+6, n+3 );
for ( int m=0; m < 6; m++ ) if ( Htarg <= htmax[m] )
{
uint32_t mask = masks[m];
do
{
*noncev = mm256_interleave_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ),
*noncev );
x17_4way_hash( hash, vdata );
x17_4way_hash( hash, vdata );
for ( int lane = 0; lane < 4; lane++ )
if ( ( ( hash7[ lane ] & mask ) == 0 ) )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
{
pdata[19] = n + lane;
nonces[ num_found++ ] = n + lane;
work_set_target_ratio( work, lane_hash );
}
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce )
for ( int lane = 0; lane < 4; lane++ )
if ( ( ( hash7[ lane ] & mask ) == 0 ) )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
{
pdata[19] = n + lane;
nonces[ num_found++ ] = n + lane;
work_set_target_ratio( work, lane_hash );
}
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce )
&& !work_restart[thr_id].restart );
break;
}
break;
}
*hashes_done = n - first_nonce + 1;
return num_found;

4
algo/x17/x17-gate.c
View File

@@ -3,10 +3,12 @@
bool register_x17_algo( algo_gate_t* gate )
{
#if defined (X17_4WAY)
init_x17_4way_ctx();
printf("register x17 4way\n");
// init_x17_4way_ctx();
gate->scanhash = (void*)&scanhash_x17_4way;
gate->hash = (void*)&x17_4way_hash;
#else
printf("register x17 no 4way\n");
init_x17_ctx();
gate->scanhash = (void*)&scanhash_x17;
gate->hash = (void*)&x17_hash;

6
algo/x17/x17-gate.h
View File

@@ -15,16 +15,16 @@ bool register_x17_algo( algo_gate_t* gate );
void x17_4way_hash( void *state, const void *input );
int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
uint64_t *hashes_done, struct thr_info *mythr );
void init_x17_4way_ctx();
//void init_x17_4way_ctx();
#endif
void x17_hash( void *state, const void *input );
int scanhash_x17( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
uint64_t *hashes_done, struct thr_info *mythr );
void init_x17_ctx();

233
algo/x17/x17.c
View File

@@ -5,7 +5,6 @@
#include <stdio.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
@@ -13,14 +12,11 @@
#include "algo/luffa/sph_luffa.h"
#include "algo/cubehash/sph_cubehash.h"
#include "algo/simd/sph_simd.h"
#include "algo/echo/sph_echo.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/sha/sph_sha2.h"
#include "algo/haval/sph-haval.h"
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
@@ -30,18 +26,21 @@
#include "algo/skein/sse2/skein.c"
#include "algo/jh/sse2/jh_sse2_opt64.h"
#include <openssl/sha.h>
#ifndef NO_AES_NI
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#else
#include "algo/groestl/sph_groestl.h"
#include "algo/echo/sph_echo.h"
#endif
typedef struct {
#ifdef NO_AES_NI
sph_groestl512_context groestl;
sph_echo512_context echo;
#else
hashState_echo echo;
#if defined(__AES__)
hashState_groestl groestl;
hashState_echo echo;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
#endif
hashState_luffa luffa;
cubehashParam cubehash;
@@ -51,11 +50,7 @@ typedef struct {
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
#ifndef USE_SPH_SHA
SHA512_CTX sha512;
#else
sph_sha512_context sha512;
#endif
sph_haval256_5_context haval;
} x17_ctx_holder;
@@ -63,12 +58,12 @@ x17_ctx_holder x17_ctx __attribute__ ((aligned (64)));
void init_x17_ctx()
{
#ifdef NO_AES_NI
#if defined(__AES__)
init_groestl( &x17_ctx.groestl, 64 );
init_echo( &x17_ctx.echo, 512 );
#else
sph_groestl512_init(&x17_ctx.groestl );
sph_echo512_init(&x17_ctx.echo);
#else
init_echo( &x17_ctx.echo, 512 );
init_groestl( &x17_ctx.groestl, 64 );
#endif
init_luffa( &x17_ctx.luffa, 512 );
cubehashInit( &x17_ctx.cubehash, 512, 16, 32 );
@@ -78,11 +73,7 @@ void init_x17_ctx()
sph_fugue512_init( &x17_ctx.fugue );
sph_shabal512_init( &x17_ctx.shabal );
sph_whirlpool_init( &x17_ctx.whirlpool );
#ifndef USE_SPH_SHA
SHA512_Init( &x17_ctx.sha512 );
#else
sph_sha512_init(&x17_ctx.sha512);
#endif
sph_haval256_5_init(&x17_ctx.haval);
};
@@ -123,12 +114,12 @@ void x17_hash(void *output, const void *input)
//---groestl----
#ifdef NO_AES_NI
sph_groestl512(&ctx.groestl, hash, 64);
sph_groestl512_close(&ctx.groestl, hash);
#else
#if defined(__AES__)
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, 512 );
#else
sph_groestl512( &ctx.groestl, hash, 64 );
sph_groestl512_close( &ctx.groestl, hash );
#endif
//---skein4---
@@ -151,134 +142,136 @@ void x17_hash(void *output, const void *input)
KEC_C;
//--- luffa7
update_and_final_luffa( &ctx.luffa, (BitSequence*)hashB,
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
// 8 Cube
cubehashUpdateDigest( &ctx.cubehash, (byte*) hash,
(const byte*)hashB, 64 );
(const byte*)hash, 64 );
// 9 Shavite
sph_shavite512( &ctx.shavite, hash, 64);
sph_shavite512_close( &ctx.shavite, hashB);
sph_shavite512_close( &ctx.shavite, hash);
// 10 Simd
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hashB, 512 );
update_final_sd( &ctx.simd, (BitSequence*)hash,
(const BitSequence*)hash, 512 );
//11---echo---
#ifdef NO_AES_NI
sph_echo512(&ctx.echo, hash, 64);
sph_echo512_close(&ctx.echo, hashB);
#if defined(__AES__)
update_final_echo ( &ctx.echo, (BitSequence*)hash,
(const BitSequence*)hash, 512 );
#else
update_final_echo ( &ctx.echo, (BitSequence *)hashB,
(const BitSequence *)hash, 512 );
sph_echo512( &ctx.echo, hash, 64 );
sph_echo512_close( &ctx.echo, hash );
#endif
// X13 algos
// 12 Hamsi
sph_hamsi512(&ctx.hamsi, hashB, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
sph_hamsi512( &ctx.hamsi, hash, 64 );
sph_hamsi512_close( &ctx.hamsi, hash );
// 13 Fugue
sph_fugue512(&ctx.fugue, hash, 64);
sph_fugue512_close(&ctx.fugue, hashB);
sph_fugue512(&ctx.fugue, hash, 64 );
sph_fugue512_close(&ctx.fugue, hash );
// X14 Shabal
sph_shabal512(&ctx.shabal, hashB, 64);
sph_shabal512_close(&ctx.shabal, hash);
sph_shabal512(&ctx.shabal, hash, 64);
sph_shabal512_close( &ctx.shabal, hash );
// X15 Whirlpool
sph_whirlpool(&ctx.whirlpool, hash, 64);
sph_whirlpool_close(&ctx.whirlpool, hashB);
sph_whirlpool( &ctx.whirlpool, hash, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash );
#ifndef USE_SPH_SHA
SHA512_Update( &ctx.sha512, hashB, 64 );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
#else
sph_sha512(&ctx.sha512,(const void*) hashB, 64);
sph_sha512_close(&ctx.sha512,(void*) hash);
#endif
sph_haval256_5(&ctx.haval,(const void*) hash, 64);
sph_haval256_5_close(&ctx.haval,hashB);
SHA512_Update( &ctx.sha512, hash, 64 );
SHA512_Final( (unsigned char*)hash, &ctx.sha512 );
asm volatile ("emms");
memcpy(output, hashB, 32);
sph_haval256_5( &ctx.haval, (const void*)hash, 64 );
sph_haval256_5_close( &ctx.haval, output );
}
int scanhash_x17(int thr_id, struct work *work,
uint32_t max_nonce, uint64_t *hashes_done)
int scanhash_x17( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr)
{
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
const uint32_t Htarg = ptarget[7];
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
uint64_t htmax[] = {
0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000
};
uint32_t masks[] = {
0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0
};
uint64_t htmax[] =
{
0,
0xF,
0xFF,
0xFFF,
0xFFFF,
0x10000000
};
uint32_t masks[] =
{
0xFFFFFFFF,
0xFFFFFFF0,
0xFFFFFF00,
0xFFFFF000,
0xFFFF0000,
0
};
// we need bigendian data...
swab32_array( endiandata, pdata, 20 );
// we need bigendian data...
casti_m128i( endiandata, 0 ) = mm128_bswap_32( casti_m128i( pdata, 0 ) );
casti_m128i( endiandata, 1 ) = mm128_bswap_32( casti_m128i( pdata, 1 ) );
casti_m128i( endiandata, 2 ) = mm128_bswap_32( casti_m128i( pdata, 2 ) );
casti_m128i( endiandata, 3 ) = mm128_bswap_32( casti_m128i( pdata, 3 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
#ifdef DEBUG_ALGO
if (Htarg != 0)
printf("[%d] Htarg=%X\n", thr_id, Htarg);
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);
x17_hash(hash64, endiandata);
for ( int m = 0; m < 6; m++ )
{
if ( Htarg <= htmax[m] )
{
uint32_t mask = masks[m];
do
{
pdata[19] = ++n;
be32enc( &endiandata[19], n );
x17_hash( hash64, endiandata );
#ifndef DEBUG_ALGO
if (!(hash64[7] & mask))
{
if ( fulltest(hash64, ptarget)) {
*hashes_done = n - first_nonce + 1;
return true;
}
// else
// {
// applog(LOG_INFO, "Result does not validate on CPU!");
// }
}
if ( !( hash64[7] & mask ) )
{
if ( fulltest( hash64, ptarget ) )
{
*hashes_done = n - first_nonce + 1;
return true;
}
// else
// applog(LOG_INFO, "Result does not validate on CPU!");
}
#else
if (!(n % 0x1000) && !thr_id) printf(".");
if (!(hash64[7] & mask)) {
printf("[%d]",thr_id);
if (fulltest(hash64, ptarget)) {
work_set_target_ratio( work, hash64 );
*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;
if ( !( n % 0x1000 ) && !thr_id ) printf(".");
if ( !( hash64[7] & mask ) )
{
printf("[%d]",thr_id);
if ( fulltest( hash64, ptarget ) )
{
work_set_target_ratio( work, hash64 );
*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;
}
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;
}

77
algo/x17/xevan.c
View File

@@ -16,17 +16,16 @@
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/sha/sph_sha2.h"
#include "algo/haval/sph-haval.h"
#include "algo/simd/nist.h"
#include "algo/cubehash/cubehash_sse2.h"
#include <openssl/sha.h>
#ifdef NO_AES_NI
#include "algo/groestl/sph_groestl.h"
#include "algo/echo/sph_echo.h"
#else
#if defined(__AES__)
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/echo/aes_ni/hash_api.h"
#else
#include "algo/groestl/sph_groestl.h"
#include "algo/echo/sph_echo.h"
#endif
typedef struct {
@@ -43,18 +42,14 @@ typedef struct {
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
#ifndef USE_SPH_SHA
SHA512_CTX sha512;
#else
sph_sha512_context sha512;
#endif
sph_haval256_5_context haval;
#ifdef NO_AES_NI
sph_groestl512_context groestl;
sph_echo512_context echo;
#else
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
#endif
} xevan_ctx_holder;
@@ -77,18 +72,14 @@ void init_xevan_ctx()
sph_fugue512_init( &xevan_ctx.fugue );
sph_shabal512_init( &xevan_ctx.shabal );
sph_whirlpool_init( &xevan_ctx.whirlpool );
#ifndef USE_SPH_SHA
SHA512_Init( &xevan_ctx.sha512 );
#else
sph_sha512_init(&xevan_ctx.sha512);
#endif
sph_haval256_5_init(&xevan_ctx.haval);
#ifdef NO_AES_NI
sph_groestl512_init( &xevan_ctx.groestl );
sph_echo512_init( &xevan_ctx.echo );
#else
#if defined(__AES__)
init_groestl( &xevan_ctx.groestl, 64 );
init_echo( &xevan_ctx.echo, 512 );
#else
sph_groestl512_init( &xevan_ctx.groestl );
sph_echo512_init( &xevan_ctx.echo );
#endif
};
@@ -117,12 +108,12 @@ void xevan_hash(void *output, const void *input)
sph_bmw512(&ctx.bmw, hash, dataLen);
sph_bmw512_close(&ctx.bmw, hash);
#ifdef NO_AES_NI
#if defined(__AES__)
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, dataLen*8 );
#else
sph_groestl512(&ctx.groestl, hash, dataLen);
sph_groestl512_close(&ctx.groestl, hash);
#else
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const char*)hash, dataLen*8 );
#endif
sph_skein512(&ctx.skein, hash, dataLen);
@@ -146,12 +137,12 @@ void xevan_hash(void *output, const void *input)
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, dataLen*8 );
#ifdef NO_AES_NI
#if defined(__AES__)
update_final_echo( &ctx.echo, (BitSequence *) hash,
(const BitSequence *) hash, dataLen*8 );
#else
sph_echo512(&ctx.echo, hash, dataLen);
sph_echo512_close(&ctx.echo, hash);
#else
update_final_echo( &ctx.echo, (BitSequence *) hash,
(const BitSequence *) hash, dataLen*8 );
#endif
sph_hamsi512(&ctx.hamsi, hash, dataLen);
@@ -166,13 +157,9 @@ void xevan_hash(void *output, const void *input)
sph_whirlpool(&ctx.whirlpool, hash, dataLen);
sph_whirlpool_close(&ctx.whirlpool, hash);
#ifndef USE_SPH_SHA
SHA512_Update( &ctx.sha512, hash, dataLen );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
#else
sph_sha512(&ctx.sha512,(const void*) hash, dataLen);
sph_sha512_close(&ctx.sha512,(void*) hash);
#endif
sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
sph_haval256_5_close(&ctx.haval, hash);
@@ -186,12 +173,12 @@ void xevan_hash(void *output, const void *input)
sph_bmw512(&ctx.bmw, hash, dataLen);
sph_bmw512_close(&ctx.bmw, hash);
#ifdef NO_AES_NI
sph_groestl512(&ctx.groestl, hash, dataLen);
sph_groestl512_close(&ctx.groestl, hash);
#else
#if defined(__AES__)
update_and_final_groestl( &ctx.groestl, (char*)hash,
(const BitSequence*)hash, dataLen*8 );
#else
sph_groestl512(&ctx.groestl, hash, dataLen);
sph_groestl512_close(&ctx.groestl, hash);
#endif
sph_skein512(&ctx.skein, hash, dataLen);
@@ -214,12 +201,12 @@ void xevan_hash(void *output, const void *input)
update_final_sd( &ctx.simd, (BitSequence *)hash,
(const BitSequence *)hash, dataLen*8 );
#ifdef NO_AES_NI
sph_echo512(&ctx.echo, hash, dataLen);
sph_echo512_close(&ctx.echo, hash);
#else
#if defined(__AES__)
update_final_echo( &ctx.echo, (BitSequence *) hash,
(const BitSequence *) hash, dataLen*8 );
#else
sph_echo512(&ctx.echo, hash, dataLen);
sph_echo512_close(&ctx.echo, hash);
#endif
sph_hamsi512(&ctx.hamsi, hash, dataLen);
@@ -234,13 +221,9 @@ void xevan_hash(void *output, const void *input)
sph_whirlpool(&ctx.whirlpool, hash, dataLen);
sph_whirlpool_close(&ctx.whirlpool, hash);
#ifndef USE_SPH_SHA
SHA512_Update( &ctx.sha512, hash, dataLen );
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
#else
sph_sha512(&ctx.sha512,(const void*) hash, dataLen);
sph_sha512_close(&ctx.sha512,(void*) hash);
#endif
sph_haval256_5(&ctx.haval,(const void*) hash, dataLen);
sph_haval256_5_close(&ctx.haval, hash);