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

v3.11.0

Browse Source
This commit is contained in:
Jay D Dee
2020-01-02 23:54:08 -05:00
parent 241bc26767
commit 3572cb53c4
118 changed files with 7030 additions and 1575 deletions
Download Patch File Download Diff File Expand all files Collapse all files

112
algo/x11/c11-4way.c
View File

@@ -14,21 +14,32 @@
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/simd-hash-2way.h"
#include "algo/echo/aes_ni/hash_api.h"
#if defined(__VAES__)
#include "algo/groestl/groestl512-hash-4way.h"
#include "algo/shavite/shavite-hash-4way.h"
#include "algo/echo/echo-hash-4way.h"
#endif
#if defined (C11_8WAY)
typedef struct {
blake512_8way_context blake;
bmw512_8way_context bmw;
hashState_groestl groestl;
skein512_8way_context skein;
jh512_8way_context jh;
keccak512_8way_context keccak;
luffa_4way_context luffa;
cube_4way_context cube;
sph_shavite512_context shavite;
simd_4way_context simd;
#if defined(__VAES__)
groestl512_4way_context groestl;
shavite512_4way_context shavite;
echo_4way_context echo;
#else
hashState_groestl groestl;
sph_shavite512_context shavite;
hashState_echo echo;
#endif
} c11_8way_ctx_holder;
c11_8way_ctx_holder c11_8way_ctx;
@@ -37,22 +48,28 @@ void init_c11_8way_ctx()
{
blake512_8way_init( &c11_8way_ctx.blake );
bmw512_8way_init( &c11_8way_ctx.bmw );
init_groestl( &c11_8way_ctx.groestl, 64 );
skein512_8way_init( &c11_8way_ctx.skein );
jh512_8way_init( &c11_8way_ctx.jh );
keccak512_8way_init( &c11_8way_ctx.keccak );
luffa_4way_init( &c11_8way_ctx.luffa, 512 );
cube_4way_init( &c11_8way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &c11_8way_ctx.shavite );
simd_4way_init( &c11_8way_ctx.simd, 512 );
#if defined(__VAES__)
groestl512_4way_init( &c11_8way_ctx.groestl, 64 );
shavite512_4way_init( &c11_8way_ctx.shavite );
echo_4way_init( &c11_8way_ctx.echo, 512 );
#else
init_groestl( &c11_8way_ctx.groestl, 64 );
sph_shavite512_init( &c11_8way_ctx.shavite );
init_echo( &c11_8way_ctx.echo, 512 );
#endif
}
void c11_8way_hash( void *state, const void *input )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
uint64_t vhash0[4*8] __attribute__ ((aligned (64)));
uint64_t vhash1[4*8] __attribute__ ((aligned (64)));
uint64_t vhashA[4*8] __attribute__ ((aligned (64)));
uint64_t vhashB[4*8] __attribute__ ((aligned (64)));
uint64_t hash0[8] __attribute__ ((aligned (64)));
uint64_t hash1[8] __attribute__ ((aligned (64)));
uint64_t hash2[8] __attribute__ ((aligned (64)));
@@ -72,11 +89,21 @@ void c11_8way_hash( void *state, const void *input )
bmw512_8way_update( &ctx.bmw, vhash, 64 );
bmw512_8way_close( &ctx.bmw, vhash );
// Serial
#if defined(__VAES__)
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
groestl512_4way_update_close( &ctx.groestl, vhashA, vhashA, 512 );
groestl512_4way_init( &ctx.groestl, 64 );
groestl512_4way_update_close( &ctx.groestl, vhashB, vhashB, 512 );
rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_8x64_512( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
vhash );
// 3 Groestl
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
memcpy( &ctx.groestl, &c11_8way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash1, (char*)hash1, 512 );
@@ -93,10 +120,11 @@ void c11_8way_hash( void *state, const void *input )
memcpy( &ctx.groestl, &c11_8way_ctx.groestl, sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );
// 4way
intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7 );
#endif
// 4 JH
jh512_8way_update( &ctx.jh, vhash, 64 );
jh512_8way_close( &ctx.jh, vhash );
@@ -109,20 +137,27 @@ void c11_8way_hash( void *state, const void *input )
skein512_8way_update( &ctx.skein, vhash, 64 );
skein512_8way_close( &ctx.skein, vhash );
rintrlv_8x64_4x128( vhash0, vhash1, vhash, 512 );
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
luffa_4way_update_close( &ctx.luffa, vhash0, vhash0, 64 );
luffa_4way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_4way_init( &ctx.luffa, 512 );
luffa_4way_update_close( &ctx.luffa, vhash1, vhash1, 64 );
luffa_4way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_4way_update_close( &ctx.cube, vhash0, vhash0, 64 );
cube_4way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_4way_init( &ctx.cube, 512, 16, 32 );
cube_4way_update_close( &ctx.cube, vhash1, vhash1, 64 );
cube_4way_update_close( &ctx.cube, vhashB, vhashB, 64 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash0 );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash1 );
#if defined(__VAES__)
shavite512_4way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_4way_init( &ctx.shavite );
shavite512_4way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
// 9 Shavite
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &c11_8way_ctx.shavite,
@@ -154,16 +189,29 @@ void c11_8way_hash( void *state, const void *input )
sph_shavite512( &ctx.shavite, hash7, 64 );
sph_shavite512_close( &ctx.shavite, hash7 );
// 10 Simd
intrlv_4x128( vhash, hash0, hash1, hash2, hash3, 512 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128( hash0, hash1, hash2, hash3, vhash, 512 );
intrlv_4x128( vhash, hash4, hash5, hash6, hash7, 512 );
simd_4way_init( &ctx.simd, 512 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128( hash4, hash5, hash6, hash7, vhash, 512 );
intrlv_4x128_512( vhashA, hash0, hash1, hash2, hash3 );
intrlv_4x128_512( vhashB, hash4, hash5, hash6, hash7 );
// 11 Echo
#endif
simd_4way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_4way_init( &ctx.simd, 512 );
simd_4way_update_close( &ctx.simd, vhashB, vhashB, 512 );
#if defined(__VAES__)
echo_4way_update_close( &ctx.echo, vhashA, vhashA, 512 );
echo_4way_init( &ctx.echo, 512 );
echo_4way_update_close( &ctx.echo, vhashB, vhashB, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &c11_8way_ctx.echo, sizeof(hashState_echo) );
@@ -188,6 +236,8 @@ void c11_8way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash7,
(const BitSequence *) hash7, 512 );
#endif
memcpy( state, hash0, 32 );
memcpy( state+ 32, hash1, 32 );
memcpy( state+ 64, hash2, 32 );
@@ -282,11 +332,11 @@ void c11_4way_hash( void *state, const void *input )
memcpy( &ctx, &c11_4way_ctx, sizeof(c11_4way_ctx) );
// 1 Blake 4way
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_update( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
// 2 Bmw
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
@@ -305,15 +355,15 @@ void c11_4way_hash( void *state, const void *input )
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 JH
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_update( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
// 5 Keccak
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// 6 Skein
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_update( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
// Serial

2
algo/x11/c11-gate.c
View File

@@ -15,7 +15,7 @@ bool register_c11_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_c11;
gate->hash = (void*)&c11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
return true;
};

10
algo/x11/timetravel-4way.c
View File

@@ -84,13 +84,13 @@ void timetravel_4way_hash(void *output, const void *input)
switch ( permutation[i] )
{
case 0:
blake512_4way( &ctx.blake, vhashA, dataLen );
blake512_4way_update( &ctx.blake, vhashA, dataLen );
blake512_4way_close( &ctx.blake, vhashB );
if ( i == 7 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 1:
bmw512_4way( &ctx.bmw, vhashA, dataLen );
bmw512_4way_update( &ctx.bmw, vhashA, dataLen );
bmw512_4way_close( &ctx.bmw, vhashB );
if ( i == 7 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
@@ -112,19 +112,19 @@ void timetravel_4way_hash(void *output, const void *input)
intrlv_4x64( vhashB, hash0, hash1, hash2, hash3, dataLen<<3 );
break;
case 3:
skein512_4way( &ctx.skein, vhashA, dataLen );
skein512_4way_update( &ctx.skein, vhashA, dataLen );
skein512_4way_close( &ctx.skein, vhashB );
if ( i == 7 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 4:
jh512_4way( &ctx.jh, vhashA, dataLen );
jh512_4way_update( &ctx.jh, vhashA, dataLen );
jh512_4way_close( &ctx.jh, vhashB );
if ( i == 7 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 5:
keccak512_4way( &ctx.keccak, vhashA, dataLen );
keccak512_4way_update( &ctx.keccak, vhashA, dataLen );
keccak512_4way_close( &ctx.keccak, vhashB );
if ( i == 7 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );

10
algo/x11/timetravel10-4way.c
View File

@@ -90,13 +90,13 @@ void timetravel10_4way_hash(void *output, const void *input)
switch ( permutation[i] )
{
case 0:
blake512_4way( &ctx.blake, vhashA, dataLen );
blake512_4way_update( &ctx.blake, vhashA, dataLen );
blake512_4way_close( &ctx.blake, vhashB );
if ( i == 9 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 1:
bmw512_4way( &ctx.bmw, vhashA, dataLen );
bmw512_4way_update( &ctx.bmw, vhashA, dataLen );
bmw512_4way_close( &ctx.bmw, vhashB );
if ( i == 9 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
@@ -118,19 +118,19 @@ void timetravel10_4way_hash(void *output, const void *input)
intrlv_4x64( vhashB, hash0, hash1, hash2, hash3, dataLen<<3 );
break;
case 3:
skein512_4way( &ctx.skein, vhashA, dataLen );
skein512_4way_update( &ctx.skein, vhashA, dataLen );
skein512_4way_close( &ctx.skein, vhashB );
if ( i == 9 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 4:
jh512_4way( &ctx.jh, vhashA, dataLen );
jh512_4way_update( &ctx.jh, vhashA, dataLen );
jh512_4way_close( &ctx.jh, vhashB );
if ( i == 9 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );
break;
case 5:
keccak512_4way( &ctx.keccak, vhashA, dataLen );
keccak512_4way_update( &ctx.keccak, vhashA, dataLen );
keccak512_4way_close( &ctx.keccak, vhashB );
if ( i == 9 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhashB, dataLen<<3 );

26
algo/x11/tribus-4way.c
View File

@@ -6,6 +6,9 @@
#include "algo/jh/jh-hash-4way.h"
#include "algo/keccak/keccak-hash-4way.h"
#include "algo/echo/aes_ni/hash_api.h"
#if defined(__VAES__)
#include "algo/echo/echo-hash-4way.h"
#endif
#if defined(TRIBUS_8WAY)
@@ -14,6 +17,8 @@ static __thread jh512_8way_context ctx_mid;
void tribus_hash_8way( void *state, const void *input )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
uint64_t vhashA[4*8] __attribute__ ((aligned (64)));
uint64_t vhashB[4*8] __attribute__ ((aligned (64)));
uint64_t hash0[8] __attribute__ ((aligned (64)));
uint64_t hash1[8] __attribute__ ((aligned (64)));
uint64_t hash2[8] __attribute__ ((aligned (64)));
@@ -24,7 +29,11 @@ void tribus_hash_8way( void *state, const void *input )
uint64_t hash7[8] __attribute__ ((aligned (64)));
jh512_8way_context ctx_jh;
keccak512_8way_context ctx_keccak;
#if defined(__VAES__)
echo_4way_context ctx_echo;
#else
hashState_echo ctx_echo;
#endif
memcpy( &ctx_jh, &ctx_mid, sizeof(ctx_mid) );
jh512_8way_update( &ctx_jh, input + (64<<3), 16 );
@@ -34,10 +43,23 @@ void tribus_hash_8way( void *state, const void *input )
keccak512_8way_update( &ctx_keccak, vhash, 64 );
keccak512_8way_close( &ctx_keccak, vhash );
#if defined(__VAES__)
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
echo_4way_init( &ctx_echo, 512 );
echo_4way_update_close( &ctx_echo, vhashA, vhashA, 512 );
echo_4way_init( &ctx_echo, 512 );
echo_4way_update_close( &ctx_echo, vhashB, vhashB, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
#else
dintrlv_8x64( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
vhash, 512 );
// hash echo serially
init_echo( &ctx_echo, 512 );
update_final_echo( &ctx_echo, (BitSequence *) hash0,
(const BitSequence *) hash0, 512 );
@@ -63,6 +85,8 @@ void tribus_hash_8way( void *state, const void *input )
update_final_echo( &ctx_echo, (BitSequence *) hash7,
(const BitSequence *) hash7, 512 );
#endif
memcpy( state, hash0, 32 );
memcpy( state+32, hash1, 32 );
memcpy( state+64, hash2, 32 );

2
algo/x11/tribus-gate.c
View File

@@ -2,7 +2,7 @@
bool register_tribus_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
#if defined (TRIBUS_8WAY)
gate->scanhash = (void*)&scanhash_tribus_8way;
gate->hash = (void*)&tribus_hash_8way;

107
algo/x11/x11-4way.c
View File

@@ -14,21 +14,32 @@
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/simd-hash-2way.h"
#include "algo/echo/aes_ni/hash_api.h"
#if defined(__VAES__)
#include "algo/groestl/groestl512-hash-4way.h"
#include "algo/shavite/shavite-hash-4way.h"
#include "algo/echo/echo-hash-4way.h"
#endif
#if defined (X11_8WAY)
typedef struct {
blake512_8way_context blake;
bmw512_8way_context bmw;
hashState_groestl groestl;
skein512_8way_context skein;
jh512_8way_context jh;
keccak512_8way_context keccak;
luffa_4way_context luffa;
cube_4way_context cube;
sph_shavite512_context shavite;
simd_4way_context simd;
#if defined(__VAES__)
groestl512_4way_context groestl;
shavite512_4way_context shavite;
echo_4way_context echo;
#else
hashState_groestl groestl;
sph_shavite512_context shavite;
hashState_echo echo;
#endif
} x11_8way_ctx_holder;
x11_8way_ctx_holder x11_8way_ctx;
@@ -37,22 +48,28 @@ void init_x11_8way_ctx()
{
blake512_8way_init( &x11_8way_ctx.blake );
bmw512_8way_init( &x11_8way_ctx.bmw );
init_groestl( &x11_8way_ctx.groestl, 64 );
skein512_8way_init( &x11_8way_ctx.skein );
jh512_8way_init( &x11_8way_ctx.jh );
keccak512_8way_init( &x11_8way_ctx.keccak );
luffa_4way_init( &x11_8way_ctx.luffa, 512 );
cube_4way_init( &x11_8way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &x11_8way_ctx.shavite );
simd_4way_init( &x11_8way_ctx.simd, 512 );
#if defined(__VAES__)
groestl512_4way_init( &x11_8way_ctx.groestl, 64 );
shavite512_4way_init( &x11_8way_ctx.shavite );
echo_4way_init( &x11_8way_ctx.echo, 512 );
#else
init_groestl( &x11_8way_ctx.groestl, 64 );
sph_shavite512_init( &x11_8way_ctx.shavite );
init_echo( &x11_8way_ctx.echo, 512 );
#endif
}
void x11_8way_hash( void *state, const void *input )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
uint64_t vhash0[4*8] __attribute__ ((aligned (64)));
uint64_t vhash1[4*8] __attribute__ ((aligned (64)));
uint64_t vhashA[4*8] __attribute__ ((aligned (64)));
uint64_t vhashB[4*8] __attribute__ ((aligned (64)));
uint64_t hash0[8] __attribute__ ((aligned (64)));
uint64_t hash1[8] __attribute__ ((aligned (64)));
uint64_t hash2[8] __attribute__ ((aligned (64)));
@@ -61,7 +78,6 @@ void x11_8way_hash( void *state, const void *input )
uint64_t hash5[8] __attribute__ ((aligned (64)));
uint64_t hash6[8] __attribute__ ((aligned (64)));
uint64_t hash7[8] __attribute__ ((aligned (64)));
x11_8way_ctx_holder ctx;
memcpy( &ctx, &x11_8way_ctx, sizeof(x11_8way_ctx) );
blake512_8way_update( &ctx.blake, input, 80 );
@@ -70,7 +86,18 @@ void x11_8way_hash( void *state, const void *input )
bmw512_8way_update( &ctx.bmw, vhash, 64 );
bmw512_8way_close( &ctx.bmw, vhash );
// Serial
#if defined(__VAES__)
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
groestl512_4way_update_close( &ctx.groestl, vhashA, vhashA, 512 );
groestl512_4way_init( &ctx.groestl, 64 );
groestl512_4way_update_close( &ctx.groestl, vhashB, vhashB, 512 );
rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_8x64_512( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
vhash );
@@ -97,10 +124,11 @@ void x11_8way_hash( void *state, const void *input )
sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );
// 4way
intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7 );
#endif
skein512_8way_update( &ctx.skein, vhash, 64 );
skein512_8way_close( &ctx.skein, vhash );
@@ -110,18 +138,26 @@ void x11_8way_hash( void *state, const void *input )
keccak512_8way_update( &ctx.keccak, vhash, 64 );
keccak512_8way_close( &ctx.keccak, vhash );
rintrlv_8x64_4x128( vhash0, vhash1, vhash, 512 );
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
luffa_4way_update_close( &ctx.luffa, vhash0, vhash0, 64 );
luffa_4way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_4way_init( &ctx.luffa, 512 );
luffa_4way_update_close( &ctx.luffa, vhash1, vhash1, 64 );
luffa_4way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
cube_4way_update_close( &ctx.cube, vhash0, vhash0, 64 );
cube_4way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_4way_init( &ctx.cube, 512, 16, 32 );
cube_4way_update_close( &ctx.cube, vhash1, vhash1, 64 );
cube_4way_update_close( &ctx.cube, vhashB, vhashB, 64 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash0 );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash1 );
#if defined(__VAES__)
shavite512_4way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_4way_init( &ctx.shavite );
shavite512_4way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
sph_shavite512( &ctx.shavite, hash0, 64 );
sph_shavite512_close( &ctx.shavite, hash0 );
@@ -154,13 +190,28 @@ void x11_8way_hash( void *state, const void *input )
sph_shavite512( &ctx.shavite, hash7, 64 );
sph_shavite512_close( &ctx.shavite, hash7 );
intrlv_4x128_512( vhash, hash0, hash1, hash2, hash3 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash );
intrlv_4x128_512( vhash, hash4, hash5, hash6, hash7 );
intrlv_4x128_512( vhashA, hash0, hash1, hash2, hash3 );
intrlv_4x128_512( vhashB, hash4, hash5, hash6, hash7 );
#endif
simd_4way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_4way_init( &ctx.simd, 512 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash );
simd_4way_update_close( &ctx.simd, vhashB, vhashB, 512 );
#if defined(__VAES__)
echo_4way_update_close( &ctx.echo, vhashA, vhashA, 512 );
echo_4way_init( &ctx.echo, 512 );
echo_4way_update_close( &ctx.echo, vhashB, vhashB, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
@@ -186,6 +237,8 @@ void x11_8way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash7,
(const BitSequence *) hash7, 512 );
#endif
memcpy( state, hash0, 32 );
memcpy( state+ 32, hash1, 32 );
memcpy( state+ 64, hash2, 32 );
@@ -282,11 +335,11 @@ void x11_4way_hash( void *state, const void *input )
memcpy( &ctx, &x11_4way_ctx, sizeof(x11_4way_ctx) );
// 1 Blake 4way
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_update( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
// 2 Bmw
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
@@ -305,15 +358,15 @@ void x11_4way_hash( void *state, const void *input )
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_update( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
// 5 JH
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_update( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
// 6 Keccak
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );

2
algo/x11/x11-gate.c
View File

@@ -15,7 +15,7 @@ bool register_x11_algo( algo_gate_t *gate )
gate->scanhash = (void*)&scanhash_x11;
gate->hash = (void*)&x11_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT ;
return true;
};

10
algo/x11/x11evo-4way.c
View File

@@ -85,12 +85,12 @@ void x11evo_4way_hash( void *state, const void *input )
switch ( idx )
{
case 0:
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_update( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 64<<3 );
break;
case 1:
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
if ( i >= len-1 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 64<<3 );
@@ -112,19 +112,19 @@ void x11evo_4way_hash( void *state, const void *input )
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 64<<3 );
break;
case 3:
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_update( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
if ( i >= len-1 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 64<<3 );
break;
case 4:
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_update( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
if ( i >= len-1 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 64<<3 );
break;
case 5:
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
if ( i >= len-1 )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 64<<3 );

114
algo/x11/x11gost-4way.c
View File

@@ -15,22 +15,33 @@
#include "algo/shavite/sph_shavite.h"
#include "algo/simd/simd-hash-2way.h"
#include "algo/echo/aes_ni/hash_api.h"
#if defined(__VAES__)
#include "algo/groestl/groestl512-hash-4way.h"
#include "algo/shavite/shavite-hash-4way.h"
#include "algo/echo/echo-hash-4way.h"
#endif
#if defined (X11GOST_8WAY)
typedef struct {
blake512_8way_context blake;
bmw512_8way_context bmw;
hashState_groestl groestl;
skein512_8way_context skein;
jh512_8way_context jh;
keccak512_8way_context keccak;
sph_gost512_context gost;
luffa_4way_context luffa;
cube_4way_context cube;
sph_shavite512_context shavite;
simd_4way_context simd;
#if defined(__VAES__)
groestl512_4way_context groestl;
shavite512_4way_context shavite;
echo_4way_context echo;
#else
hashState_groestl groestl;
sph_shavite512_context shavite;
hashState_echo echo;
#endif
} x11gost_8way_ctx_holder;
x11gost_8way_ctx_holder x11gost_8way_ctx;
@@ -39,21 +50,29 @@ void init_x11gost_8way_ctx()
{
blake512_8way_init( &x11gost_8way_ctx.blake );
bmw512_8way_init( &x11gost_8way_ctx.bmw );
init_groestl( &x11gost_8way_ctx.groestl, 64 );
skein512_8way_init( &x11gost_8way_ctx.skein );
jh512_8way_init( &x11gost_8way_ctx.jh );
keccak512_8way_init( &x11gost_8way_ctx.keccak );
sph_gost512_init( &x11gost_8way_ctx.gost );
luffa_4way_init( &x11gost_8way_ctx.luffa, 512 );
cube_4way_init( &x11gost_8way_ctx.cube, 512, 16, 32 );
sph_shavite512_init( &x11gost_8way_ctx.shavite );
simd_4way_init( &x11gost_8way_ctx.simd, 512 );
#if defined(__VAES__)
groestl512_4way_init( &x11gost_8way_ctx.groestl, 64 );
shavite512_4way_init( &x11gost_8way_ctx.shavite );
echo_4way_init( &x11gost_8way_ctx.echo, 512 );
#else
init_groestl( &x11gost_8way_ctx.groestl, 64 );
sph_shavite512_init( &x11gost_8way_ctx.shavite );
init_echo( &x11gost_8way_ctx.echo, 512 );
#endif
}
void x11gost_8way_hash( void *state, const void *input )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
uint64_t vhashA[4*8] __attribute__ ((aligned (64)));
uint64_t vhashB[4*8] __attribute__ ((aligned (64)));
uint64_t hash0[8] __attribute__ ((aligned (64)));
uint64_t hash1[8] __attribute__ ((aligned (64)));
uint64_t hash2[8] __attribute__ ((aligned (64)));
@@ -72,7 +91,18 @@ void x11gost_8way_hash( void *state, const void *input )
bmw512_8way_update( &ctx.bmw, vhash, 64 );
bmw512_8way_close( &ctx.bmw, vhash );
// Serial
#if defined(__VAES__)
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
groestl512_4way_update_close( &ctx.groestl, vhashA, vhashA, 512 );
groestl512_4way_init( &ctx.groestl, 64 );
groestl512_4way_update_close( &ctx.groestl, vhashB, vhashB, 512 );
rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 );
#else
dintrlv_8x64_512( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7,
vhash );
@@ -99,10 +129,11 @@ void x11gost_8way_hash( void *state, const void *input )
sizeof(hashState_groestl) );
update_and_final_groestl( &ctx.groestl, (char*)hash7, (char*)hash7, 512 );
// 4way
intrlv_8x64_512( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7 );
#endif
skein512_8way_update( &ctx.skein, vhash, 64 );
skein512_8way_close( &ctx.skein, vhash );
@@ -140,20 +171,28 @@ void x11gost_8way_hash( void *state, const void *input )
sph_gost512( &ctx.gost, hash7, 64 );
sph_gost512_close( &ctx.gost, hash7 );
intrlv_4x128_512( vhashA, hash0, hash1, hash2, hash3 );
intrlv_4x128_512( vhashB, hash4, hash5, hash6, hash7 );
// Luffa + Cube
intrlv_4x128_512( vhash, hash0, hash1, hash2, hash3 );
luffa_4way_update_close( &ctx.luffa, vhash, vhash, 64 );
cube_4way_update_close( &ctx.cube, vhash, vhash, 64 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash );
intrlv_4x128_512( vhash, hash4, hash5, hash6, hash7 );
luffa_4way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
luffa_4way_init( &ctx.luffa, 512 );
cube_4way_init( &ctx.cube, 512, 16, 32 );
luffa_4way_update_close( &ctx.luffa, vhash, vhash, 64 );
cube_4way_update_close( &ctx.cube, vhash, vhash, 64 );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash );
luffa_4way_update_close( &ctx.luffa, vhashB, vhashB, 64 );
sph_shavite512( &ctx.shavite, hash0, 64 );
cube_4way_update_close( &ctx.cube, vhashA, vhashA, 64 );
cube_4way_init( &ctx.cube, 512, 16, 32 );
cube_4way_update_close( &ctx.cube, vhashB, vhashB, 64 );
#if defined(__VAES__)
shavite512_4way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
shavite512_4way_init( &ctx.shavite );
shavite512_4way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
sph_shavite512_close( &ctx.shavite, hash0 );
memcpy( &ctx.shavite, &x11gost_8way_ctx.shavite,
sizeof(sph_shavite512_context) );
@@ -184,14 +223,29 @@ void x11gost_8way_hash( void *state, const void *input )
sph_shavite512( &ctx.shavite, hash7, 64 );
sph_shavite512_close( &ctx.shavite, hash7 );
intrlv_4x128_512( vhash, hash0, hash1, hash2, hash3 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhash );
intrlv_4x128_512( vhash, hash4, hash5, hash6, hash7 );
simd_4way_init( &ctx.simd, 512 );
simd_4way_update_close( &ctx.simd, vhash, vhash, 512 );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhash );
intrlv_4x128_512( vhashA, hash0, hash1, hash2, hash3 );
intrlv_4x128_512( vhashB, hash4, hash5, hash6, hash7 );
#endif
simd_4way_update_close( &ctx.simd, vhashA, vhashA, 512 );
simd_4way_init( &ctx.simd, 512 );
simd_4way_update_close( &ctx.simd, vhashB, vhashB, 512 );
#if defined(__VAES__)
echo_4way_update_close( &ctx.echo, vhashA, vhashA, 512 );
echo_4way_init( &ctx.echo, 512 );
echo_4way_update_close( &ctx.echo, vhashB, vhashB, 512 );
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
#else
dintrlv_4x128_512( hash0, hash1, hash2, hash3, vhashA );
dintrlv_4x128_512( hash4, hash5, hash6, hash7, vhashB );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
(const BitSequence *) hash0, 512 );
memcpy( &ctx.echo, &x11gost_8way_ctx.echo, sizeof(hashState_echo) );
@@ -216,6 +270,8 @@ void x11gost_8way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash7,
(const BitSequence *) hash7, 512 );
#endif
memcpy( state, hash0, 32 );
memcpy( state+ 32, hash1, 32 );
memcpy( state+ 64, hash2, 32 );
@@ -310,10 +366,10 @@ void x11gost_4way_hash( void *state, const void *input )
x11gost_4way_ctx_holder ctx;
memcpy( &ctx, &x11gost_4way_ctx, sizeof(x11gost_4way_ctx) );
blake512_4way( &ctx.blake, input, 80 );
blake512_4way_update( &ctx.blake, input, 80 );
blake512_4way_close( &ctx.blake, vhash );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_update( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
@@ -333,13 +389,13 @@ void x11gost_4way_hash( void *state, const void *input )
// 4way
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way( &ctx.skein, vhash, 64 );
skein512_4way_update( &ctx.skein, vhash, 64 );
skein512_4way_close( &ctx.skein, vhash );
jh512_4way( &ctx.jh, vhash, 64 );
jh512_4way_update( &ctx.jh, vhash, 64 );
jh512_4way_close( &ctx.jh, vhash );
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
// Serial