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

v3.12.4

Browse Source
This commit is contained in:
Jay D Dee
2020-02-21 16:34:53 -05:00
parent b93be8816a
commit 6d66051de6
8 changed files with 299 additions and 125 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
algo/lyra2/lyra2-gate.c
View File

@@ -94,12 +94,12 @@ bool lyra2rev2_thread_init()
const int64_t ROW_LEN_BYTES = ROW_LEN_INT64 * 8;
int size = (int64_t)ROW_LEN_BYTES * 4; // nRows;
#if defined (LYRA2REV2_8WAY)
#if defined (LYRA2REV2_16WAY)
l2v2_wholeMatrix = _mm_malloc( 2 * size, 64 ); // 2 way
init_lyra2rev2_8way_ctx();;
#elif defined (LYRA2REV2_4WAY)
init_lyra2rev2_16way_ctx();;
#elif defined (LYRA2REV2_8WAY)
l2v2_wholeMatrix = _mm_malloc( size, 64 );
init_lyra2rev2_4way_ctx();;
init_lyra2rev2_8way_ctx();;
#else
l2v2_wholeMatrix = _mm_malloc( size, 64 );
init_lyra2rev2_ctx();
@@ -109,12 +109,12 @@ bool lyra2rev2_thread_init()
bool register_lyra2rev2_algo( algo_gate_t* gate )
{
#if defined (LYRA2REV2_8WAY)
#if defined (LYRA2REV2_16WAY)
gate->scanhash = (void*)&scanhash_lyra2rev2_16way;
gate->hash = (void*)&lyra2rev2_16way_hash;
#elif defined (LYRA2REV2_8WAY)
gate->scanhash = (void*)&scanhash_lyra2rev2_8way;
gate->hash = (void*)&lyra2rev2_8way_hash;
#elif defined (LYRA2REV2_4WAY)
gate->scanhash = (void*)&scanhash_lyra2rev2_4way;
gate->hash = (void*)&lyra2rev2_4way_hash;
#else
gate->scanhash = (void*)&scanhash_lyra2rev2;
gate->hash = (void*)&lyra2rev2_hash;

20
algo/lyra2/lyra2-gate.h
View File

@@ -51,30 +51,32 @@ bool init_lyra2rev3_ctx();
//////////////////////////////////
#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
#define LYRA2REV2_8WAY 1
#define LYRA2REV2_16WAY 1
#elif defined(__AVX2__)
#define LYRA2REV2_4WAY 1
#define LYRA2REV2_8WAY 1
#endif
extern __thread uint64_t* l2v2_wholeMatrix;
bool register_lyra2rev2_algo( algo_gate_t* gate );
#if defined(LYRA2REV2_8WAY)
#if defined(LYRA2REV2_16WAY)
void lyra2rev2_16way_hash( void *state, const void *input );
int scanhash_lyra2rev2_16way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool init_lyra2rev2_16way_ctx();
#elif defined(LYRA2REV2_8WAY)
void lyra2rev2_8way_hash( void *state, const void *input );
int scanhash_lyra2rev2_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool init_lyra2rev2_8way_ctx();
#elif defined(LYRA2REV2_4WAY)
void lyra2rev2_4way_hash( void *state, const void *input );
int scanhash_lyra2rev2_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool init_lyra2rev2_4way_ctx();
#else
void lyra2rev2_hash( void *state, const void *input );
int scanhash_lyra2rev2( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );

282
algo/lyra2/lyra2rev2-4way.c
View File

@@ -8,12 +8,30 @@
#include "algo/cubehash/cube-hash-2way.h"
#if 0
void lyra2rev2_8way_hash( void *state, const void *input )
#if defined (LYRA2REV2_16WAY)
typedef struct {
blake256_16way_context blake;
keccak256_8way_context keccak;
cubehashParam cube;
skein256_8way_context skein;
bmw256_16way_context bmw;
} lyra2v2_16way_ctx_holder __attribute__ ((aligned (64)));
static lyra2v2_16way_ctx_holder l2v2_16way_ctx;
bool init_lyra2rev2_16way_ctx()
{
uint32_t vhash[8*8] __attribute__ ((aligned (128)));
uint32_t vhashA[8*8] __attribute__ ((aligned (64)));
uint32_t vhashB[8*8] __attribute__ ((aligned (64)));
keccak256_8way_init( &l2v2_16way_ctx.keccak );
cubehashInit( &l2v2_16way_ctx.cube, 256, 16, 32 );
skein256_8way_init( &l2v2_16way_ctx.skein );
bmw256_16way_init( &l2v2_16way_ctx.bmw );
return true;
}
void lyra2rev2_16way_hash( void *state, const void *input )
{
uint32_t vhash[8*16] __attribute__ ((aligned (128)));
uint32_t hash0[8] __attribute__ ((aligned (64)));
uint32_t hash1[8] __attribute__ ((aligned (64)));
uint32_t hash2[8] __attribute__ ((aligned (64)));
@@ -22,35 +40,60 @@ void lyra2rev2_8way_hash( void *state, const void *input )
uint32_t hash5[8] __attribute__ ((aligned (64)));
uint32_t hash6[8] __attribute__ ((aligned (64)));
uint32_t hash7[8] __attribute__ ((aligned (64)));
lyra2v2_8way_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &l2v2_8way_ctx, sizeof(l2v2_8way_ctx) );
uint32_t hash8[8] __attribute__ ((aligned (64)));
uint32_t hash9[8] __attribute__ ((aligned (64)));
uint32_t hash10[8] __attribute__ ((aligned (64)));
uint32_t hash11[8] __attribute__ ((aligned (64)));
uint32_t hash12[8] __attribute__ ((aligned (64)));
uint32_t hash13[8] __attribute__ ((aligned (64)));
uint32_t hash14[8] __attribute__ ((aligned (64)));
uint32_t hash15[8] __attribute__ ((aligned (64)));
lyra2v2_16way_ctx_holder ctx __attribute__ ((aligned (64)));
memcpy( &ctx, &l2v2_16way_ctx, sizeof(l2v2_16way_ctx) );
blake256_8way_update( &ctx.blake, input + (64<<3), 16 );
blake256_8way_close( &ctx.blake, vhash );
blake256_16way_update( &ctx.blake, input + (64<<4), 16 );
blake256_16way_close( &ctx.blake, vhash );
rintrlv_8x32_8x64( vhashA, vhash, 256 );
dintrlv_16x32( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7,
hash8, hash9, hash10, hash11,
hash12, hash13, hash14, hash15, vhash, 256 );
keccak256_8way_update( &ctx.keccak, vhashA, 32 );
intrlv_8x64( vhash, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, 256 );
keccak256_8way_update( &ctx.keccak, vhash, 32 );
keccak256_8way_close( &ctx.keccak, vhash );
dintrlv_8x64( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, vhash, 256 );
intrlv_8x64( vhash, hash8, hash9, hash10, hash11,
hash12, hash13, hash14, hash15, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
cubehash_full( &ctx.cube, (byte*) hash2, 256, (const byte*) hash2, 32 );
cubehash_full( &ctx.cube, (byte*) hash3, 256, (const byte*) hash3, 32 );
cubehash_full( &ctx.cube, (byte*) hash4, 256, (const byte*) hash4, 32 );
cubehash_full( &ctx.cube, (byte*) hash5, 256, (const byte*) hash5, 32 );
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
keccak256_8way_init( &ctx.keccak );
keccak256_8way_update( &ctx.keccak, vhash, 32 );
keccak256_8way_close( &ctx.keccak, vhash );
dintrlv_8x64( hash8, hash9, hash10, hash11,
hash12, hash13, hash14, hash5, vhash, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
cubehash_full( &ctx.cube, (byte*) hash2, 256, (const byte*) hash2, 32 );
cubehash_full( &ctx.cube, (byte*) hash3, 256, (const byte*) hash3, 32 );
cubehash_full( &ctx.cube, (byte*) hash4, 256, (const byte*) hash4, 32 );
cubehash_full( &ctx.cube, (byte*) hash5, 256, (const byte*) hash5, 32 );
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
cubehash_full( &ctx.cube, (byte*) hash8, 256, (const byte*) hash8, 32 );
cubehash_full( &ctx.cube, (byte*) hash9, 256, (const byte*) hash9, 32 );
cubehash_full( &ctx.cube, (byte*) hash10, 256, (const byte*) hash10, 32 );
cubehash_full( &ctx.cube, (byte*) hash11, 256, (const byte*) hash11, 32 );
cubehash_full( &ctx.cube, (byte*) hash12, 256, (const byte*) hash12, 32 );
cubehash_full( &ctx.cube, (byte*) hash13, 256, (const byte*) hash13, 32 );
cubehash_full( &ctx.cube, (byte*) hash14, 256, (const byte*) hash14, 32 );
cubehash_full( &ctx.cube, (byte*) hash15, 256, (const byte*) hash15, 32 );
// cube_4way_update_close( &ctx.cube, vhashA, vhashA, 32 );
// cube_4way_init( &ctx.cube, 256, 16, 32 );
// cube_4way_update_close( &ctx.cube, vhashB, vhashB, 32 );
//
// dintrlv_4x128( hash0, hash1, hash2, hash3, vhashA, 256 );
// dintrlv_4x128( hash4, hash5, hash6, hash7, vhashB, 256 );
intrlv_2x256( vhash, hash0, hash1, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
@@ -64,61 +107,127 @@ void lyra2rev2_8way_hash( void *state, const void *input )
intrlv_2x256( vhash, hash6, hash7, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash6, hash7, vhash, 256 );
intrlv_2x256( vhash, hash8, hash9, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash8, hash9, vhash, 256 );
intrlv_2x256( vhash, hash10, hash11, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash10, hash11, vhash, 256 );
intrlv_2x256( vhash, hash12, hash13, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash12, hash13, vhash, 256 );
intrlv_2x256( vhash, hash14, hash15, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash14, hash15, vhash, 256 );
intrlv_8x64( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7, 256 );
intrlv_8x64( vhash, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, 256 );
skein256_8way_update( &ctx.skein, vhash, 32 );
skein256_8way_close( &ctx.skein, vhash );
dintrlv_8x64( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, vhash, 256 );
intrlv_8x64( vhash, hash8, hash9, hash10, hash11, hash12,
hash13, hash14, hash15, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
cubehash_full( &ctx.cube, (byte*) hash2, 256, (const byte*) hash2, 32 );
cubehash_full( &ctx.cube, (byte*) hash3, 256, (const byte*) hash3, 32 );
cubehash_full( &ctx.cube, (byte*) hash4, 256, (const byte*) hash4, 32 );
cubehash_full( &ctx.cube, (byte*) hash5, 256, (const byte*) hash5, 32 );
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
skein256_8way_init( &ctx.skein );
skein256_8way_update( &ctx.skein, vhash, 32 );
skein256_8way_close( &ctx.skein, vhash );
// cube_4way_init( &ctx.cube, 256, 16, 32 );
// cube_4way_update_close( &ctx.cube, vhashA, vhashA, 32 );
// cube_4way_init( &ctx.cube, 256, 16, 32 );
// cube_4way_update_close( &ctx.cube, vhashB, vhashB, 32 );
//
// dintrlv_4x128( hash0, hash1, hash2, hash3, vhashA, 256 );
// dintrlv_4x128( hash4, hash5, hash6, hash7, vhashB, 256 );
dintrlv_8x64( hash8, hash9, hash10, hash11,
hash12, hash13, hash14, hash15, vhash, 256 );
intrlv_8x32( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
cubehash_full( &ctx.cube, (byte*) hash2, 256, (const byte*) hash2, 32 );
cubehash_full( &ctx.cube, (byte*) hash3, 256, (const byte*) hash3, 32 );
cubehash_full( &ctx.cube, (byte*) hash4, 256, (const byte*) hash4, 32 );
cubehash_full( &ctx.cube, (byte*) hash5, 256, (const byte*) hash5, 32 );
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
cubehash_full( &ctx.cube, (byte*) hash8, 256, (const byte*) hash8, 32 );
cubehash_full( &ctx.cube, (byte*) hash9, 256, (const byte*) hash9, 32 );
cubehash_full( &ctx.cube, (byte*) hash10, 256, (const byte*) hash10, 32 );
cubehash_full( &ctx.cube, (byte*) hash11, 256, (const byte*) hash11, 32 );
cubehash_full( &ctx.cube, (byte*) hash12, 256, (const byte*) hash12, 32 );
cubehash_full( &ctx.cube, (byte*) hash13, 256, (const byte*) hash13, 32 );
cubehash_full( &ctx.cube, (byte*) hash14, 256, (const byte*) hash14, 32 );
cubehash_full( &ctx.cube, (byte*) hash15, 256, (const byte*) hash15, 32 );
bmw256_8way_update( &ctx.bmw, vhash, 32 );
bmw256_8way_close( &ctx.bmw, state );
intrlv_16x32( vhash, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7,
hash8, hash9, hash10, hash11,
hash12, hash13, hash14, hash15, 256 );
bmw256_16way_update( &ctx.bmw, vhash, 32 );
bmw256_16way_close( &ctx.bmw, state );
}
#endif
int scanhash_lyra2rev2_16way( struct work *work, const uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[8*16] __attribute__ ((aligned (128)));
uint32_t vdata[20*16] __attribute__ ((aligned (64)));
uint32_t *hashd7 = &hash[7*16];
uint32_t lane_hash[8] __attribute__ ((aligned (32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
const uint32_t last_nonce = max_nonce - 16;
uint32_t n = first_nonce;
const uint32_t targ32 = ptarget[7];
__m512i *noncev = (__m512i*)vdata + 19;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
if ( bench ) ptarget[7] = 0x0000ff;
mm512_bswap32_intrlv80_16x32( vdata, pdata );
*noncev = _mm512_set_epi32( n+15, n+14, n+13, n+12, n+11, n+10, n+ 9, n+ 8,
n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n+ 1, n );
blake256_16way_init( &l2v2_16way_ctx.blake );
blake256_16way_update( &l2v2_16way_ctx.blake, vdata, 64 );
#if defined (LYRA2REV2_8WAY)
do
{
lyra2rev2_16way_hash( hash, vdata );
for ( int lane = 0; lane < 16; lane++ )
if ( unlikely( hashd7[lane] <= targ32 ) )
{
extr_lane_16x32( lane_hash, hash, lane, 256 );
if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n + lane );
submit_lane_solution( work, lane_hash, mythr, lane );
}
}
*noncev = _mm512_add_epi32( *noncev, m512_const1_32( 16 ) );
n += 16;
} while ( likely( (n < last_nonce) && !work_restart[thr_id].restart ) );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
#elif defined (LYRA2REV2_8WAY)
typedef struct {
blake256_8way_context blake;
keccak256_8way_context keccak;
keccak256_4way_context keccak;
cubehashParam cube;
skein256_8way_context skein;
bmw256_8way_context bmw;
skein256_4way_context skein;
bmw256_8way_context bmw;
} lyra2v2_8way_ctx_holder __attribute__ ((aligned (64)));
static lyra2v2_8way_ctx_holder l2v2_8way_ctx;
bool init_lyra2rev2_8way_ctx()
{
keccak256_8way_init( &l2v2_8way_ctx.keccak );
keccak256_4way_init( &l2v2_8way_ctx.keccak );
cubehashInit( &l2v2_8way_ctx.cube, 256, 16, 32 );
skein256_8way_init( &l2v2_8way_ctx.skein );
skein256_4way_init( &l2v2_8way_ctx.skein );
bmw256_8way_init( &l2v2_8way_ctx.bmw );
return true;
}
@@ -126,7 +235,6 @@ bool init_lyra2rev2_8way_ctx()
void lyra2rev2_8way_hash( void *state, const void *input )
{
uint32_t vhash[8*8] __attribute__ ((aligned (128)));
uint32_t vhashA[8*8] __attribute__ ((aligned (64)));
uint32_t hash0[8] __attribute__ ((aligned (64)));
uint32_t hash1[8] __attribute__ ((aligned (64)));
uint32_t hash2[8] __attribute__ ((aligned (64)));
@@ -141,14 +249,19 @@ void lyra2rev2_8way_hash( void *state, const void *input )
blake256_8way_update( &ctx.blake, input + (64<<3), 16 );
blake256_8way_close( &ctx.blake, vhash );
rintrlv_8x32_8x64( vhashA, vhash, 256 );
keccak256_8way_update( &ctx.keccak, vhashA, 32 );
keccak256_8way_close( &ctx.keccak, vhash );
dintrlv_8x64( hash0, hash1, hash2, hash3,
dintrlv_8x32( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, vhash, 256 );
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 256 );
keccak256_4way_update( &ctx.keccak, vhash, 32 );
keccak256_4way_close( &ctx.keccak, vhash );
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 256 );
intrlv_4x64( vhash, hash4, hash5, hash6, hash7, 256 );
keccak256_4way_init( &ctx.keccak );
keccak256_4way_update( &ctx.keccak, vhash, 32 );
keccak256_4way_close( &ctx.keccak, vhash );
dintrlv_4x64( hash4, hash5, hash6, hash7, vhash, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
cubehash_full( &ctx.cube, (byte*) hash2, 256, (const byte*) hash2, 32 );
@@ -158,27 +271,25 @@ void lyra2rev2_8way_hash( void *state, const void *input )
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
intrlv_2x256( vhash, hash0, hash1, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash0, hash1, vhash, 256 );
intrlv_2x256( vhash, hash2, hash3, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash2, hash3, vhash, 256 );
intrlv_2x256( vhash, hash4, hash5, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash4, hash5, vhash, 256 );
intrlv_2x256( vhash, hash6, hash7, 256 );
LYRA2REV2_2WAY( l2v2_wholeMatrix, vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash6, hash7, vhash, 256 );
intrlv_8x64( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7, 256 );
skein256_8way_update( &ctx.skein, vhash, 32 );
skein256_8way_close( &ctx.skein, vhash );
dintrlv_8x64( hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, vhash, 256 );
LYRA2REV2( l2v2_wholeMatrix, hash0, 32, hash0, 32, hash0, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash1, 32, hash1, 32, hash1, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash2, 32, hash2, 32, hash2, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash3, 32, hash3, 32, hash3, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash4, 32, hash4, 32, hash4, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash5, 32, hash5, 32, hash5, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash6, 32, hash6, 32, hash6, 32, 1, 4, 4 );
LYRA2REV2( l2v2_wholeMatrix, hash7, 32, hash7, 32, hash7, 32, 1, 4, 4 );
intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 256 );
skein256_4way_update( &ctx.skein, vhash, 32 );
skein256_4way_close( &ctx.skein, vhash );
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 256 );
intrlv_4x64( vhash, hash4, hash5, hash6, hash7, 256 );
skein256_4way_init( &ctx.skein );
skein256_4way_update( &ctx.skein, vhash, 32 );
skein256_4way_close( &ctx.skein, vhash );
dintrlv_4x64( hash4, hash5, hash6, hash7, vhash, 256 );
cubehash_full( &ctx.cube, (byte*) hash0, 256, (const byte*) hash0, 32 );
cubehash_full( &ctx.cube, (byte*) hash1, 256, (const byte*) hash1, 32 );
@@ -189,8 +300,8 @@ void lyra2rev2_8way_hash( void *state, const void *input )
cubehash_full( &ctx.cube, (byte*) hash6, 256, (const byte*) hash6, 32 );
cubehash_full( &ctx.cube, (byte*) hash7, 256, (const byte*) hash7, 32 );
intrlv_8x32( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7, 256 );
intrlv_8x32( vhash, hash0, hash1, hash2, hash3,
hash4, hash5, hash6, hash7, 256 );
bmw256_8way_update( &ctx.bmw, vhash, 32 );
bmw256_8way_close( &ctx.bmw, state );
@@ -223,7 +334,6 @@ int scanhash_lyra2rev2_8way( struct work *work, const uint32_t max_nonce,
do
{
lyra2rev2_8way_hash( hash, vdata );
pdata[19] = n;
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( hashd7[lane] <= targ32 ) )
@@ -243,6 +353,9 @@ int scanhash_lyra2rev2_8way( struct work *work, const uint32_t max_nonce,
return 0;
}
#endif
/*
#elif defined (LYRA2REV2_4WAY)
typedef struct {
@@ -367,3 +480,4 @@ int scanhash_lyra2rev2_4way( struct work *work, uint32_t max_nonce,
}
#endif
*/