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

v3.4.8 release

Browse Source
This commit is contained in:
Jay D Dee
2016-10-17 14:16:28 -04:00
parent 4ecf9555da
commit 6f28a67f9c
22 changed files with 1117 additions and 656 deletions
Download Patch File Download Diff File Expand all files Collapse all files

141
algo/lyra2/sponge.h
View File

@@ -23,42 +23,34 @@
#define SPONGE_H_
#include <stdint.h>
#include "avxdefs.h"
/* Blake2b IV Array */
#if defined(__GNUC__)
#define ALIGN __attribute__ ((aligned(32)))
#elif defined(_MSC_VER)
#define ALIGN __declspec(align(32))
#else
#define ALIGN
#endif
/*Blake2b IV Array*/
static const uint64_t blake2b_IV[8] =
{
0x6a09e667f3bcc908ULL, 0xbb67ae8584caa73bULL,
0x3c6ef372fe94f82bULL, 0xa54ff53a5f1d36f1ULL,
0x510e527fade682d1ULL, 0x9b05688c2b3e6c1fULL,
0x1f83d9abfb41bd6bULL, 0x5be0cd19137e2179ULL
0x6a09e667f3bcc908ULL, 0xbb67ae8584caa73bULL,
0x3c6ef372fe94f82bULL, 0xa54ff53a5f1d36f1ULL,
0x510e527fade682d1ULL, 0x9b05688c2b3e6c1fULL,
0x1f83d9abfb41bd6bULL, 0x5be0cd19137e2179ULL
};
/* Blake2b's rotation */
static __inline uint64_t rotr64(const uint64_t w, const unsigned c) {
#ifdef _MSC_VER
return _rotr64(w, c);
#else
return ( w >> c ) | ( w << ( 64 - c ) );
#endif
/*Blake2b's rotation*/
static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
return ( w >> c ) | ( w << ( 64 - c ) );
}
#if defined __AVX2__
// only available with avx2
// rotate each uint64 c bits
// returns _m256i
#define mm256_rotr_64(w,c) _mm256_or_si256(_mm256_srli_epi64(w, c), \
_mm256_slli_epi64(w, 64 - c))
// Rotate 4 uint64 (256 bits) by one uint64 (64 bits)
// returns __m256i
#define mm256_rotl256_1x64(s) _mm256_permute4x64_epi64( s, 0x39 )
#define mm256_rotr256_1x64(s) _mm256_permute4x64_epi64( s, 0x93 )
// swap hi and lo 128 bits in 256 bit vector
// returns _m256i
#define mm256_swap128(s) _mm256_permute2f128_si256( s, s, 1 )
// init vectors from memory
// returns void, updates defines and inits implicit args a, b, c, d
#define LYRA_INIT_AVX2 \
@@ -88,15 +80,29 @@ static __inline uint64_t rotr64(const uint64_t w, const unsigned c) {
c = _mm256_add_epi64( c, d ); \
b = mm256_rotr_64( _mm256_xor_si256( b, c ), 63 );
#define LYRA_ROUND_AVX2 \
G_4X64( a[0], a[1], a[2], a[3] ); \
a[1] = mm256_rotl256_1x64( a[1]); \
a[2] = mm256_swap128( a[2] ); \
a[3] = mm256_rotr256_1x64( a[3] ); \
G_4X64( a[0], a[1], a[2], a[3] ); \
a[1] = mm256_rotr256_1x64( a[1] ); \
a[2] = mm256_swap128( a[2] ); \
a[3] = mm256_rotl256_1x64( a[3] );
#define LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
G_4X64( s0, s1, s2, s3 ); \
s1 = mm256_rotl256_1x64( s1); \
s2 = mm256_swap128( s2 ); \
s3 = mm256_rotr256_1x64( s3 ); \
G_4X64( s0, s1, s2, s3 ); \
s1 = mm256_rotr256_1x64( s1 ); \
s2 = mm256_swap128( s2 ); \
s3 = mm256_rotl256_1x64( s3 );
#define LYRA_12_ROUNDS_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
#else
// only available with avx
@@ -148,6 +154,7 @@ static __inline uint64_t rotr64(const uint64_t w, const unsigned c) {
#endif // AVX2
/*
#if defined __AVX__
// can coexist with AVX2
@@ -161,7 +168,7 @@ static __inline uint64_t rotr64(const uint64_t w, const unsigned c) {
#define mm128_swap128(s0, s1) s0 = _mm_xor_si128(s0, s1); \
s1 = _mm_xor_si128(s0, s1); \
s0 = _mm_xor_si128(s0, s1);
// swap uint64 in 128 bit source vector, equivalent of rotating 128 bits by
// 64 bits (8 bytes)
// __m128i
@@ -193,30 +200,33 @@ static __inline uint64_t rotr64(const uint64_t w, const unsigned c) {
} while(0)
#endif // AVX
*/
/* Blake2b's G function */
#define G(r,i,a,b,c,d) do { \
a = a + b; \
d = rotr64(d ^ a, 32); \
c = c + d; \
b = rotr64(b ^ c, 24); \
a = a + b; \
d = rotr64(d ^ a, 16); \
c = c + d; \
b = rotr64(b ^ c, 63); \
// Scalar
//Blake2b's G function
#define G(r,i,a,b,c,d) \
do { \
a = a + b; \
d = rotr64(d ^ a, 32); \
c = c + d; \
b = rotr64(b ^ c, 24); \
a = a + b; \
d = rotr64(d ^ a, 16); \
c = c + d; \
b = rotr64(b ^ c, 63); \
} while(0)
/*One Round of the Blake2b's compression function*/
#define ROUND_LYRA(r) \
G(r,0,v[ 0],v[ 4],v[ 8],v[12]); \
G(r,1,v[ 1],v[ 5],v[ 9],v[13]); \
G(r,2,v[ 2],v[ 6],v[10],v[14]); \
G(r,3,v[ 3],v[ 7],v[11],v[15]); \
G(r,4,v[ 0],v[ 5],v[10],v[15]); \
G(r,5,v[ 1],v[ 6],v[11],v[12]); \
G(r,6,v[ 2],v[ 7],v[ 8],v[13]); \
G(r,7,v[ 3],v[ 4],v[ 9],v[14]);
#define ROUND_LYRA(r) \
G(r,0,v[ 0],v[ 4],v[ 8],v[12]); \
G(r,1,v[ 1],v[ 5],v[ 9],v[13]); \
G(r,2,v[ 2],v[ 6],v[10],v[14]); \
G(r,3,v[ 3],v[ 7],v[11],v[15]); \
G(r,4,v[ 0],v[ 5],v[10],v[15]); \
G(r,5,v[ 1],v[ 6],v[11],v[12]); \
G(r,6,v[ 2],v[ 7],v[ 8],v[13]); \
G(r,7,v[ 3],v[ 4],v[ 9],v[14]);
//---- Housekeeping
@@ -224,18 +234,31 @@ void initState(uint64_t state[/*16*/]);
//---- Squeezes
void squeeze(uint64_t *state, unsigned char *out, unsigned int len);
void reducedSqueezeRow0(uint64_t* state, uint64_t* row, const uint32_t nCols);
void reducedSqueezeRow0(uint64_t* state, uint64_t* row, uint64_t nCols);
//---- Absorbs
void absorbBlock(uint64_t *state, const uint64_t *in);
void absorbBlockBlake2Safe(uint64_t *state, const uint64_t *in);
//---- Duplexes
void reducedDuplexRow1(uint64_t *state, uint64_t *rowIn, uint64_t *rowOut, const uint32_t nCols);
void reducedDuplexRowSetup(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut, const uint32_t nCols);
void reducedDuplexRow(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut, const uint32_t nCols);
void reducedDuplexRow1(uint64_t *state, uint64_t *rowIn, uint64_t *rowOut, uint64_t nCols);
void reducedDuplexRowSetup(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut, uint64_t nCols);
void reducedDuplexRow(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut, uint64_t nCols);
//---- Misc
void printArray(unsigned char *array, unsigned int size, char *name);
////////////////////////////////////////////////////////////////////////////////////////////////
////TESTS////
//void reducedDuplexRowc(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut);
//void reducedDuplexRowd(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut);
//void reducedDuplexRowSetupv4(uint64_t *state, uint64_t *rowIn1, uint64_t *rowIn2, uint64_t *rowOut1, uint64_t *rowOut2);
//void reducedDuplexRowSetupv5(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut);
//void reducedDuplexRowSetupv5c(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut);
//void reducedDuplexRowSetupv5d(uint64_t *state, uint64_t *rowIn, uint64_t *rowInOut, uint64_t *rowOut);
/////////////
#endif /* SPONGE_H_ */