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

v23.5

Browse Source
This commit is contained in:
Jay D Dee
2023-10-25 20:36:20 -04:00
parent 31c4dedf59
commit 160608cce5
180 changed files with 10318 additions and 13097 deletions
Download Patch File Download Diff File Expand all files Collapse all files

413
algo/skein/skein-hash-4way.c
View File

@@ -30,20 +30,10 @@
* @author Thomas Pornin <thomas.pornin@cryptolog.com>
*/
#if defined (__AVX2__)
#include <stddef.h>
#include <string.h>
#include "skein-hash-4way.h"
#ifdef __cplusplus
extern "C"{
#endif
#ifdef _MSC_VER
#pragma warning (disable: 4146)
#endif
/*
static const uint64_t IV256[] = {
0xCCD044A12FDB3E13, 0xE83590301A79A9EB,
@@ -414,6 +404,8 @@ do { \
#endif // AVX512
#if defined (__AVX2__)
#define TFBIG_KINIT_4WAY( k0, k1, k2, k3, k4, k5, k6, k7, k8, t0, t1, t2 ) \
k8 = mm256_xor3( mm256_xor3( k0, k1, k2 ), \
mm256_xor3( k3, k4, k5 ), \
@@ -517,6 +509,8 @@ do { \
__m256i h0, h1, h2, h3, h4, h5, h6, h7; \
uint64_t bcount;
#endif // AVX2
#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
void skein256_8way_init( skein256_8way_context *sc )
@@ -807,6 +801,7 @@ skein512_8way_close(void *cc, void *dst)
#endif // AVX512
#if defined(__AVX2__)
void skein256_4way_init( skein256_4way_context *sc )
{
@@ -1099,8 +1094,398 @@ skein512_4way_close(void *cc, void *dst)
skein_big_close_4way(cc, 0, 0, dst, 64);
}
#ifdef __cplusplus
}
#endif
#endif // AVX2
// SSE2 & NEON
#define TFBIG_KINIT_2WAY( k0, k1, k2, k3, k4, k5, k6, k7, k8, t0, t1, t2 ) \
k8 = v128_xor3( v128_xor3( k0, k1, k2 ), \
v128_xor3( k3, k4, k5 ), \
v128_xor3( k6, k7, v128_64( 0x1BD11BDAA9FC1A22 ) ) ); \
t2 = t0 ^ t1;
#define TFBIG_ADDKEY_2WAY(w0, w1, w2, w3, w4, w5, w6, w7, k, t, s) \
w0 = v128_add64( w0, SKBI(k,s,0) ); \
w1 = v128_add64( w1, SKBI(k,s,1) ); \
w2 = v128_add64( w2, SKBI(k,s,2) ); \
w3 = v128_add64( w3, SKBI(k,s,3) ); \
w4 = v128_add64( w4, SKBI(k,s,4) ); \
w5 = v128_add64( w5, v128_add64( SKBI(k,s,5), v128_64( SKBT(t,s,0) ) ) ); \
w6 = v128_add64( w6, v128_add64( SKBI(k,s,6), v128_64( SKBT(t,s,1) ) ) ); \
w7 = v128_add64( w7, v128_add64( SKBI(k,s,7), v128_64(s) ) );
#define TFBIG_MIX_2WAY(x0, x1, rc) \
x0 = v128_add64( x0, x1 ); \
x1 = v128_xor( v128_rol64( x1, rc ), x0 );
#define TFBIG_MIX8_2WAY(w0, w1, w2, w3, w4, w5, w6, w7, rc0, rc1, rc2, rc3) \
TFBIG_MIX_2WAY(w0, w1, rc0); \
TFBIG_MIX_2WAY(w2, w3, rc1); \
TFBIG_MIX_2WAY(w4, w5, rc2); \
TFBIG_MIX_2WAY(w6, w7, rc3);
#define TFBIG_2WAY_4e(s) \
TFBIG_ADDKEY_2WAY(p0, p1, p2, p3, p4, p5, p6, p7, h, t, s); \
TFBIG_MIX8_2WAY(p0, p1, p2, p3, p4, p5, p6, p7, 46, 36, 19, 37); \
TFBIG_MIX8_2WAY(p2, p1, p4, p7, p6, p5, p0, p3, 33, 27, 14, 42); \
TFBIG_MIX8_2WAY(p4, p1, p6, p3, p0, p5, p2, p7, 17, 49, 36, 39); \
TFBIG_MIX8_2WAY(p6, p1, p0, p7, p2, p5, p4, p3, 44, 9, 54, 56);
#define TFBIG_2WAY_4o(s) \
TFBIG_ADDKEY_2WAY(p0, p1, p2, p3, p4, p5, p6, p7, h, t, s); \
TFBIG_MIX8_2WAY(p0, p1, p2, p3, p4, p5, p6, p7, 39, 30, 34, 24); \
TFBIG_MIX8_2WAY(p2, p1, p4, p7, p6, p5, p0, p3, 13, 50, 10, 17); \
TFBIG_MIX8_2WAY(p4, p1, p6, p3, p0, p5, p2, p7, 25, 29, 39, 43); \
TFBIG_MIX8_2WAY(p6, p1, p0, p7, p2, p5, p4, p3, 8, 35, 56, 22);
// scale buf offset by 4
#define UBI_BIG_2WAY(etype, extra) \
do { \
uint64_t t0, t1, t2; \
v128u64_t h8; \
v128u64_t m0 = buf[0]; \
v128u64_t m1 = buf[1]; \
v128u64_t m2 = buf[2]; \
v128u64_t m3 = buf[3]; \
v128u64_t m4 = buf[4]; \
v128u64_t m5 = buf[5]; \
v128u64_t m6 = buf[6]; \
v128u64_t m7 = buf[7]; \
\
v128u64_t p0 = m0; \
v128u64_t p1 = m1; \
v128u64_t p2 = m2; \
v128u64_t p3 = m3; \
v128u64_t p4 = m4; \
v128u64_t p5 = m5; \
v128u64_t p6 = m6; \
v128u64_t p7 = m7; \
t0 = (uint64_t)(bcount << 6) + (uint64_t)(extra); \
t1 = (bcount >> 58) + ((uint64_t)(etype) << 55); \
TFBIG_KINIT_2WAY(h0, h1, h2, h3, h4, h5, h6, h7, h8, t0, t1, t2); \
TFBIG_2WAY_4e(0); \
TFBIG_2WAY_4o(1); \
TFBIG_2WAY_4e(2); \
TFBIG_2WAY_4o(3); \
TFBIG_2WAY_4e(4); \
TFBIG_2WAY_4o(5); \
TFBIG_2WAY_4e(6); \
TFBIG_2WAY_4o(7); \
TFBIG_2WAY_4e(8); \
TFBIG_2WAY_4o(9); \
TFBIG_2WAY_4e(10); \
TFBIG_2WAY_4o(11); \
TFBIG_2WAY_4e(12); \
TFBIG_2WAY_4o(13); \
TFBIG_2WAY_4e(14); \
TFBIG_2WAY_4o(15); \
TFBIG_2WAY_4e(16); \
TFBIG_2WAY_4o(17); \
TFBIG_ADDKEY_2WAY(p0, p1, p2, p3, p4, p5, p6, p7, h, t, 18); \
h0 = v128_xor( m0, p0 );\
h1 = v128_xor( m1, p1 );\
h2 = v128_xor( m2, p2 );\
h3 = v128_xor( m3, p3 );\
h4 = v128_xor( m4, p4 );\
h5 = v128_xor( m5, p5 );\
h6 = v128_xor( m6, p6 );\
h7 = v128_xor( m7, p7 );\
} while (0)
#define DECL_STATE_BIG_2WAY \
v128u64_t h0, h1, h2, h3, h4, h5, h6, h7; \
uint64_t bcount;
void skein256_2x64_init( skein256_2x64_context *sc )
{
sc->h0 = v128_64( 0xCCD044A12FDB3E13 );
sc->h1 = v128_64( 0xE83590301A79A9EB );
sc->h2 = v128_64( 0x55AEA0614F816E6F );
sc->h3 = v128_64( 0x2A2767A4AE9B94DB );
sc->h4 = v128_64( 0xEC06025E74DD7683 );
sc->h5 = v128_64( 0xE7A436CDC4746251 );
sc->h6 = v128_64( 0xC36FBAF9393AD185 );
sc->h7 = v128_64( 0x3EEDBA1833EDFC13 );
sc->bcount = 0;
sc->ptr = 0;
}
void skein512_2x64_init( skein512_2x64_context *sc )
{
sc->h0 = v128_64( 0x4903ADFF749C51CE );
sc->h1 = v128_64( 0x0D95DE399746DF03 );
sc->h2 = v128_64( 0x8FD1934127C79BCE );
sc->h3 = v128_64( 0x9A255629FF352CB1 );
sc->h4 = v128_64( 0x5DB62599DF6CA7B0 );
sc->h5 = v128_64( 0xEABE394CA9D5C3F4 );
sc->h6 = v128_64( 0x991112C71A75B523 );
sc->h7 = v128_64( 0xAE18A40B660FCC33 );
sc->bcount = 0;
sc->ptr = 0;
}
static void
skein_big_core_2way( skein512_2x64_context *sc, const void *data,
size_t len )
{
v128u64_t *vdata = (v128u64_t*)data;
v128u64_t *buf;
size_t ptr;
unsigned first;
DECL_STATE_BIG_2WAY
buf = sc->buf;
ptr = sc->ptr;
const int buf_size = 64; // 64 * _m256i
if ( len <= buf_size - ptr )
{
v128_memcpy( buf + (ptr>>3), vdata, len>>3 );
sc->ptr = ptr + len;
if ( ptr < buf_size ) return;
}
READ_STATE_BIG( sc );
first = ( bcount == 0 ) << 7;
do {
size_t clen;
if ( ptr == buf_size )
{
bcount ++;
UBI_BIG_2WAY( 96 + first, 0 );
first = 0;
ptr = 0;
}
clen = buf_size - ptr;
if ( clen > len )
clen = len;
len -= clen;
if ( len == 0 ) break;
v128_memcpy( buf + (ptr>>3), vdata, clen>>3 );
ptr += clen;
vdata += (clen>>3);
len -= clen;
} while ( len > 0 );
WRITE_STATE_BIG( sc );
sc->ptr = ptr;
}
static void
skein_big_close_2way( skein512_2x64_context *sc, unsigned ub, unsigned n,
void *dst, size_t out_len )
{
v128u64_t *buf;
size_t ptr;
unsigned et;
DECL_STATE_BIG_2WAY
buf = sc->buf;
ptr = sc->ptr;
const int buf_size = 64;
READ_STATE_BIG(sc);
if ( ptr )
{
v128_memset_zero( buf + (ptr>>3), (buf_size - ptr) >> 3 );
et = 352 + ((bcount == 0) << 7);
UBI_BIG_2WAY( et, ptr );
}
v128_memset_zero( buf, buf_size >> 3 );
bcount = 0;
UBI_BIG_2WAY( 510, 8 );
buf[0] = h0;
buf[1] = h1;
buf[2] = h2;
buf[3] = h3;
buf[4] = h4;
buf[5] = h5;
buf[6] = h6;
buf[7] = h7;
v128_memcpy( dst, buf, out_len >> 3 );
}
void
skein512_2x64_full( skein512_2x64_context *sc, void *out, const void *data,
size_t len )
{
v128u64_t h0, h1, h2, h3, h4, h5, h6, h7;
v128u64_t *vdata = (v128u64_t*)data;
v128u64_t *buf = sc->buf;
size_t ptr = 0;
unsigned first;
const int buf_size = 64; // 64 * __m256i
uint64_t bcount = 0;
h0 = v128_64( 0x4903ADFF749C51CE );
h1 = v128_64( 0x0D95DE399746DF03 );
h2 = v128_64( 0x8FD1934127C79BCE );
h3 = v128_64( 0x9A255629FF352CB1 );
h4 = v128_64( 0x5DB62599DF6CA7B0 );
h5 = v128_64( 0xEABE394CA9D5C3F4 );
h6 = v128_64( 0x991112C71A75B523 );
h7 = v128_64( 0xAE18A40B660FCC33 );
// Update
if ( len <= buf_size - ptr )
{
v128_memcpy( buf + (ptr>>3), vdata, len>>3 );
ptr += len;
}
else
{
first = ( bcount == 0 ) << 7;
do {
size_t clen;
if ( ptr == buf_size )
{
bcount ++;
UBI_BIG_2WAY( 96 + first, 0 );
first = 0;
ptr = 0;
}
clen = buf_size - ptr;
if ( clen > len )
clen = len;
v128_memcpy( buf + (ptr>>3), vdata, clen>>3 );
ptr += clen;
vdata += (clen>>3);
len -= clen;
} while ( len > 0 );
}
// Close
unsigned et;
v128_memset_zero( buf + (ptr>>3), (buf_size - ptr) >> 3 );
et = 352 + ((bcount == 0) << 7);
UBI_BIG_2WAY( et, ptr );
v128_memset_zero( buf, buf_size >> 3 );
bcount = 0;
UBI_BIG_2WAY( 510, 8 );
casti_v128u64( out, 0 ) = h0;
casti_v128u64( out, 1 ) = h1;
casti_v128u64( out, 2 ) = h2;
casti_v128u64( out, 3 ) = h3;
casti_v128u64( out, 4 ) = h4;
casti_v128u64( out, 5 ) = h5;
casti_v128u64( out, 6 ) = h6;
casti_v128u64( out, 7 ) = h7;
}
void
skein512_2x64_prehash64( skein512_2x64_context *sc, const void *data )
{
v128u64_t *vdata = (v128u64_t*)data;
v128u64_t *buf = sc->buf;
buf[0] = vdata[0];
buf[1] = vdata[1];
buf[2] = vdata[2];
buf[3] = vdata[3];
buf[4] = vdata[4];
buf[5] = vdata[5];
buf[6] = vdata[6];
buf[7] = vdata[7];
register v128u64_t h0 = v128_64( 0x4903ADFF749C51CE );
register v128u64_t h1 = v128_64( 0x0D95DE399746DF03 );
register v128u64_t h2 = v128_64( 0x8FD1934127C79BCE );
register v128u64_t h3 = v128_64( 0x9A255629FF352CB1 );
register v128u64_t h4 = v128_64( 0x5DB62599DF6CA7B0 );
register v128u64_t h5 = v128_64( 0xEABE394CA9D5C3F4 );
register v128u64_t h6 = v128_64( 0x991112C71A75B523 );
register v128u64_t h7 = v128_64( 0xAE18A40B660FCC33 );
uint64_t bcount = 1;
UBI_BIG_2WAY( 224, 0 );
sc->h0 = h0;
sc->h1 = h1;
sc->h2 = h2;
sc->h3 = h3;
sc->h4 = h4;
sc->h5 = h5;
sc->h6 = h6;
sc->h7 = h7;
}
void
skein512_2x64_final16( skein512_2x64_context *sc, void *out, const void *data )
{
v128u64_t *vdata = (v128u64_t*)data;
v128u64_t *buf = sc->buf;
register v128u64_t h0 = sc->h0;
register v128u64_t h1 = sc->h1;
register v128u64_t h2 = sc->h2;
register v128u64_t h3 = sc->h3;
register v128u64_t h4 = sc->h4;
register v128u64_t h5 = sc->h5;
register v128u64_t h6 = sc->h6;
register v128u64_t h7 = sc->h7;
const v128u64_t zero = v128_zero;
buf[0] = vdata[0];
buf[1] = vdata[1];
buf[2] = zero;
buf[3] = zero;
buf[4] = zero;
buf[5] = zero;
buf[6] = zero;
buf[7] = zero;
uint64_t bcount = 1;
UBI_BIG_2WAY( 352, 16 );
buf[0] = zero;
buf[1] = zero;
bcount = 0;
UBI_BIG_2WAY( 510, 8 );
casti_v128u64( out, 0 ) = h0;
casti_v128u64( out, 1 ) = h1;
casti_v128u64( out, 2 ) = h2;
casti_v128u64( out, 3 ) = h3;
casti_v128u64( out, 4 ) = h4;
casti_v128u64( out, 5 ) = h5;
casti_v128u64( out, 6 ) = h6;
casti_v128u64( out, 7 ) = h7;
}
// Broken for 80 bytes, use prehash.
void
skein256_2x64_update(void *cc, const void *data, size_t len)
{
skein_big_core_2way(cc, data, len);
}
void
skein256_2x64_close(void *cc, void *dst)
{
skein_big_close_2way(cc, 0, 0, dst, 32);
}
// Broken for 80 & 128 bytes, use prehash or full
void
skein512_2x64_update(void *cc, const void *data, size_t len)
{
skein_big_core_2way(cc, data, len);
}
void
skein512_2x64_close(void *cc, void *dst)
{
skein_big_close_2way(cc, 0, 0, dst, 64);
}
#endif