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

v3.10.4

Browse Source
This commit is contained in:
Jay D Dee
2019-12-17 00:57:35 -05:00
parent a17ff6f189
commit d741f1c9a9
51 changed files with 5473 additions and 911 deletions
Download Patch File Download Diff File Expand all files Collapse all files

243
algo/lyra2/sponge-2way.c
View File

@@ -19,7 +19,7 @@
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include "algo-gate.h"
//#include "algo-gate.h"
#include <string.h>
#include <stdio.h>
#include <time.h>
@@ -31,21 +31,31 @@
inline void squeeze_2way( uint64_t *State, byte *Out, unsigned int len )
{
const int len_m256i = len / 32;
const int fullBlocks = len_m256i / BLOCK_LEN_M256I;
const int fullBlocks = len / 32;
__m512i* state = (__m512i*)State;
__m512i* out = (__m512i*)Out;
int i;
//printf("squeeze 1, len= %d, full %d\n", len,fullBlocks);
//Squeezes full blocks
for ( i = 0; i < fullBlocks; i++ )
{
//printf("squeeze 1, %d\n",i);
memcpy_512( out, state, BLOCK_LEN_M256I*2 );
LYRA_ROUND_2WAY_AVX2( state[0], state[1], state[2], state[3] );
out += BLOCK_LEN_M256I*2;
//printf("squeeze 2\n");
LYRA_ROUND_2WAY_AVX512( state[0], state[1], state[2], state[3] );
//printf("squeeze 2\n");
out += BLOCK_LEN_M256I;
}
//Squeezes remaining bytes
memcpy_512( out, state, ( (len_m256i % BLOCK_LEN_M256I) * 2 ) );
// memcpy_512( out, state, ( (len * 2 ) );
}
inline void absorbBlock_2way( uint64_t *State, const uint64_t *In )
@@ -90,7 +100,7 @@ inline void absorbBlockBlake2Safe_2way( uint64_t *State, const uint64_t *In,
state1 = _mm512_xor_si512( state1, in[1] );
LYRA_12_ROUNDS_2WAY_AVX512( state0, state1, state2, state3 );
In += block_len * 2;
In += block_len*2;
}
_mm512_store_si512( (__m512i*)State, state0 );
@@ -109,7 +119,7 @@ inline void reducedSqueezeRow0_2way( uint64_t* State, uint64_t* rowOut,
register __m512i state0, state1, state2, state3;
__m512i* out = (__m512i*)rowOut + ( (nCols-1) * BLOCK_LEN_M256I * 2 );
__m512i* out = (__m512i*)rowOut + ( (nCols-1) * BLOCK_LEN_M256I );
state0 = _mm512_load_si512( (__m512i*)State );
state1 = _mm512_load_si512( (__m512i*)State + 1 );
@@ -132,7 +142,7 @@ inline void reducedSqueezeRow0_2way( uint64_t* State, uint64_t* rowOut,
out[2] = state2;
//Goes to next block (column) that will receive the squeezed data
out -= BLOCK_LEN_M256I * 2;
out -= BLOCK_LEN_M256I;
LYRA_ROUND_2WAY_AVX512( state0, state1, state2, state3 );
}
@@ -143,15 +153,14 @@ inline void reducedSqueezeRow0_2way( uint64_t* State, uint64_t* rowOut,
_mm512_store_si512( (__m512i*)State + 3, state3 );
}
// This function has to deal with gathering 2 256 bit rowin vectors from
// non-contiguous memory. Extra work and performance penalty.
inline void reducedDuplexRow1_2way( uint64_t *State, uint64_t *rowIn,
uint64_t *rowOut, uint64_t nCols )
{
int i;
register __m512i state0, state1, state2, state3;
__m512i *in = (__m256i*)rowIn;
__m512i *in = (__m512i*)rowIn;
__m512i *out = (__m512i*)rowOut + ( (nCols-1) * BLOCK_LEN_M256I );
state0 = _mm512_load_si512( (__m512i*)State );
state1 = _mm512_load_si512( (__m512i*)State + 1 );
@@ -171,17 +180,15 @@ inline void reducedDuplexRow1_2way( uint64_t *State, uint64_t *rowIn,
out[2] = _mm512_xor_si512( state2, in[2] );
//Input: next column (i.e., next block in sequence)
in0 += BLOCK_LEN_M256I;
in1 += BLOCK_LEN_M256I;
in += BLOCK_LEN_M256I;
//Output: goes to previous column
out -= BLOCK_LEN_M256I * 2;
out -= BLOCK_LEN_M256I;
}
_mm512_store_si256( (__m512i*)State, state0 );
_mm512_store_si256( (__m512i*)State + 1, state1 );
_mm512_store_si256( (__m512i*)State + 2, state2 );
_mm512_store_si256( (__m512i*)State + 3, state3 );
}
_mm512_store_si512( (__m512i*)State, state0 );
_mm512_store_si512( (__m512i*)State + 1, state1 );
_mm512_store_si512( (__m512i*)State + 2, state2 );
_mm512_store_si512( (__m512i*)State + 3, state3 );
}
inline void reducedDuplexRowSetup_2way( uint64_t *State, uint64_t *rowIn,
@@ -192,7 +199,7 @@ inline void reducedDuplexRowSetup_2way( uint64_t *State, uint64_t *rowIn,
register __m512i state0, state1, state2, state3;
__m512i* in = (__m512i*)rowIn;
__m512i* inout = (__m512i*)rowInOut;
__m512i* out = (__m512i*)rowOut + ( (nCols-1) * BLOCK_LEN_M256I * 2 );
__m512i* out = (__m512i*)rowOut + ( (nCols-1) * BLOCK_LEN_M256I );
__m512i t0, t1, t2;
state0 = _mm512_load_si512( (__m512i*)State );
@@ -209,7 +216,7 @@ inline void reducedDuplexRowSetup_2way( uint64_t *State, uint64_t *rowIn,
state2 = _mm512_xor_si512( state2,
_mm512_add_epi64( in[2], inout[2] ) );
LYRA_ROUND_2WAY AVX512( state0, state1, state2, state3 );
LYRA_ROUND_2WAY_AVX512( state0, state1, state2, state3 );
out[0] = _mm512_xor_si512( state0, in[0] );
out[1] = _mm512_xor_si512( state1, in[1] );
@@ -221,17 +228,17 @@ inline void reducedDuplexRowSetup_2way( uint64_t *State, uint64_t *rowIn,
t2 = _mm512_permutex_epi64( state2, 0x93 );
inout[0] = _mm512_xor_si512( inout[0],
_mm512_mask_blend_epi32( t0, t2, 0x03 ) );
_mm512_mask_blend_epi32( 0x03, t0, t2 ) );
inout[1] = _mm512_xor_si512( inout[1],
_mm512_mask_blend_epi32( t1, t0, 0x03 ) );
_mm512_mask_blend_epi32( 0x03, t1, t0 ) );
inout[2] = _mm512_xor_si512( inout[2],
_mm512_mask_blend_epi32( t2, t1, 0x03 ) );
_mm512_mask_blend_epi32( 0x03, t2, t1 ) );
//Inputs: next column (i.e., next block in sequence)
in += BLOCK_LEN_M256I * 2;
inout += BLOCK_LEN_M256I * 2;
in += BLOCK_LEN_M256I;
inout += BLOCK_LEN_M256I;
//Output: goes to previous column
out -= BLOCK_LEN_M256I * 2;
out -= BLOCK_LEN_M256I;
}
_mm512_store_si512( (__m512i*)State, state0 );
@@ -240,53 +247,99 @@ inline void reducedDuplexRowSetup_2way( uint64_t *State, uint64_t *rowIn,
_mm512_store_si512( (__m512i*)State + 3, state3 );
}
inline void reducedDuplexRow_2way( uint64_t *State, uint64_t *rowIn1,
uint64_t *rowIn0, uint64_t *rowInOut, uint64_t *rowOut,
uint64_t nCols )
// big ugly workaound for pointer aliasing, use a union of pointers.
// Access matrix using m512i for in and out, m256i for inout
inline void reducedDuplexRow_2way( uint64_t *State, povly matrix,
uint64_t rowIn,
uint64_t rowInOut0, uint64_t rowInOut1,
uint64_t rowOut, uint64_t nCols )
{
int i;
register __m512i state0, state1, state2, state3;
__m256i *in0 = (__m256i*)rowIn0;
__m256i *in0 = (__m256i*)rowIn0;
__m2512* in = (__m512i*)rowIn;
__m2512* inout = (__m512i*)rowInOut;
__m512i* out = (__m512i*)rowOut;
__m512i t0, t1, t2;
const uint64_t ROW_LEN_M256I = BLOCK_LEN_INT64 * nCols / 4;
__m512i state0, state1, state2, state3;
// register __m512i state0, state1, state2, state3;
__m512i *in = &matrix.v512[ rowIn * ROW_LEN_M256I ];
__m256i *inout0 = &matrix.v256[ 2 * rowInOut0 * ROW_LEN_M256I ];
__m256i *inout1 = &matrix.v256[ 2 * rowInOut1 * ROW_LEN_M256I ];
__m512i *out = &matrix.v512[ rowOut * ROW_LEN_M256I ];
__m512i io[3];
povly inout;
inout.v512 = &io[0];
__m512i t0, t1, t2;
_mm_prefetch( in0, _MM_HINT_T0 );
_mm_prefetch( in1, _MM_HINT_T0 );
_mm_prefetch( in0 + 2, _MM_HINT_T0 );
_mm_prefetch( in1 + 2, _MM_HINT_T0 );
_mm_prefetch( in0 + 4, _MM_HINT_T0 );
_mm_prefetch( in1 + 4, _MM_HINT_T0 );
_mm_prefetch( in0 + 6, _MM_HINT_T0 );
_mm_prefetch( in1 + 6, _MM_HINT_T0 );
state0 = _mm512_load_si512( (__m512i*)State );
state1 = _mm512_load_si512( (__m512i*)State + 1 );
state2 = _mm512_load_si512( (__m512i*)State + 2 );
state3 = _mm512_load_si512( (__m512i*)State + 3 );
_mm_prefetch( in, _MM_HINT_T0 );
_mm_prefetch( inout0, _MM_HINT_T0 );
_mm_prefetch( inout1, _MM_HINT_T0 );
_mm_prefetch( in + 2, _MM_HINT_T0 );
_mm_prefetch( inout0 + 2, _MM_HINT_T0 );
_mm_prefetch( inout1 + 2, _MM_HINT_T0 );
_mm_prefetch( in + 4, _MM_HINT_T0 );
_mm_prefetch( inout0 + 4, _MM_HINT_T0 );
_mm_prefetch( inout1 + 4, _MM_HINT_T0 );
_mm_prefetch( in + 6, _MM_HINT_T0 );
_mm_prefetch( inout0 + 6, _MM_HINT_T0 );
_mm_prefetch( inout1 + 6, _MM_HINT_T0 );
//uint64_t *ii = (uint64_t*)in0;
//printf("RDRV0 IO %016lx %016lx %016lx %016lx\n",ii[0],ii[1],ii[2],ii[3]);
for ( i = 0; i < nCols; i++ )
{
/*
//printf("RDR: loop %d\n",i);
uint64_t *io1 = (uint64_t*)inout1;
printf("RDRV0 col= %d\n", i);
printf("RDRV0 IO1 %016lx %016lx %016lx %016lx\n",io1[0],io1[1],io1[2],io1[3]);
printf("RDRV0 IO1 %016lx %016lx %016lx %016lx\n",io1[4],io1[5],io1[6],io1[7]);
printf("RDRV0 IO1 %016lx %016lx %016lx %016lx\n",io1[8],io1[9],io1[10],io1[11]);
printf("RDRV0 IO1 %016lx %016lx %016lx %016lx\n",io1[12],io1[13],io1[14],io1[153]);
*/
//Absorbing "M[prev] [+] M[row*]"
inout.v256[0] = inout0[0];
inout.v256[1] = inout1[1];
inout.v256[2] = inout0[2];
inout.v256[3] = inout1[3];
inout.v256[4] = inout0[4];
inout.v256[5] = inout1[5];
/*
uint64_t *io = (uint64_t*)inout.u64;
uint64_t *ii = (uint64_t*)in;
printf("RDRV1 col= %d\n", i);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[0],io[1],io[2],io[3]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[4],io[5],io[6],io[7]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[8],io[9],io[10],io[11]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[12],io[13],io[14],io[15]);
printf("RDRV1 IN %016lx %016lx %016lx %016lx\n",ii[0],ii[1],ii[2],ii[3]);
printf("RDRV1 IN %016lx %016lx %016lx %016lx\n",ii[4],ii[5],ii[6],ii[7]);
printf("RDRV1 IN %016lx %016lx %016lx %016lx\n",ii[8],ii[9],ii[10],ii[11]);
printf("RDRV1 IN %016lx %016lx %016lx %016lx\n",ii[12],ii[13],ii[14],ii[15]);
*/
// state0 = _mm512_xor_si512( state0, mm512_concat_256( in1[0], in0[0] );
// state1 = _mm512_xor_si512( state1, mm512_concat_256( in1[1], in0[1] );
// state2 = _mm512_xor_si512( state2, mm512_concat_256( in1[2], in0[2] );
t0 = mm512_concat_256( in1[0], in0[0] );
t1 = mm512_concat_256( in1[1], in0[1] );
t2 = mm512_concat_256( in1[2], in0[2] );
state0 = _mm512_xor_si512( state0,
_mm512_add_epi64( t0, inout[0] ) );
_mm512_add_epi64( in[0], inout.v512[0] ) );
state1 = _mm512_xor_si512( state1,
_mm512_add_epi64( t1, inout[1] ) );
_mm512_add_epi64( in[1], inout.v512[1] ) );
state2 = _mm512_xor_si512( state2,
_mm512_add_epi64( t2, inout[2] ) );
_mm512_add_epi64( in[2], inout.v512[2] ) );
//printf("RDR: round\n");
//Applies the reduced-round transformation f to the sponge's state
LYRA_ROUND_2WAY_AVX512( state0, state1, state2, state3 );
//printf("RDR 3\n");
//M[rowOut][col] = M[rowOut][col] XOR rand
out[0] = _mm512_xor_si512( out[0], state0 );
out[1] = _mm512_xor_si512( out[1], state1 );
@@ -296,18 +349,76 @@ inline void reducedDuplexRow_2way( uint64_t *State, uint64_t *rowIn1,
t0 = _mm512_permutex_epi64( state0, 0x93 );
t1 = _mm512_permutex_epi64( state1, 0x93 );
t2 = _mm512_permutex_epi64( state2, 0x93 );
/*
uint64_t *st = (uint64_t*)&state0;
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
st = (uint64_t*)&state1;
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
st = (uint64_t*)&state2;
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
inout[0] = _mm512_xor_si512( inout[0],
_mm512_mask_blend_epi32( t0, t2, 0x03 ) );
inout[1] = _mm512_xor_si512( inout[1],
_mm512_mask_blend_epi32( t1, t0, 0x03 ) );
inout[2] = _mm512_xor_si512( inout[2],
_mm512_mask_blend_epi32( t2, t1, 0x03 ) );
st = (uint64_t*)&t0;
printf("RDRV2 t0 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 t0 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
st = (uint64_t*)&t1;
printf("RDRV2 t1 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 t1 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
st = (uint64_t*)&t2;
printf("RDRV2 t2 %016lx %016lx %016lx %016lx\n",st[0],st[1],st[2],st[3]);
printf("RDRv2 t2 %016lx %016lx %016lx %016lx\n",st[4],st[5],st[6],st[7]);
*/
/*
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[8],st[9],st[10],st[11]);
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[12],st[13],st[14],st[15]);
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[16],st[17],st[18],st[19]);
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[20],st[21],st[22],st[23]);
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[24],st[25],st[26],st[271]);
printf("RDRV2 %016lx %016lx %016lx %016lx\n",st[28],st[29],st[30],st[31]);
*/
//printf("RDR 4\n");
/*
//uint64_t *io = (uint64_t*)&inout;
printf("RDRV1 col= %d\n", i);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[0],io[1],io[2],io[3]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[4],io[5],io[6],io[7]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[8],io[9],io[10],io[11]);
printf("RDRV1 IO %016lx %016lx %016lx %016lx\n",io[12],io[13],io[14],io[15]);
*/
// need to split inout for write
inout.v512[0] = _mm512_xor_si512( inout.v512[0],
_mm512_mask_blend_epi32( 0x03, t0, t2 ) );
inout.v512[1] = _mm512_xor_si512( inout.v512[1],
_mm512_mask_blend_epi32( 0x03, t1, t0 ) );
inout.v512[2] = _mm512_xor_si512( inout.v512[2],
_mm512_mask_blend_epi32( 0x03, t2, t1 ) );
/*
printf("RDRV3 IO %016lx %016lx %016lx %016lx\n",io[0],io[1],io[2],io[3]);
printf("RDRV3 IO %016lx %016lx %016lx %016lx\n",io[4],io[5],io[6],io[7]);
printf("RDRV3 IO %016lx %016lx %016lx %016lx\n",io[8],io[9],io[10],io[11]);
printf("RDRV3 IO %016lx %016lx %016lx %016lx\n",io[12],io[13],io[14],io[153]);
*/
inout0[0] = inout.v256[0];
inout1[1] = inout.v256[1];
inout0[2] = inout.v256[2];
inout1[3] = inout.v256[3];
inout0[4] = inout.v256[4];
inout1[5] = inout.v256[5];
//printf("RDR 5\n");
//Goes to next block
in += BLOCK_LEN_M256I * 2;
out += BLOCK_LEN_M256I * 2;
inout += BLOCK_LEN_M256I * 2;
in += BLOCK_LEN_M256I;
inout0 += BLOCK_LEN_M256I * 2;
inout1 += BLOCK_LEN_M256I * 2;
out += BLOCK_LEN_M256I;
}
_mm512_store_si512( (__m512i*)State, state0 );