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

v3.8.4

Browse Source
This commit is contained in:
Jay D Dee
2018-03-18 12:51:03 -04:00
parent 157508bd07
commit 20fe05054c
19 changed files with 830 additions and 289 deletions
Download Patch File Download Diff File Expand all files Collapse all files

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

@@ -76,7 +76,6 @@ char* hodl_malloc_txs_request( struct work *work )
return req;
}
void hodl_build_block_header( struct work* g_work, uint32_t version,
uint32_t *prevhash, uint32_t *merkle_tree,
uint32_t ntime, uint32_t nbits )
@@ -88,16 +87,16 @@ void hodl_build_block_header( struct work* g_work, uint32_t version,
if ( have_stratum )
for ( i = 0; i < 8; i++ )
g_work->data[1 + i] = le32dec( prevhash + i );
g_work->data[ 1+i ] = le32dec( prevhash + i );
else
for (i = 0; i < 8; i++)
g_work->data[ 8-i ] = le32dec( prevhash + i );
for ( i = 0; i < 8; i++ )
g_work->data[9 + i] = be32dec( merkle_tree + i );
g_work->data[ 9+i ] = be32dec( merkle_tree + i );
g_work->data[ algo_gate.ntime_index ] = ntime;
g_work->data[ algo_gate.nbits_index ] = nbits;
g_work->data[ algo_gate.ntime_index ] = ntime;
g_work->data[ algo_gate.nbits_index ] = nbits;
g_work->data[22] = 0x80000000;
g_work->data[31] = 0x00000280;
}
@@ -194,8 +193,13 @@ bool register_hodl_algo( algo_gate_t* gate )
applog( LOG_ERR, "Only CPUs with AES are supported, use legacy version.");
return false;
#endif
// if ( TOTAL_CHUNKS % opt_n_threads )
// {
// applog(LOG_ERR,"Thread count must be power of 2.");
// return false;
// }
pthread_barrier_init( &hodl_barrier, NULL, opt_n_threads );
gate->optimizations = SSE2_OPT | AES_OPT | AVX_OPT | AVX2_OPT;
gate->optimizations = AES_OPT | AVX_OPT | AVX2_OPT;
gate->scanhash = (void*)&hodl_scanhash;
gate->get_new_work = (void*)&hodl_get_new_work;
gate->longpoll_rpc_call = (void*)&hodl_longpoll_rpc_call;

31
algo/hodl/hodl-wolf.c
View File

@@ -10,23 +10,26 @@
#ifndef NO_AES_NI
void GenerateGarbageCore(CacheEntry *Garbage, int ThreadID, int ThreadCount, void *MidHash)
void GenerateGarbageCore( CacheEntry *Garbage, int ThreadID, int ThreadCount,
void *MidHash )
{
#ifdef __AVX__
uint64_t* TempBufs[SHA512_PARALLEL_N] ;
uint64_t* desination[SHA512_PARALLEL_N];
const int Chunk = TOTAL_CHUNKS / ThreadCount;
const uint32_t StartChunk = ThreadID * Chunk;
const uint32_t EndChunk = StartChunk + Chunk;
for ( int i=0; i<SHA512_PARALLEL_N; ++i )
#ifdef __AVX__
uint64_t* TempBufs[ SHA512_PARALLEL_N ] ;
uint64_t* desination[ SHA512_PARALLEL_N ];
for ( int i=0; i < SHA512_PARALLEL_N; ++i )
{
TempBufs[i] = (uint64_t*)malloc(32);
memcpy(TempBufs[i], MidHash, 32);
TempBufs[i] = (uint64_t*)malloc( 32 );
memcpy( TempBufs[i], MidHash, 32 );
}
uint32_t StartChunk = ThreadID * (TOTAL_CHUNKS / ThreadCount);
for ( uint32_t i = StartChunk;
i < StartChunk + (TOTAL_CHUNKS / ThreadCount); i+= SHA512_PARALLEL_N )
for ( uint32_t i = StartChunk; i < EndChunk; i += SHA512_PARALLEL_N )
{
for ( int j=0; j<SHA512_PARALLEL_N; ++j )
for ( int j = 0; j < SHA512_PARALLEL_N; ++j )
{
( (uint32_t*)TempBufs[j] )[0] = i + j;
desination[j] = (uint64_t*)( (uint8_t *)Garbage + ( (i+j)
@@ -35,15 +38,13 @@ void GenerateGarbageCore(CacheEntry *Garbage, int ThreadID, int ThreadCount, voi
sha512Compute32b_parallel( TempBufs, desination );
}
for ( int i=0; i<SHA512_PARALLEL_N; ++i )
for ( int i = 0; i < SHA512_PARALLEL_N; ++i )
free( TempBufs[i] );
#else
uint32_t TempBuf[8];
memcpy( TempBuf, MidHash, 32 );
uint32_t StartChunk = ThreadID * (TOTAL_CHUNKS / ThreadCount);
for ( uint32_t i = StartChunk;
i < StartChunk + (TOTAL_CHUNKS / ThreadCount); ++i )
for ( uint32_t i = StartChunk; i < EndChunk; ++i )
{
TempBuf[0] = i;
SHA512( ( uint8_t *)TempBuf, 32,