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

v23.15

Browse Source
This commit is contained in:
Jay D Dee
2023-11-30 14:36:47 -05:00
parent 4e3f1b926f
commit 9d3a46c355
29 changed files with 3081 additions and 2234 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
algo/quark/hmq1725.c
View File

@@ -6,7 +6,7 @@
#include <stdint.h>
#include "algo/blake/blake512-hash.h"
#include "algo/bmw/sph_bmw.h"
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
#include "algo/fugue/fugue-aesni.h"
#else
#include "algo/fugue/sph_fugue.h"
@@ -35,7 +35,7 @@ union _hmq1725_ctx_holder
{
blake512_context blake;
sph_bmw512_context bmw;
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
hashState_fugue fugue;
#else
sph_fugue512_context fugue;
@@ -177,7 +177,7 @@ extern void hmq1725hash(void *state, const void *input)
sph_hamsi512( &ctx.hamsi, hashA, 64 ); //3
sph_hamsi512_close( &ctx.hamsi, hashB ); //4
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
fugue512_full( &ctx.fugue, hashA, hashB, 64 );
#else
sph_fugue512_init( &ctx.fugue );
@@ -208,7 +208,7 @@ extern void hmq1725hash(void *state, const void *input)
if ( hashB[0] & mask ) //7
{
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
fugue512_full( &ctx.fugue, hashA, hashB, 64 );
#else
sph_fugue512_init( &ctx.fugue );
@@ -259,30 +259,18 @@ extern void hmq1725hash(void *state, const void *input)
int scanhash_hmq1725( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
// uint32_t endiandata[32] __attribute__((aligned(64)));
uint32_t endiandata[20] __attribute__((aligned(32)));
uint32_t hash64[8] __attribute__((aligned(32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t endiandata[20] __attribute__((aligned(32)));
uint32_t hash64[8] __attribute__((aligned(32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
int thr_id = mythr->id; // thr_id arg is deprecated
//const uint32_t Htarg = ptarget[7];
//we need bigendian data...
// for (int k = 0; k < 32; k++)
for (int k = 0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
for (int k = 0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
// hmq_bmw512_midstate( endiandata );
// if (opt_debug)
// {
// applog(LOG_DEBUG, "Thr: %02d, firstN: %08x, maxN: %08x, ToDo: %d", thr_id, first_nonce, max_nonce, max_nonce-first_nonce);
// }
/* I'm to lazy to put the loop in an inline function... so dirty copy'n'paste.... */
/* i know that i could set a variable, but i don't know how the compiler will optimize it, not that then the cpu needs to load the value *everytime* in a register */
if (ptarget[7]==0) {
do {
pdata[19] = ++n;

3
algo/quark/quark-gate.c
View File

@@ -14,7 +14,8 @@ bool register_quark_algo( algo_gate_t* gate )
gate->scanhash = (void*)&scanhash_quark;
gate->hash = (void*)&quark_hash;
#endif
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT | VAES_OPT
| NEON_OPT;
return true;
};

22
algo/quark/quark.c
View File

@@ -7,12 +7,12 @@
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "algo/blake/sph_blake.h"
#include "algo/blake/blake512-hash.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
#include "algo/groestl/aes_ni/hash-groestl.h"
#else
#include "algo/groestl/sph_groestl.h"
@@ -21,9 +21,9 @@
void quark_hash(void *state, const void *input)
{
uint32_t hash[16] __attribute__((aligned(64)));
sph_blake512_context ctx_blake;
blake512_context ctx_blake;
sph_bmw512_context ctx_bmw;
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
hashState_groestl ctx_groestl;
#else
sph_groestl512_context ctx_groestl;
@@ -33,17 +33,15 @@ void quark_hash(void *state, const void *input)
sph_keccak512_context ctx_keccak;
uint32_t mask = 8;
sph_blake512_init( &ctx_blake );
sph_blake512( &ctx_blake, input, 80 );
sph_blake512_close( &ctx_blake, hash );
blake512_full( &ctx_blake, hash, input, 80 );
sph_bmw512_init( &ctx_bmw );
sph_bmw512( &ctx_bmw, hash, 64 );
sph_bmw512_close( &ctx_bmw, hash );
if ( hash[0] & mask )
{
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
init_groestl( &ctx_groestl, 64 );
update_and_final_groestl( &ctx_groestl, (char*)hash,
(const char*)hash, 512 );
@@ -60,7 +58,7 @@ void quark_hash(void *state, const void *input)
sph_skein512_close( &ctx_skein, hash );
}
#if defined(__AES__)
#if defined(__AES__) || defined(__ARM_FEATURE_AES)
init_groestl( &ctx_groestl, 64 );
update_and_final_groestl( &ctx_groestl, (char*)hash,
(const char*)hash, 512 );
@@ -76,9 +74,7 @@ void quark_hash(void *state, const void *input)
if ( hash[0] & mask )
{
sph_blake512_init( &ctx_blake );
sph_blake512( &ctx_blake, hash, 64 );
sph_blake512_close( &ctx_blake, hash );
blake512_full( &ctx_blake, hash, hash, 64 );
}
else
{