This commit is contained in:
Jay D Dee
2020-10-02 10:48:37 -04:00
parent cdd587537e
commit c85fb3842b
44 changed files with 1351 additions and 747 deletions

View File

@@ -6,30 +6,6 @@
*/
#include "x16r-gate.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
#include "algo/echo/sph_echo.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include <openssl/sha.h>
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#endif
static void hex_getAlgoString(const uint32_t* prevblock, char *output)
{
char *sptr = output;
@@ -47,34 +23,6 @@ static void hex_getAlgoString(const uint32_t* prevblock, char *output)
*sptr = '\0';
}
/*
union _hex_context_overlay
{
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
#endif
sph_blake512_context blake;
sph_bmw512_context bmw;
sph_skein512_context skein;
sph_jh512_context jh;
sph_keccak512_context keccak;
hashState_luffa luffa;
cubehashParam cube;
shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;
};
typedef union _hex_context_overlay hex_context_overlay;
*/
static __thread x16r_context_overlay hex_ctx;
int hex_hash( void* output, const void* input, int thrid )
@@ -187,8 +135,12 @@ int hex_hash( void* output, const void* input, int thrid )
sph_hamsi512_close( &ctx.hamsi, hash );
break;
case FUGUE:
#if defined(__AES__)
fugue512_full( &ctx.fugue, hash, in, size );
#else
sph_fugue512_full( &ctx.fugue, hash, in, size );
break;
#endif
break;
case SHABAL:
if ( i == 0 )
sph_shabal512( &ctx.shabal, in+64, 16 );

View File

@@ -15,16 +15,17 @@
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include <openssl/sha.h>
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/fugue/fugue-aesni.h"
#else
#include "algo/echo/sph_echo.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/fugue/sph_fugue.h"
#endif
// Config
@@ -34,13 +35,16 @@ typedef struct TortureNode TortureNode;
typedef struct TortureGarden TortureGarden;
// Graph of hash algos plus SPH contexts
struct TortureGarden {
struct TortureGarden
{
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
hashState_fugue fugue;
#else
sph_echo512_context echo;
sph_groestl512_context groestl;
sph_echo512_context echo;
sph_groestl512_context groestl;
sph_fugue512_context fugue;
#endif
sph_blake512_context blake;
sph_bmw512_context bmw;
@@ -52,15 +56,13 @@ struct TortureGarden {
shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;
struct TortureNode {
unsigned int algo;
TortureNode *childLeft;
TortureNode *childRight;
TortureNode *child[2];
} nodes[22];
} __attribute__ ((aligned (64)));
@@ -97,10 +99,12 @@ static void get_hash( void *output, const void *input, TortureGarden *garden,
#endif
break;
case 4:
sph_fugue512_init(&garden->fugue);
sph_fugue512(&garden->fugue, input, 64);
sph_fugue512_close(&garden->fugue, hash);
break;
#if defined(__AES__)
fugue512_full( &garden->fugue, hash, input, 64 );
#else
sph_fugue512_full( &garden->fugue, hash, input, 64 );
#endif
break;
case 5:
#if defined(__AES__)
groestl512_full( &garden->groestl, (char*)hash, (char*)input, 512 );
@@ -162,68 +166,61 @@ static void get_hash( void *output, const void *input, TortureGarden *garden,
break;
}
// Output the hash
memcpy(output, hash, 64);
}
// Recursively traverse a given torture garden starting with a given hash and given node within the garden. The hash is overwritten with the final hash.
static void traverse_garden( TortureGarden *garden, void *hash,
TortureNode *node )
{
unsigned char partialHash[64] __attribute__ ((aligned (64)));
get_hash(partialHash, hash, garden, node->algo);
if ( partialHash[63] % 2 == 0 )
{ // Last byte of output hash is even
if ( node->childLeft != NULL )
traverse_garden( garden, partialHash, node->childLeft );
}
else
{ // Last byte of output hash is odd
if ( node->childRight != NULL )
traverse_garden( garden, partialHash, node->childRight );
}
memcpy( hash, partialHash, 64 );
}
// Associate child nodes with a parent node
static inline void link_nodes( TortureNode *parent, TortureNode *childLeft,
TortureNode *childRight )
{
parent->childLeft = childLeft;
parent->childRight = childRight;
}
static __thread TortureGarden garden;
bool initialize_torture_garden()
{
// Create torture garden nodes. Note that both sides of 19 and 20 lead to 21, and 21 has no children (to make traversal complete).
link_nodes(&garden.nodes[0], &garden.nodes[1], &garden.nodes[2]);
link_nodes(&garden.nodes[1], &garden.nodes[3], &garden.nodes[4]);
link_nodes(&garden.nodes[2], &garden.nodes[5], &garden.nodes[6]);
link_nodes(&garden.nodes[3], &garden.nodes[7], &garden.nodes[8]);
link_nodes(&garden.nodes[4], &garden.nodes[9], &garden.nodes[10]);
link_nodes(&garden.nodes[5], &garden.nodes[11], &garden.nodes[12]);
link_nodes(&garden.nodes[6], &garden.nodes[13], &garden.nodes[14]);
link_nodes(&garden.nodes[7], &garden.nodes[15], &garden.nodes[16]);
link_nodes(&garden.nodes[8], &garden.nodes[15], &garden.nodes[16]);
link_nodes(&garden.nodes[9], &garden.nodes[15], &garden.nodes[16]);
link_nodes(&garden.nodes[10], &garden.nodes[15], &garden.nodes[16]);
link_nodes(&garden.nodes[11], &garden.nodes[17], &garden.nodes[18]);
link_nodes(&garden.nodes[12], &garden.nodes[17], &garden.nodes[18]);
link_nodes(&garden.nodes[13], &garden.nodes[17], &garden.nodes[18]);
link_nodes(&garden.nodes[14], &garden.nodes[17], &garden.nodes[18]);
link_nodes(&garden.nodes[15], &garden.nodes[19], &garden.nodes[20]);
link_nodes(&garden.nodes[16], &garden.nodes[19], &garden.nodes[20]);
link_nodes(&garden.nodes[17], &garden.nodes[19], &garden.nodes[20]);
link_nodes(&garden.nodes[18], &garden.nodes[19], &garden.nodes[20]);
link_nodes(&garden.nodes[19], &garden.nodes[21], &garden.nodes[21]);
link_nodes(&garden.nodes[20], &garden.nodes[21], &garden.nodes[21]);
garden.nodes[21].childLeft = NULL;
garden.nodes[21].childRight = NULL;
return true;
garden.nodes[ 0].child[0] = &garden.nodes[ 1];
garden.nodes[ 0].child[1] = &garden.nodes[ 2];
garden.nodes[ 1].child[0] = &garden.nodes[ 3];
garden.nodes[ 1].child[1] = &garden.nodes[ 4];
garden.nodes[ 2].child[0] = &garden.nodes[ 5];
garden.nodes[ 2].child[1] = &garden.nodes[ 6];
garden.nodes[ 3].child[0] = &garden.nodes[ 7];
garden.nodes[ 3].child[1] = &garden.nodes[ 8];
garden.nodes[ 4].child[0] = &garden.nodes[ 9];
garden.nodes[ 4].child[1] = &garden.nodes[10];
garden.nodes[ 5].child[0] = &garden.nodes[11];
garden.nodes[ 5].child[1] = &garden.nodes[12];
garden.nodes[ 6].child[0] = &garden.nodes[13];
garden.nodes[ 6].child[1] = &garden.nodes[14];
garden.nodes[ 7].child[0] = &garden.nodes[15];
garden.nodes[ 7].child[1] = &garden.nodes[16];
garden.nodes[ 8].child[0] = &garden.nodes[15];
garden.nodes[ 8].child[1] = &garden.nodes[16];
garden.nodes[ 9].child[0] = &garden.nodes[15];
garden.nodes[ 9].child[1] = &garden.nodes[16];
garden.nodes[10].child[0] = &garden.nodes[15];
garden.nodes[10].child[1] = &garden.nodes[16];
garden.nodes[11].child[0] = &garden.nodes[17];
garden.nodes[11].child[1] = &garden.nodes[18];
garden.nodes[12].child[0] = &garden.nodes[17];
garden.nodes[12].child[1] = &garden.nodes[18];
garden.nodes[13].child[0] = &garden.nodes[17];
garden.nodes[13].child[1] = &garden.nodes[18];
garden.nodes[14].child[0] = &garden.nodes[17];
garden.nodes[14].child[1] = &garden.nodes[18];
garden.nodes[15].child[0] = &garden.nodes[19];
garden.nodes[15].child[1] = &garden.nodes[20];
garden.nodes[16].child[0] = &garden.nodes[19];
garden.nodes[16].child[1] = &garden.nodes[20];
garden.nodes[17].child[0] = &garden.nodes[19];
garden.nodes[17].child[1] = &garden.nodes[20];
garden.nodes[18].child[0] = &garden.nodes[19];
garden.nodes[18].child[1] = &garden.nodes[20];
garden.nodes[19].child[0] = &garden.nodes[21];
garden.nodes[19].child[1] = &garden.nodes[21];
garden.nodes[20].child[0] = &garden.nodes[21];
garden.nodes[20].child[1] = &garden.nodes[21];
garden.nodes[21].child[0] = NULL;
garden.nodes[21].child[1] = NULL;
return true;
}
// Produce a 32-byte hash from 80-byte input data
@@ -236,20 +233,67 @@ int minotaur_hash( void *output, const void *input, int thr_id )
SHA512_Update( &garden.sha512, input, 80 );
SHA512_Final( (unsigned char*) hash, &garden.sha512 );
// algo 6 (Hamsi) is very slow. It's faster to skip hashing this nonce
// if Hamsi is needed but only the first and last functions are
// currently known. Abort if either is Hamsi.
if ( ( ( hash[ 0] % MINOTAUR_ALGO_COUNT ) == 6 )
|| ( ( hash[21] % MINOTAUR_ALGO_COUNT ) == 6 ) )
return 0;
// Assign algos to torture garden nodes based on initial hash
for ( int i = 0; i < 22; i++ )
garden.nodes[i].algo = hash[i] % MINOTAUR_ALGO_COUNT;
// Send the initial hash through the torture garden
traverse_garden( &garden, hash, &garden.nodes[0] );
TortureNode *node = &garden.nodes[0];
while ( node )
{
get_hash( hash, hash, &garden, node->algo );
node = node->child[ hash[63] & 1 ];
}
memcpy( output, hash, 32 );
return 1;
}
int scanhash_minotaur( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t edata[20] __attribute__((aligned(64)));
uint32_t hash[8] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
const uint32_t last_nonce = max_nonce - 1;
uint32_t n = first_nonce;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
uint64_t skipped = 0;
mm128_bswap32_80( edata, pdata );
do
{
edata[19] = n;
if ( likely( algo_gate.hash( hash, edata, thr_id ) ) )
{
if ( unlikely( valid_hash( hash, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n );
submit_solution( work, hash, mythr );
}
}
else skipped++;
n++;
} while ( n < last_nonce && !work_restart[thr_id].restart );
*hashes_done = n - first_nonce - skipped;
pdata[19] = n;
return 0;
}
bool register_minotaur_algo( algo_gate_t* gate )
{
gate->scanhash = (void*)&scanhash_minotaur;
gate->hash = (void*)&minotaur_hash;
gate->optimizations = SSE2_OPT | AES_OPT | AVX2_OPT | AVX512_OPT;
gate->miner_thread_init = (void*)&initialize_torture_garden;

View File

@@ -347,14 +347,14 @@ int x16r_8way_hash_generic( void* output, const void* input, int thrid )
hash7, vhash );
break;
case FUGUE:
sph_fugue512_full( &ctx.fugue, hash0, in0, size );
sph_fugue512_full( &ctx.fugue, hash1, in1, size );
sph_fugue512_full( &ctx.fugue, hash2, in2, size );
sph_fugue512_full( &ctx.fugue, hash3, in3, size );
sph_fugue512_full( &ctx.fugue, hash4, in4, size );
sph_fugue512_full( &ctx.fugue, hash5, in5, size );
sph_fugue512_full( &ctx.fugue, hash6, in6, size );
sph_fugue512_full( &ctx.fugue, hash7, in7, size );
fugue512_full( &ctx.fugue, hash0, in0, size );
fugue512_full( &ctx.fugue, hash1, in1, size );
fugue512_full( &ctx.fugue, hash2, in2, size );
fugue512_full( &ctx.fugue, hash3, in3, size );
fugue512_full( &ctx.fugue, hash4, in4, size );
fugue512_full( &ctx.fugue, hash5, in5, size );
fugue512_full( &ctx.fugue, hash6, in6, size );
fugue512_full( &ctx.fugue, hash7, in7, size );
break;
case SHABAL:
intrlv_8x32( vhash, in0, in1, in2, in3, in4, in5, in6, in7,
@@ -747,10 +747,10 @@ int x16r_4way_hash_generic( void* output, const void* input, int thrid )
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
break;
case FUGUE:
sph_fugue512_full( &ctx.fugue, hash0, in0, size );
sph_fugue512_full( &ctx.fugue, hash1, in1, size );
sph_fugue512_full( &ctx.fugue, hash2, in2, size );
sph_fugue512_full( &ctx.fugue, hash3, in3, size );
fugue512_full( &ctx.fugue, hash0, in0, size );
fugue512_full( &ctx.fugue, hash1, in1, size );
fugue512_full( &ctx.fugue, hash2, in2, size );
fugue512_full( &ctx.fugue, hash3, in3, size );
break;
case SHABAL:
intrlv_4x32( vhash, in0, in1, in2, in3, size<<3 );

View File

@@ -24,6 +24,7 @@
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/fugue/fugue-aesni.h"
#endif
#if defined (__AVX2__)
#include "algo/blake/blake-hash-4way.h"
@@ -111,7 +112,7 @@ union _x16r_8way_context_overlay
cubehashParam cube;
simd_4way_context simd;
hamsi512_8way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_8way_context shabal;
sph_whirlpool_context whirlpool;
sha512_8way_context sha512;
@@ -155,7 +156,7 @@ union _x16r_4way_context_overlay
shavite512_context shavite;
simd_2way_context simd;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_4way_context shabal;
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
@@ -180,9 +181,11 @@ union _x16r_context_overlay
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
hashState_fugue fugue;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
sph_fugue512_context fugue;
#endif
sph_blake512_context blake;
sph_bmw512_context bmw;
@@ -194,7 +197,6 @@ union _x16r_context_overlay
shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;

View File

@@ -151,8 +151,12 @@ int x16r_hash_generic( void* output, const void* input, int thrid )
sph_hamsi512_close( &ctx.hamsi, hash );
break;
case FUGUE:
sph_fugue512_full( &ctx.fugue, hash, in, size );
break;
#if defined(__AES__)
fugue512_full( &ctx.fugue, hash, in, size );
#else
sph_fugue512_full( &ctx.fugue, hash, in, size );
#endif
break;
case SHABAL:
if ( i == 0 )
sph_shabal512( &ctx.shabal, in+64, 16 );

View File

@@ -8,30 +8,7 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "algo/blake/blake-hash-4way.h"
#include "algo/bmw/bmw-hash-4way.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#include "algo/skein/skein-hash-4way.h"
#include "algo/jh/jh-hash-4way.h"
#include "algo/keccak/keccak-hash-4way.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/luffa/luffa-hash-2way.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/cubehash/cube-hash-2way.h"
#include "algo/simd/simd-hash-2way.h"
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/hamsi/hamsi-hash-4way.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/shabal-hash-4way.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include "algo/sha/sha-hash-4way.h"
#include "algo/tiger/sph_tiger.h"
#if defined(__VAES__)
#include "algo/groestl/groestl512-hash-4way.h"
#include "algo/shavite/shavite-hash-4way.h"
#include "algo/echo/echo-hash-4way.h"
#endif
#if defined (X16RV2_8WAY)
@@ -46,7 +23,7 @@ union _x16rv2_8way_context_overlay
cubehashParam cube;
simd_4way_context simd;
hamsi512_8way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_8way_context shabal;
sph_whirlpool_context whirlpool;
sha512_8way_context sha512;
@@ -432,14 +409,14 @@ int x16rv2_8way_hash( void* output, const void* input, int thrid )
hash7, vhash );
break;
case FUGUE:
sph_fugue512_full( &ctx.fugue, hash0, in0, size );
sph_fugue512_full( &ctx.fugue, hash1, in1, size );
sph_fugue512_full( &ctx.fugue, hash2, in2, size );
sph_fugue512_full( &ctx.fugue, hash3, in3, size );
sph_fugue512_full( &ctx.fugue, hash4, in4, size );
sph_fugue512_full( &ctx.fugue, hash5, in5, size );
sph_fugue512_full( &ctx.fugue, hash6, in6, size );
sph_fugue512_full( &ctx.fugue, hash7, in7, size );
fugue512_full( &ctx.fugue, hash0, in0, size );
fugue512_full( &ctx.fugue, hash1, in1, size );
fugue512_full( &ctx.fugue, hash2, in2, size );
fugue512_full( &ctx.fugue, hash3, in3, size );
fugue512_full( &ctx.fugue, hash4, in4, size );
fugue512_full( &ctx.fugue, hash5, in5, size );
fugue512_full( &ctx.fugue, hash6, in6, size );
fugue512_full( &ctx.fugue, hash7, in7, size );
break;
case SHABAL:
intrlv_8x32( vhash, in0, in1, in2, in3, in4, in5, in6, in7,
@@ -705,7 +682,7 @@ union _x16rv2_4way_context_overlay
shavite512_context shavite;
simd_2way_context simd;
hamsi512_4way_context hamsi;
sph_fugue512_context fugue;
hashState_fugue fugue;
shabal512_4way_context shabal;
sph_whirlpool_context whirlpool;
sha512_4way_context sha512;
@@ -946,10 +923,10 @@ int x16rv2_4way_hash( void* output, const void* input, int thrid )
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
break;
case FUGUE:
sph_fugue512_full( &ctx.fugue, hash0, in0, size );
sph_fugue512_full( &ctx.fugue, hash1, in1, size );
sph_fugue512_full( &ctx.fugue, hash2, in2, size );
sph_fugue512_full( &ctx.fugue, hash3, in3, size );
fugue512_full( &ctx.fugue, hash0, in0, size );
fugue512_full( &ctx.fugue, hash1, in1, size );
fugue512_full( &ctx.fugue, hash2, in2, size );
fugue512_full( &ctx.fugue, hash3, in3, size );
break;
case SHABAL:
intrlv_4x32( vhash, in0, in1, in2, in3, size<<3 );

View File

@@ -8,41 +8,18 @@
#if !defined(X16R_8WAY) && !defined(X16R_4WAY)
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#include "algo/shavite/sph_shavite.h"
#include "algo/luffa/luffa_for_sse2.h"
#include "algo/cubehash/cubehash_sse2.h"
#include "algo/simd/nist.h"
#include "algo/echo/sph_echo.h"
#include "algo/hamsi/sph_hamsi.h"
#include "algo/fugue/sph_fugue.h"
#include "algo/shabal/sph_shabal.h"
#include "algo/whirlpool/sph_whirlpool.h"
#include <openssl/sha.h>
#include "algo/tiger/sph_tiger.h"
#if defined(__AES__)
#include "algo/echo/aes_ni/hash_api.h"
#include "algo/groestl/aes_ni/hash-groestl.h"
#endif
static __thread uint32_t s_ntime = UINT32_MAX;
union _x16rv2_context_overlay
{
#if defined(__AES__)
hashState_echo echo;
hashState_groestl groestl;
hashState_fugue fugue;
#else
sph_groestl512_context groestl;
sph_echo512_context echo;
sph_fugue512_context fugue;
#endif
sph_blake512_context blake;
sph_bmw512_context bmw;
@@ -54,7 +31,6 @@ union _x16rv2_context_overlay
shavite512_context shavite;
hashState_sd simd;
sph_hamsi512_context hamsi;
sph_fugue512_context fugue;
sph_shabal512_context shabal;
sph_whirlpool_context whirlpool;
SHA512_CTX sha512;
@@ -160,8 +136,12 @@ int x16rv2_hash( void* output, const void* input, int thrid )
sph_hamsi512_close( &ctx.hamsi, hash );
break;
case FUGUE:
#if defined(__AES__)
fugue512_full( &ctx.fugue, hash, in, size );
#else
sph_fugue512_full( &ctx.fugue, hash, in, size );
break;
#endif
break;
case SHABAL:
sph_shabal512_init( &ctx.shabal );
sph_shabal512( &ctx.shabal, in, size );