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

v3.9.2.5

Browse Source
This commit is contained in:
Jay D Dee
2019-06-13 11:20:27 -04:00
parent 7fec680835
commit b2331375a3
70 changed files with 4413 additions and 4360 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
algo/x17/hmq1725.c
View File

@@ -299,7 +299,7 @@ extern void hmq1725hash(void *state, const void *input)
}
int scanhash_hmq1725( int thr_id, struct work *work, int32_t max_nonce,
uint64_t *hashes_done )
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t endiandata[32] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(64)));
@@ -307,6 +307,7 @@ int scanhash_hmq1725( int thr_id, struct work *work, int32_t max_nonce,
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...

182
algo/x17/sonoa-4way.c
View File

@@ -47,30 +47,6 @@ union _sonoa_4way_context_overlay
};
typedef union _sonoa_4way_context_overlay sonoa_4way_context_overlay;
/*
sonoa_4way_ctx_holder sonoa_4way_ctx __attribute__ ((aligned (64)));
void init_sonoa_4way_ctx()
{
blake512_4way_init( &sonoa_4way_ctx.blake );
bmw512_4way_init( &sonoa_4way_ctx.bmw );
init_groestl( &sonoa_4way_ctx.groestl, 64 );
skein512_4way_init( &sonoa_4way_ctx.skein );
jh512_4way_init( &sonoa_4way_ctx.jh );
keccak512_4way_init( &sonoa_4way_ctx.keccak );
luffa_2way_init( &sonoa_4way_ctx.luffa, 512 );
cube_2way_init( &sonoa_4way_ctx.cube, 512, 16, 32 );
shavite512_2way_init( &sonoa_4way_ctx.shavite );
simd_2way_init( &sonoa_4way_ctx.simd, 512 );
init_echo( &sonoa_4way_ctx.echo, 512 );
hamsi512_4way_init( &sonoa_4way_ctx.hamsi );
sph_fugue512_init( &sonoa_4way_ctx.fugue );
shabal512_4way_init( &sonoa_4way_ctx.shabal );
sph_whirlpool_init( &sonoa_4way_ctx.whirlpool );
sha512_4way_init( &sonoa_4way_ctx.sha512 );
haval256_5_4way_init( &sonoa_4way_ctx.haval );
};
*/
void sonoa_4way_hash( void *state, const void *input )
{
@@ -82,8 +58,6 @@ void sonoa_4way_hash( void *state, const void *input )
uint64_t vhashA[8*4] __attribute__ ((aligned (64)));
uint64_t vhashB[8*4] __attribute__ ((aligned (64)));
sonoa_4way_context_overlay ctx;
// sonoa_4way_ctx_holder ctx __attribute__ ((aligned (64)));
// memcpy( &ctx, &sonoa_4way_ctx, sizeof(sonoa_4way_ctx) );
// 1
@@ -95,7 +69,7 @@ void sonoa_4way_hash( void *state, const void *input )
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -106,7 +80,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -120,7 +94,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -142,8 +116,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -160,13 +134,13 @@ void sonoa_4way_hash( void *state, const void *input )
// 2
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -177,7 +151,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -191,7 +165,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -213,8 +187,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -229,7 +203,7 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
@@ -241,7 +215,7 @@ void sonoa_4way_hash( void *state, const void *input )
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -252,7 +226,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -266,7 +240,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -288,8 +262,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -304,13 +278,13 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -326,13 +300,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_fugue512_close( &ctx.fugue, hash3 );
// 4
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -343,7 +317,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -357,7 +331,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -379,8 +353,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -395,13 +369,13 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -416,19 +390,19 @@ void sonoa_4way_hash( void *state, const void *input )
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm256_reinterleave_4x32_4x64( vhashB, vhash, 512 );
mm256_rintrlv_4x32_4x64( vhashB, vhash, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhashB, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -443,8 +417,8 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_2x128( vhashA, hash0, hash1, 512 );
mm256_interleave_2x128( vhashB, hash2, hash3, 512 );
mm256_intrlv_2x128( vhashA, hash0, hash1, 512 );
mm256_intrlv_2x128( vhashB, hash2, hash3, 512 );
shavite512_2way_init( &ctx.shavite );
shavite512_2way_update_close( &ctx.shavite, vhashA, vhashA, 64 );
@@ -452,19 +426,19 @@ void sonoa_4way_hash( void *state, const void *input )
shavite512_2way_update_close( &ctx.shavite, vhashB, vhashB, 64 );
// 5
mm256_reinterleave_2x128_4x64( vhash, vhashA, vhashB, 512 );
mm256_rintrlv_2x128_4x64( vhash, vhashA, vhashB, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_reinterleave_4x64_4x32( vhashB, vhash, 512 );
mm256_rintrlv_4x64_4x32( vhashB, vhash, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhashB, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -475,7 +449,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -489,7 +463,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -511,8 +485,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -527,13 +501,13 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -548,13 +522,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
@@ -571,13 +545,13 @@ void sonoa_4way_hash( void *state, const void *input )
// 6
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -588,7 +562,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -602,7 +576,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -624,8 +598,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -640,13 +614,13 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -661,13 +635,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
@@ -682,13 +656,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
@@ -705,13 +679,13 @@ void sonoa_4way_hash( void *state, const void *input )
// 7
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
bmw512_4way_init( &ctx.bmw );
bmw512_4way( &ctx.bmw, vhash, 64 );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -722,7 +696,7 @@ void sonoa_4way_hash( void *state, const void *input )
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, 64 );
@@ -736,7 +710,7 @@ void sonoa_4way_hash( void *state, const void *input )
keccak512_4way( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -758,8 +732,8 @@ void sonoa_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -774,13 +748,13 @@ void sonoa_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -795,13 +769,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_fugue512( &ctx.fugue, hash3, 64 );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, 64 );
@@ -816,13 +790,13 @@ void sonoa_4way_hash( void *state, const void *input )
sph_whirlpool( &ctx.whirlpool, hash3, 64 );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
mm256_reinterleave_4x64_4x32( vhashB, vhash, 512 );
mm256_rintrlv_4x64_4x32( vhashB, vhash, 512 );
haval256_5_4way_init( &ctx.haval );
haval256_5_4way( &ctx.haval, vhashB, 64 );
@@ -836,7 +810,6 @@ int scanhash_sonoa_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *hash7 = &(hash[7<<2]);
uint32_t lane_hash[8];
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
@@ -850,19 +823,13 @@ int scanhash_sonoa_4way( int thr_id, struct work *work, uint32_t max_nonce,
0xFFFFF000, 0xFFFF0000, 0 };
// Need big endian data
casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_bswap_intrlv80_4x64( vdata, pdata );
for ( int m=0; m < 6; m++ ) if ( Htarg <= htmax[m] )
{
uint32_t mask = masks[m];
do
{
*noncev = mm256_interleave_blend_32( mm256_bswap_32(
*noncev = mm256_intrlv_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ),
*noncev );
sonoa_4way_hash( hash, vdata );
@@ -871,17 +838,10 @@ int scanhash_sonoa_4way( int thr_id, struct work *work, uint32_t max_nonce,
if ( ( ( hash7[ lane ] & mask ) == 0 ) )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
{
pdata[19] = n + lane;
work_set_target_ratio( work, lane_hash );
if ( submit_work( mythr, work ) )
applog( LOG_NOTICE,
"Share %d submitted by thread %d, lane %d.",
accepted_share_count + rejected_share_count + 1,
thr_id, lane );
else
applog( LOG_WARNING, "Failed to submit share." );
submit_solution( work, lane_hash, mythr, lane );
}
}
n += 4;

76
algo/x17/x17-4way.c
View File

@@ -24,7 +24,6 @@
#include "algo/haval/haval-hash-4way.h"
#include "algo/sha/sha2-hash-4way.h"
//typedef struct {
union _x17_4way_context_overlay
{
blake512_4way_context blake;
@@ -47,30 +46,6 @@ union _x17_4way_context_overlay
};
typedef union _x17_4way_context_overlay x17_4way_context_overlay;
/*
x17_4way_ctx_holder x17_4way_ctx __attribute__ ((aligned (64)));
void init_x17_4way_ctx()
{
blake512_4way_init( &x17_4way_ctx.blake );
bmw512_4way_init( &x17_4way_ctx.bmw );
init_groestl( &x17_4way_ctx.groestl, 64 );
skein512_4way_init( &x17_4way_ctx.skein );
jh512_4way_init( &x17_4way_ctx.jh );
keccak512_4way_init( &x17_4way_ctx.keccak );
luffa_2way_init( &x17_4way_ctx.luffa, 512 );
cube_2way_init( &x17_4way_ctx.cube, 512, 16, 32 );
shavite512_2way_init( &x17_4way_ctx.shavite );
simd_2way_init( &x17_4way_ctx.simd, 512 );
init_echo( &x17_4way_ctx.echo, 512 );
hamsi512_4way_init( &x17_4way_ctx.hamsi );
sph_fugue512_init( &x17_4way_ctx.fugue );
shabal512_4way_init( &x17_4way_ctx.shabal );
sph_whirlpool_init( &x17_4way_ctx.whirlpool );
sha512_4way_init( &x17_4way_ctx.sha512 );
haval256_5_4way_init( &x17_4way_ctx.haval );
};
*/
void x17_4way_hash( void *state, const void *input )
{
uint64_t hash0[8] __attribute__ ((aligned (64)));
@@ -81,7 +56,6 @@ void x17_4way_hash( void *state, const void *input )
uint64_t vhashA[8*4] __attribute__ ((aligned (64)));
uint64_t vhashB[8*4] __attribute__ ((aligned (64)));
x17_4way_context_overlay ctx;
// memcpy( &ctx, &x17_4way_ctx, sizeof(x17_4way_ctx) );
// 1 Blake parallel 4 way 64 bit
blake512_4way_init( &ctx.blake );
@@ -94,7 +68,7 @@ void x17_4way_hash( void *state, const void *input )
bmw512_4way_close( &ctx.bmw, vhash );
// Serialize
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 3 Groestl
init_groestl( &ctx.groestl, 64 );
@@ -107,7 +81,7 @@ void x17_4way_hash( void *state, const void *input )
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// Parallellize
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein parallel 4 way 64 bit
skein512_4way_init( &ctx.skein );
@@ -125,7 +99,7 @@ void x17_4way_hash( void *state, const void *input )
keccak512_4way_close( &ctx.keccak, vhash );
// 7 Luffa parallel 2 way 128 bit
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, 512 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, 64 );
@@ -150,8 +124,8 @@ void x17_4way_hash( void *state, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhashA, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhashB, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, 512 );
// 11 Echo serial
init_echo( &ctx.echo, 512 );
@@ -168,13 +142,13 @@ void x17_4way_hash( void *state, const void *input )
(const BitSequence *) hash3, 512 );
// 12 Hamsi parallel 4 way 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 13 Fugue serial
sph_fugue512_init( &ctx.fugue );
@@ -191,13 +165,13 @@ void x17_4way_hash( void *state, const void *input )
sph_fugue512_close( &ctx.fugue, hash3 );
// 14 Shabal, parallel 4 way 32 bit
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, 64 );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, 512 );
// 15 Whirlpool serial
sph_whirlpool_init( &ctx.whirlpool );
@@ -214,19 +188,18 @@ void x17_4way_hash( void *state, const void *input )
sph_whirlpool_close( &ctx.whirlpool, hash3 );
// 16 SHA512 parallel 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, 64 );
sha512_4way_close( &ctx.sha512, vhash );
// 17 Haval parallel 32 bit
mm256_reinterleave_4x64_4x32( vhashB, vhash, 512 );
mm256_rintrlv_4x64_4x32( vhashB, vhash, 512 );
haval256_5_4way_init( &ctx.haval );
haval256_5_4way( &ctx.haval, vhashB, 64 );
haval256_5_4way_close( &ctx.haval, state );
}
int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
@@ -236,7 +209,6 @@ int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *hash7 = &(hash[7<<2]);
uint32_t lane_hash[8];
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
@@ -250,38 +222,24 @@ int scanhash_x17_4way( int thr_id, struct work *work, uint32_t max_nonce,
0xFFFFF000, 0xFFFF0000, 0 };
// Need big endian data
casti_m256i( endiandata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
casti_m256i( endiandata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
casti_m128i( endiandata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_bswap_intrlv80_4x64( vdata, pdata );
for ( int m = 0; m < 6; m++ ) if ( Htarg <= htmax[m] )
{
uint32_t mask = masks[ m ];
do
{
*noncev = mm256_interleave_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ),
*noncev );
*noncev = mm256_intrlv_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
x17_4way_hash( hash, vdata );
for ( int lane = 0; lane < 4; lane++ )
for ( int lane = 0; lane < 4; lane++ )
if ( ( hash7[ lane ] & mask ) == 0 )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
{
pdata[19] = n + lane;
work_set_target_ratio( work, lane_hash );
if ( submit_work( mythr, work ) )
applog( LOG_NOTICE,
"Share %d submitted by thread %d, lane %d.",
accepted_share_count + rejected_share_count + 1,
thr_id, lane );
else
applog( LOG_WARNING, "Failed to submit share." );
submit_solution( work, lane_hash, mythr, lane );
}
}
n += 4;

68
algo/x17/xevan-4way.c
View File

@@ -71,7 +71,7 @@ void xevan_4way_hash( void *output, const void *input )
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0,
@@ -87,7 +87,7 @@ void xevan_4way_hash( void *output, const void *input )
dataLen<<3 );
// Parallel 4way
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, dataLen );
@@ -101,7 +101,7 @@ void xevan_4way_hash( void *output, const void *input )
keccak512_4way( &ctx.keccak, vhash, dataLen );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, dataLen );
@@ -123,8 +123,8 @@ void xevan_4way_hash( void *output, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, dataLen<<3 );
mm256_deinterleave_1x128( hash0, hash1, vhashA, dataLen<<3 );
mm256_deinterleave_1x128( hash2, hash3, vhashB, dataLen<<3 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -139,13 +139,13 @@ void xevan_4way_hash( void *output, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, dataLen<<3 );
// Parallel
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, dataLen );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, dataLen );
@@ -161,13 +161,13 @@ void xevan_4way_hash( void *output, const void *input )
sph_fugue512_close( &ctx.fugue, hash3 );
// Parallel 4way 32 bit
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, dataLen );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
// Serial
sph_whirlpool_init( &ctx.whirlpool );
@@ -183,19 +183,19 @@ void xevan_4way_hash( void *output, const void *input )
sph_whirlpool( &ctx.whirlpool, hash3, dataLen );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, dataLen );
sha512_4way_close( &ctx.sha512, vhash );
mm256_reinterleave_4x64_4x32( vhashA, vhash, dataLen<<3 );
mm256_rintrlv_4x64_4x32( vhashA, vhash, dataLen<<3 );
haval256_5_4way_init( &ctx.haval );
haval256_5_4way( &ctx.haval, vhashA, dataLen );
haval256_5_4way_close( &ctx.haval, vhashA );
mm256_reinterleave_4x32_4x64( vhash, vhashA, dataLen<<3 );
mm256_rintrlv_4x32_4x64( vhash, vhashA, dataLen<<3 );
memset( &vhash[ 4<<2 ], 0, (dataLen-32) << 2 );
@@ -207,7 +207,7 @@ void xevan_4way_hash( void *output, const void *input )
bmw512_4way( &ctx.bmw, vhash, dataLen );
bmw512_4way_close( &ctx.bmw, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
init_groestl( &ctx.groestl, 64 );
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0,
@@ -222,7 +222,7 @@ void xevan_4way_hash( void *output, const void *input )
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3,
dataLen<<3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
skein512_4way_init( &ctx.skein );
skein512_4way( &ctx.skein, vhash, dataLen );
@@ -236,7 +236,7 @@ void xevan_4way_hash( void *output, const void *input )
keccak512_4way( &ctx.keccak, vhash, dataLen );
keccak512_4way_close( &ctx.keccak, vhash );
mm256_reinterleave_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
mm256_rintrlv_4x64_2x128( vhashA, vhashB, vhash, dataLen<<3 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhashA, vhashA, dataLen );
@@ -258,8 +258,8 @@ void xevan_4way_hash( void *output, const void *input )
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhashB, vhashB, dataLen<<3 );
mm256_deinterleave_1x128( hash0, hash1, vhashA, dataLen<<3 );
mm256_deinterleave_1x128( hash2, hash3, vhashB, dataLen<<3 );
mm256_dintrlv_2x128( hash0, hash1, vhashA, dataLen<<3 );
mm256_dintrlv_2x128( hash2, hash3, vhashB, dataLen<<3 );
init_echo( &ctx.echo, 512 );
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -274,13 +274,13 @@ void xevan_4way_hash( void *output, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, dataLen<<3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way( &ctx.hamsi, vhash, dataLen );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
sph_fugue512_init( &ctx.fugue );
sph_fugue512( &ctx.fugue, hash0, dataLen );
@@ -295,13 +295,13 @@ void xevan_4way_hash( void *output, const void *input )
sph_fugue512( &ctx.fugue, hash3, dataLen );
sph_fugue512_close( &ctx.fugue, hash3 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
shabal512_4way_init( &ctx.shabal );
shabal512_4way( &ctx.shabal, vhash, dataLen );
shabal512_4way_close( &ctx.shabal, vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, vhash, dataLen<<3 );
sph_whirlpool_init( &ctx.whirlpool );
sph_whirlpool( &ctx.whirlpool, hash0, dataLen );
@@ -316,13 +316,13 @@ void xevan_4way_hash( void *output, const void *input )
sph_whirlpool( &ctx.whirlpool, hash3, dataLen );
sph_whirlpool_close( &ctx.whirlpool, hash3 );
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, dataLen<<3 );
sha512_4way_init( &ctx.sha512 );
sha512_4way( &ctx.sha512, vhash, dataLen );
sha512_4way_close( &ctx.sha512, vhash );
mm256_reinterleave_4x64_4x32( vhashA, vhash, dataLen<<3 );
mm256_rintrlv_4x64_4x32( vhashA, vhash, dataLen<<3 );
haval256_5_4way_init( &ctx.haval );
haval256_5_4way( &ctx.haval, vhashA, dataLen );
@@ -336,7 +336,6 @@ int scanhash_xevan_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *hash7 = &(hash[7<<2]);
uint32_t lane_hash[8];
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
uint32_t _ALIGN(64) endiandata[20];
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
@@ -349,15 +348,9 @@ int scanhash_xevan_4way( int thr_id, struct work *work, uint32_t max_nonce,
if ( opt_benchmark )
ptarget[7] = 0x0cff;
uint64_t *edata = (uint64_t*)endiandata;
casti_m256i( edata, 0 ) = mm256_bswap_32( casti_m256i( pdata, 0 ) );
casti_m256i( edata, 1 ) = mm256_bswap_32( casti_m256i( pdata, 1 ) );
casti_m128i( edata, 4 ) = mm128_bswap_32( casti_m128i( pdata, 4 ) );
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_bswap_intrlv80_4x64( vdata, pdata );
do {
*noncev = mm256_interleave_blend_32( mm256_bswap_32(
*noncev = mm256_intrlv_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0,n+2, 0,n+1, 0, n, 0 ) ), *noncev );
xevan_4way_hash( hash, vdata );
@@ -365,17 +358,10 @@ int scanhash_xevan_4way( int thr_id, struct work *work, uint32_t max_nonce,
if ( hash7[ lane ] <= Htarg )
{
mm128_extract_lane_4x32( lane_hash, hash, lane, 256 );
if ( fulltest( lane_hash, ptarget ) )
if ( fulltest( lane_hash, ptarget ) && !opt_benchmark )
{
pdata[19] = n + lane;
work_set_target_ratio( work, lane_hash );
if ( submit_work( mythr, work ) )
applog( LOG_NOTICE,
"Share %d submitted by thread %d, lane %d.",
accepted_share_count + rejected_share_count + 1,
thr_id, lane );
else
applog( LOG_WARNING, "Failed to submit share." );
submit_solution( work, lane_hash, mythr, lane );
}
}
n += 4;

6
algo/x17/xevan.c
View File

@@ -237,8 +237,7 @@ int scanhash_xevan( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t _ALIGN(64) endiandata[20];
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
/* int */ thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
const uint32_t first_nonce = pdata[19];
uint32_t nonce = first_nonce;
@@ -250,8 +249,7 @@ int scanhash_xevan( int thr_id, struct work *work, uint32_t max_nonce,
for (int k=0; k < 19; k++)
be32enc(&endiandata[k], pdata[k]);
xevan_blake512_midstate( endiandata );
xevan_blake512_midstate( endiandata );
do {
be32enc(&endiandata[19], nonce);
xevan_hash(hash, endiandata);