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

v3.7.4

Browse Source
This commit is contained in:
Jay D Dee
2021-09-29 17:31:16 -04:00
parent 9b905fccc8
commit 2cd1507c2e
80 changed files with 8145 additions and 2097 deletions
Download Patch File Download Diff File Expand all files Collapse all files

53
algo/lyra2/allium-4way.c
View File

@@ -16,7 +16,7 @@
typedef struct {
blake256_16way_context blake;
keccak256_8way_context keccak;
cube_4way_context cube;
cube_4way_2buf_context cube;
skein256_8way_context skein;
#if defined(__VAES__)
groestl256_4way_context groestl;
@@ -30,13 +30,7 @@ static __thread allium_16way_ctx_holder allium_16way_ctx;
bool init_allium_16way_ctx()
{
keccak256_8way_init( &allium_16way_ctx.keccak );
cube_4way_init( &allium_16way_ctx.cube, 256, 16, 32 );
skein256_8way_init( &allium_16way_ctx.skein );
#if defined(__VAES__)
groestl256_4way_init( &allium_16way_ctx.groestl, 32 );
#else
init_groestl256( &allium_16way_ctx.groestl, 32 );
#endif
return true;
}
@@ -111,12 +105,11 @@ void allium_16way_hash( void *state, const void *input )
intrlv_2x256( vhash, hash14, hash15, 256 );
LYRA2RE_2WAY( vhash, 32, vhash, 32, 1, 8, 8 );
dintrlv_2x256( hash14, hash15, vhash, 256 );
intrlv_4x128( vhashA, hash0, hash1, hash2, hash3, 256 );
intrlv_4x128( vhashB, hash4, hash5, hash6, hash7, 256 );
cube_4way_full( &ctx.cube, vhashA, 256, vhashA, 32 );
cube_4way_full( &ctx.cube, vhashB, 256, vhashB, 32 );
cube_4way_2buf_full( &ctx.cube, vhashA, vhashB, 256, vhashA, vhashB, 32 );
dintrlv_4x128( hash0, hash1, hash2, hash3, vhashA, 256 );
dintrlv_4x128( hash4, hash5, hash6, hash7, vhashB, 256 );
@@ -124,8 +117,7 @@ void allium_16way_hash( void *state, const void *input )
intrlv_4x128( vhashA, hash8, hash9, hash10, hash11, 256 );
intrlv_4x128( vhashB, hash12, hash13, hash14, hash15, 256 );
cube_4way_full( &ctx.cube, vhashA, 256, vhashA, 32 );
cube_4way_full( &ctx.cube, vhashB, 256, vhashB, 32 );
cube_4way_2buf_full( &ctx.cube, vhashA, vhashB, 256, vhashA, vhashB, 32 );
dintrlv_4x128( hash8, hash9, hash10, hash11, vhashA, 256 );
dintrlv_4x128( hash12, hash13, hash14, hash15, vhashB, 256 );
@@ -255,7 +247,7 @@ int scanhash_allium_16way( struct work *work, uint32_t max_nonce,
typedef struct {
blake256_8way_context blake;
keccak256_4way_context keccak;
cubehashParam cube;
cube_2way_context cube;
skein256_4way_context skein;
#if defined(__VAES__)
groestl256_2way_context groestl;
@@ -269,13 +261,7 @@ static __thread allium_8way_ctx_holder allium_8way_ctx;
bool init_allium_8way_ctx()
{
keccak256_4way_init( &allium_8way_ctx.keccak );
cubehashInit( &allium_8way_ctx.cube, 256, 16, 32 );
skein256_4way_init( &allium_8way_ctx.skein );
#if defined(__VAES__)
groestl256_2way_init( &allium_8way_ctx.groestl, 32 );
#else
init_groestl256( &allium_8way_ctx.groestl, 32 );
#endif
return true;
}
@@ -320,21 +306,20 @@ void allium_8way_hash( void *hash, const void *input )
LYRA2RE( hash6, 32, hash6, 32, hash6, 32, 1, 8, 8 );
LYRA2RE( hash7, 32, hash7, 32, hash7, 32, 1, 8, 8 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*)hash0, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash1, (const byte*)hash1, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash2, (const byte*)hash2, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash3, (const byte*)hash3, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash4, (const byte*)hash4, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash5, (const byte*)hash5, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash6, (const byte*)hash6, 32 );
cubehashInit( &ctx.cube, 256, 16, 32 );
cubehashUpdateDigest( &ctx.cube, (byte*)hash7, (const byte*)hash7, 32 );
intrlv_2x128( vhashA, hash0, hash1, 256 );
intrlv_2x128( vhashB, hash2, hash3, 256 );
cube_2way_full( &ctx.cube, vhashA, 256, vhashA, 32 );
cube_2way_full( &ctx.cube, vhashB, 256, vhashB, 32 );
dintrlv_2x128( hash0, hash1, vhashA, 256 );
dintrlv_2x128( hash2, hash3, vhashB, 256 );
intrlv_2x128( vhashA, hash4, hash5, 256 );
intrlv_2x128( vhashB, hash6, hash7, 256 );
cube_2way_full( &ctx.cube, vhashA, 256, vhashA, 32 );
cube_2way_full( &ctx.cube, vhashB, 256, vhashB, 32 );
dintrlv_2x128( hash4, hash5, vhashA, 256 );
dintrlv_2x128( hash6, hash7, vhashB, 256 );
LYRA2RE( hash0, 32, hash0, 32, hash0, 32, 1, 8, 8 );
LYRA2RE( hash1, 32, hash1, 32, hash1, 32, 1, 8, 8 );

30
algo/lyra2/sponge.h
View File

@@ -66,13 +66,13 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
#define LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
G2W_4X64( s0, s1, s2, s3 ); \
s1 = mm512_ror256_64( s1); \
s3 = mm512_shufll256_64( s3 ); \
s1 = mm512_shuflr256_64( s1); \
s2 = mm512_swap256_128( s2 ); \
s3 = mm512_rol256_64( s3 ); \
G2W_4X64( s0, s1, s2, s3 ); \
s1 = mm512_rol256_64( s1 ); \
s2 = mm512_swap256_128( s2 ); \
s3 = mm512_ror256_64( s3 );
s3 = mm512_shuflr256_64( s3 ); \
s1 = mm512_shufll256_64( s1 ); \
s2 = mm512_swap256_128( s2 );
#define LYRA_12_ROUNDS_2WAY_AVX512( s0, s1, s2, s3 ) \
LYRA_ROUND_2WAY_AVX512( s0, s1, s2, s3 ) \
@@ -107,13 +107,13 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
#define LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
G_4X64( s0, s1, s2, s3 ); \
s1 = mm256_ror_1x64( s1); \
s3 = mm256_shufll_64( s3 ); \
s1 = mm256_shuflr_64( s1); \
s2 = mm256_swap_128( s2 ); \
s3 = mm256_rol_1x64( s3 ); \
G_4X64( s0, s1, s2, s3 ); \
s1 = mm256_rol_1x64( s1 ); \
s2 = mm256_swap_128( s2 ); \
s3 = mm256_ror_1x64( s3 );
s3 = mm256_shuflr_64( s3 ); \
s1 = mm256_shufll_64( s1 ); \
s2 = mm256_swap_128( s2 );
#define LYRA_12_ROUNDS_AVX2( s0, s1, s2, s3 ) \
LYRA_ROUND_AVX2( s0, s1, s2, s3 ) \
@@ -148,14 +148,14 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
#define LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
G_2X64( s0, s2, s4, s6 ); \
G_2X64( s1, s3, s5, s7 ); \
mm128_ror256_64( s2, s3 ); \
mm128_vrol256_64( s6, s7 ); \
mm128_vror256_64( s2, s3 ); \
mm128_swap256_128( s4, s5 ); \
mm128_rol256_64( s6, s7 ); \
G_2X64( s0, s2, s4, s6 ); \
G_2X64( s1, s3, s5, s7 ); \
mm128_rol256_64( s2, s3 ); \
mm128_swap256_128( s4, s5 ); \
mm128_ror256_64( s6, s7 );
mm128_vror256_64( s6, s7 ); \
mm128_vrol256_64( s2, s3 ); \
mm128_swap256_128( s4, s5 );
#define LYRA_12_ROUNDS_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \