#include "cpuminer-config.h" #include "quark-gate.h" #include #include #include #include "algo/blake/blake-hash-4way.h" #include "algo/bmw/bmw-hash-4way.h" #include "algo/skein/skein-hash-4way.h" #include "algo/jh/jh-hash-4way.h" #include "algo/keccak/keccak-hash-4way.h" #include "algo/groestl/aes_ni/hash-groestl.h" #if defined(__VAES__) #include "algo/groestl/groestl512-hash-4way.h" #endif #if defined (QUARK_8WAY) typedef struct { blake512_8way_context blake; bmw512_8way_context bmw; jh512_8way_context jh; skein512_8way_context skein; keccak512_8way_context keccak; #if defined(__VAES__) groestl512_4way_context groestl; #else hashState_groestl groestl; #endif } quark_8way_ctx_holder; quark_8way_ctx_holder quark_8way_ctx __attribute__ ((aligned (128))); void init_quark_8way_ctx() { blake512_8way_init( &quark_8way_ctx.blake ); bmw512_8way_init( &quark_8way_ctx.bmw ); skein512_8way_init( &quark_8way_ctx.skein ); jh512_8way_init( &quark_8way_ctx.jh ); keccak512_8way_init( &quark_8way_ctx.keccak ); #if defined(__VAES__) groestl512_4way_init( &quark_8way_ctx.groestl, 64 ); #else init_groestl( &quark_8way_ctx.groestl, 64 ); #endif } void quark_8way_hash( void *state, const void *input ) { uint64_t vhash[8*8] __attribute__ ((aligned (128))); uint64_t vhashA[8*8] __attribute__ ((aligned (64))); uint64_t vhashB[8*8] __attribute__ ((aligned (64))); uint64_t vhashC[8*8] __attribute__ ((aligned (64))); #if !defined(__VAES__) uint64_t hash0[8] __attribute__ ((aligned (64))); uint64_t hash1[8] __attribute__ ((aligned (64))); uint64_t hash2[8] __attribute__ ((aligned (64))); uint64_t hash3[8] __attribute__ ((aligned (64))); uint64_t hash4[8] __attribute__ ((aligned (64))); uint64_t hash5[8] __attribute__ ((aligned (64))); uint64_t hash6[8] __attribute__ ((aligned (64))); uint64_t hash7[8] __attribute__ ((aligned (64))); #endif __m512i* vh = (__m512i*)vhash; __m512i* vhA = (__m512i*)vhashA; __m512i* vhB = (__m512i*)vhashB; __m512i* vhC = (__m512i*)vhashC; __mmask8 vh_mask; quark_8way_ctx_holder ctx; const uint32_t mask = 8; const __m512i bit3_mask = m512_const1_64( mask ); memcpy( &ctx, &quark_8way_ctx, sizeof(quark_8way_ctx) ); blake512_8way_full( &ctx.blake, vhash, input, 80 ); bmw512_8way_full( &ctx.bmw, vhash, vhash, 64 ); vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask ); #if defined(__VAES__) rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 ); if ( ( vh_mask & 0x0f ) != 0x0f ) groestl512_4way_full( &ctx.groestl, vhashA, vhashA, 64 ); if ( ( vh_mask & 0xf0 ) != 0xf0 ) groestl512_4way_full( &ctx.groestl, vhashB, vhashB, 64 ); rintrlv_4x128_8x64( vhashC, vhashA, vhashB, 512 ); #else dintrlv_8x64( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7, vhash, 512 ); if ( hash0[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 ); if ( hash1[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 ); if ( hash2[0] & 8) groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 ); if ( hash3[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 ); if ( hash4[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash4, (char*)hash4, 512 ); if ( hash5[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash5, (char*)hash5, 512 ); if ( hash6[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash6, (char*)hash6, 512 ); if ( hash7[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash7, (char*)hash7, 512 ); intrlv_8x64( vhashC, hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7, 512 ); #endif if ( vh_mask & 0xff ) skein512_8way_full( &ctx.skein, vhashB, vhash, 64 ); mm512_blend_hash_8x64( vh, vhC, vhB, vh_mask ); #if defined(__VAES__) rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 ); groestl512_4way_full( &ctx.groestl, vhashA, vhashA, 64 ); groestl512_4way_full( &ctx.groestl, vhashB, vhashB, 64 ); rintrlv_4x128_8x64( vhash, vhashA, vhashB, 512 ); #else dintrlv_8x64( hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7, vhash, 512 ); groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 ); groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 ); groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 ); groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 ); groestl512_full( &ctx.groestl, (char*)hash4, (char*)hash4, 512 ); groestl512_full( &ctx.groestl, (char*)hash5, (char*)hash5, 512 ); groestl512_full( &ctx.groestl, (char*)hash6, (char*)hash6, 512 ); groestl512_full( &ctx.groestl, (char*)hash7, (char*)hash7, 512 ); intrlv_8x64( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6, hash7, 512 ); #endif jh512_8way_update( &ctx.jh, vhash, 64 ); jh512_8way_close( &ctx.jh, vhash ); vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask ); if ( ( vh_mask & 0xff ) != 0xff ) blake512_8way_full( &ctx.blake, vhashA, vhash, 64 ); if ( vh_mask & 0xff ) bmw512_8way_full( &ctx.bmw, vhashB, vhash, 64 ); mm512_blend_hash_8x64( vh, vhA, vhB, vh_mask ); keccak512_8way_update( &ctx.keccak, vhash, 64 ); keccak512_8way_close( &ctx.keccak, vhash ); skein512_8way_full( &ctx.skein, vhash, vhash, 64 ); vh_mask = _mm512_testn_epi64_mask( vh[0], bit3_mask ); if ( ( vh_mask & 0xff ) != 0xff ) { keccak512_8way_init( &ctx.keccak ); keccak512_8way_update( &ctx.keccak, vhash, 64 ); keccak512_8way_close( &ctx.keccak, vhashA ); } if ( vh_mask & 0xff ) { jh512_8way_init( &ctx.jh ); jh512_8way_update( &ctx.jh, vhash, 64 ); jh512_8way_close( &ctx.jh, vhashB ); } // Final blend, directly to state, only need 32 bytes. casti_m512i( state,0 ) = _mm512_mask_blend_epi64( vh_mask, vhA[0], vhB[0] ); casti_m512i( state,1 ) = _mm512_mask_blend_epi64( vh_mask, vhA[1], vhB[1] ); casti_m512i( state,2 ) = _mm512_mask_blend_epi64( vh_mask, vhA[2], vhB[2] ); casti_m512i( state,3 ) = _mm512_mask_blend_epi64( vh_mask, vhA[3], vhB[3] ); } int scanhash_quark_8way( struct work *work, uint32_t max_nonce, uint64_t *hashes_done, struct thr_info *mythr ) { uint64_t hash64[4*8] __attribute__ ((aligned (128))); uint32_t vdata[20*8] __attribute__ ((aligned (64))); uint32_t lane_hash[8] __attribute__ ((aligned (64))); uint64_t *hash64_q3 = &(hash64[3*8]); uint32_t *ptarget = work->target; const uint64_t targ64_q3 = ((uint64_t*)ptarget)[3]; uint32_t *pdata = work->data; uint32_t n = pdata[19]; const uint32_t first_nonce = pdata[19]; const uint32_t last_nonce = max_nonce - 8; __m512i *noncev = (__m512i*)vdata + 9; const int thr_id = mythr->id; const bool bench = opt_benchmark; mm512_bswap32_intrlv80_8x64( vdata, pdata ); *noncev = mm512_intrlv_blend_32( _mm512_set_epi32( n+7, 0, n+6, 0, n+5, 0, n+4, 0, n+3, 0, n+2, 0, n+1, 0, n , 0 ), *noncev ); do { quark_8way_hash( hash64, vdata ); for ( int lane = 0; lane < 8; lane++ ) if ( unlikely( hash64_q3[ lane ] <= targ64_q3 && !bench ) ) { extr_lane_8x64( lane_hash, hash64, lane, 256 ); if ( valid_hash( lane_hash, ptarget ) ) { pdata[19] = bswap_32( n + lane ); submit_solution( work, lane_hash, mythr ); } } *noncev = _mm512_add_epi32( *noncev, m512_const1_64( 0x0000000800000000 ) ); n += 8; } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) ); pdata[19] = n; *hashes_done = n - first_nonce; return 0; } #elif defined (QUARK_4WAY) typedef struct { blake512_4way_context blake; bmw512_4way_context bmw; hashState_groestl groestl; jh512_4way_context jh; skein512_4way_context skein; keccak512_4way_context keccak; } quark_4way_ctx_holder; quark_4way_ctx_holder quark_4way_ctx __attribute__ ((aligned (64))); void init_quark_4way_ctx() { blake512_4way_init( &quark_4way_ctx.blake ); bmw512_4way_init( &quark_4way_ctx.bmw ); init_groestl( &quark_4way_ctx.groestl, 64 ); skein512_4way_init( &quark_4way_ctx.skein ); jh512_4way_init( &quark_4way_ctx.jh ); keccak512_4way_init( &quark_4way_ctx.keccak ); } void quark_4way_hash( void *state, const void *input ) { uint64_t hash0[8] __attribute__ ((aligned (64))); uint64_t hash1[8] __attribute__ ((aligned (64))); uint64_t hash2[8] __attribute__ ((aligned (64))); uint64_t hash3[8] __attribute__ ((aligned (64))); uint64_t vhash[8*4] __attribute__ ((aligned (64))); uint64_t vhashA[8*4] __attribute__ ((aligned (64))); uint64_t vhashB[8*4] __attribute__ ((aligned (64))); __m256i* vh = (__m256i*)vhash; __m256i* vhA = (__m256i*)vhashA; __m256i* vhB = (__m256i*)vhashB; __m256i vh_mask; int h_mask; quark_4way_ctx_holder ctx; const __m256i bit3_mask = m256_const1_64( 8 ); const __m256i zero = _mm256_setzero_si256(); memcpy( &ctx, &quark_4way_ctx, sizeof(quark_4way_ctx) ); blake512_4way_full( &ctx.blake, vhash, input, 80 ); bmw512_4way_update( &ctx.bmw, vhash, 64 ); bmw512_4way_close( &ctx.bmw, vhash ); vh_mask = _mm256_cmpeq_epi64( _mm256_and_si256( vh[0], bit3_mask ), zero ); h_mask = _mm256_movemask_epi8( vh_mask ); dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 ); // A if ( hash0[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 ); if ( hash1[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 ); if ( hash2[0] & 8) groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 ); if ( hash3[0] & 8 ) groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 ); intrlv_4x64( vhashA, hash0, hash1, hash2, hash3, 512 ); // B if ( likely( h_mask & 0xffffffff ) ) skein512_4way_full( &ctx.skein, vhashB, vhash, 64 ); mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask ); dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 ); groestl512_full( &ctx.groestl, (char*)hash0, (char*)hash0, 512 ); groestl512_full( &ctx.groestl, (char*)hash1, (char*)hash1, 512 ); groestl512_full( &ctx.groestl, (char*)hash2, (char*)hash2, 512 ); groestl512_full( &ctx.groestl, (char*)hash3, (char*)hash3, 512 ); intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 ); jh512_4way_update( &ctx.jh, vhash, 64 ); jh512_4way_close( &ctx.jh, vhash ); vh_mask = _mm256_cmpeq_epi64( _mm256_and_si256( vh[0], bit3_mask ), zero ); h_mask = _mm256_movemask_epi8( vh_mask ); // A if ( likely( ( h_mask & 0xffffffff ) != 0xffffffff ) ) blake512_4way_full( &ctx.blake, vhashA, vhash, 64 ); // B if ( likely( h_mask & 0xffffffff ) ) { bmw512_4way_init( &ctx.bmw ); bmw512_4way_update( &ctx.bmw, vhash, 64 ); bmw512_4way_close( &ctx.bmw, vhashB ); } mm256_blend_hash_4x64( vh, vhA, vhB, vh_mask ); keccak512_4way_update( &ctx.keccak, vhash, 64 ); keccak512_4way_close( &ctx.keccak, vhash ); skein512_4way_full( &ctx.skein, vhash, vhash, 64 ); vh_mask = _mm256_cmpeq_epi64( _mm256_and_si256( vh[0], bit3_mask ), zero ); h_mask = _mm256_movemask_epi8( vh_mask ); // A if ( likely( ( h_mask & 0xffffffff ) != 0xffffffff ) ) { keccak512_4way_init( &ctx.keccak ); keccak512_4way_update( &ctx.keccak, vhash, 64 ); keccak512_4way_close( &ctx.keccak, vhashA ); } // B if ( likely( h_mask & 0xffffffff ) ) { jh512_4way_init( &ctx.jh ); jh512_4way_update( &ctx.jh, vhash, 64 ); jh512_4way_close( &ctx.jh, vhashB ); } // Final blend, directly to state, only need 32 bytes. casti_m256i( state, 0 ) = _mm256_blendv_epi8( vhA[0], vhB[0], vh_mask ); casti_m256i( state, 1 ) = _mm256_blendv_epi8( vhA[1], vhB[1], vh_mask ); casti_m256i( state, 2 ) = _mm256_blendv_epi8( vhA[2], vhB[2], vh_mask ); casti_m256i( state, 3 ) = _mm256_blendv_epi8( vhA[3], vhB[3], vh_mask ); } int scanhash_quark_4way( struct work *work, uint32_t max_nonce, uint64_t *hashes_done, struct thr_info *mythr ) { uint64_t hash64[4*4] __attribute__ ((aligned (64))); uint32_t vdata[20*4] __attribute__ ((aligned (64))); uint32_t lane_hash[8] __attribute__ ((aligned (64))); uint64_t *hash64_q3 = &(hash64[3*4]); uint32_t *pdata = work->data; uint32_t *ptarget = work->target; const uint64_t targ64_q3 = ((uint64_t*)ptarget)[3]; uint32_t n = pdata[19]; const uint32_t first_nonce = pdata[19]; const uint32_t last_nonce = max_nonce - 4; __m256i *noncev = (__m256i*)vdata + 9; const int thr_id = mythr->id; const bool bench = opt_benchmark; mm256_bswap32_intrlv80_4x64( vdata, pdata ); *noncev = mm256_intrlv_blend_32( _mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev ); do { quark_4way_hash( hash64, vdata ); for ( int lane = 0; lane < 4; lane++ ) if ( hash64_q3[ lane ] <= targ64_q3 && !bench ) { extr_lane_4x64( lane_hash, hash64, lane, 256 ); if ( valid_hash( lane_hash, ptarget ) ) { pdata[19] = bswap_32( n + lane ); submit_solution( work, lane_hash, mythr ); } } *noncev = _mm256_add_epi32( *noncev, m256_const1_64( 0x0000000400000000 ) ); n += 4; } while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) ); pdata[19] = n; *hashes_done = n - first_nonce; return 0; } #endif