v3.21.5

algo/blake/sph_blake2b.c

@@ -103,16 +103,16 @@
    const uint8_t *sigmaR = sigma[R]; \
    BLAKE2B_G( V[0], V[2], V[4], V[6], 0, 1, 2, 3 ); \
    BLAKE2B_G( V[1], V[3], V[5], V[7], 4, 5, 6, 7 ); \
-   V2 = mm128_alignr_64( V[3], V[2] ); \
-   V3 = mm128_alignr_64( V[2], V[3] ); \
-   V6 = mm128_alignr_64( V[6], V[7] ); \
-   V7 = mm128_alignr_64( V[7], V[6] ); \
+   V2 = mm128_alignr_64( V[3], V[2], 1 ); \
+   V3 = mm128_alignr_64( V[2], V[3], 1 ); \
+   V6 = mm128_alignr_64( V[6], V[7], 1 ); \
+   V7 = mm128_alignr_64( V[7], V[6], 1 ); \
    BLAKE2B_G( V[0], V2, V[5], V6,  8,  9, 10, 11 ); \
    BLAKE2B_G( V[1], V3, V[4], V7, 12, 13, 14, 15 ); \
-   V[2] = mm128_alignr_64( V2, V3 ); \
-   V[3] = mm128_alignr_64( V3, V2 ); \
-   V[6] = mm128_alignr_64( V7, V6 ); \
-   V[7] = mm128_alignr_64( V6, V7 ); \
+   V[2] = mm128_alignr_64( V2, V3, 1 ); \
+   V[3] = mm128_alignr_64( V3, V2, 1 ); \
+   V[6] = mm128_alignr_64( V7, V6, 1 ); \
+   V[7] = mm128_alignr_64( V6, V7, 1 ); \
 }
 
 #else

algo/groestl/myr-groestl.c

@@ -73,11 +73,11 @@ int scanhash_myriad( struct work *work, uint32_t max_nonce,
       be32enc(&endiandata[19], nonce);
       myriad_hash(hash, endiandata);
 
-      if (hash[7] <= Htarg && fulltest(hash, ptarget))
+      if (hash[7] <= Htarg )
+      if ( fulltest(hash, ptarget) && !opt_benchmark )
       {
          pdata[19] = nonce;
-         *hashes_done = pdata[19] - first_nonce;
-         return 1;
+         submit_solution( work, hash, mythr );
       }
       nonce++;
 

algo/luffa/luffa_for_sse2.c

@@ -19,26 +19,34 @@
  */
 
 #include <string.h>
-#include <emmintrin.h>
 #include "simd-utils.h"
 #include "luffa_for_sse2.h"
 
-#if defined(__SSE4_1__)
+#if defined(__AVX512VL__)
+
+#define MULT2( a0, a1 ) \
+{ \
+  __m128i b = _mm_xor_si128( a0, _mm_maskz_shuffle_epi32( 0xb, a1, 0x10 ) ); \
+  a0 = _mm_alignr_epi32( a1, b, 1 ); \
+  a1 = _mm_alignr_epi32( b, a1, 1 ); \
+}
+
+#elif defined(__SSE4_1__)
 
 #define MULT2( a0, a1 ) do \
 { \
-  __m128i b =  _mm_xor_si128( a0, _mm_shuffle_epi32( mm128_mask_32( a1, 0xe ), 0x10 ) ); \
-  a0 = _mm_or_si128( _mm_srli_si128( b, 4 ), _mm_slli_si128( a1, 12 ) ); \
-  a1 = _mm_or_si128( _mm_srli_si128( a1, 4 ), _mm_slli_si128( b, 12 ) );  \
+  __m128i b = _mm_xor_si128( a0, _mm_shuffle_epi32( mm128_mask_32( a1, 0xe ), 0x10 ) ); \
+  a0 = _mm_alignr_epi8( a1, b, 4 ); \
+  a1 = _mm_alignr_epi8( b, a1, 4 ); \
 } while(0)
 
 #else
 
 #define MULT2( a0, a1 ) do \
 { \
-  __m128i b =  _mm_xor_si128( a0, _mm_shuffle_epi32( _mm_and_si128( a1, MASK ), 16 ) ); \
+  __m128i b = _mm_xor_si128( a0, _mm_shuffle_epi32( _mm_and_si128( a1, MASK ), 0x10 ) ); \
   a0 = _mm_or_si128( _mm_srli_si128( b, 4 ), _mm_slli_si128( a1, 12 ) ); \
-  a1 = _mm_or_si128( _mm_srli_si128( a1, 4 ), _mm_slli_si128( b, 12 ) );  \
+  a1 = _mm_or_si128( _mm_srli_si128( a1, 4 ), _mm_slli_si128( b, 12 ) ); \
 } while(0)
 
 #endif

algo/lyra2/sponge.h

@@ -146,14 +146,25 @@ static inline uint64_t rotr64( const uint64_t w, const unsigned c ){
    b = mm128_ror_64( _mm_xor_si128( b, c ), 63 );
 
 #define LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
+{ \
+   __m128i t; \
    G_2X64( s0, s2, s4, s6 ); \
    G_2X64( s1, s3, s5, s7 ); \
-   mm128_vrol256_64( s6, s7 ); \
-   mm128_vror256_64( s2, s3 ); \
+   t = mm128_alignr_64( s7, s6, 1 ); \
+   s6 = mm128_alignr_64( s6, s7, 1 ); \
+   s7 = t; \
+   t = mm128_alignr_64( s2, s3, 1 ); \
+   s2 =  mm128_alignr_64( s3, s2, 1 ); \
+   s3 = t; \
    G_2X64( s0, s2, s5, s6 ); \
    G_2X64( s1, s3, s4, s7 ); \
-   mm128_vror256_64( s6, s7 ); \
-   mm128_vrol256_64( s2, s3 );
+   t = mm128_alignr_64( s6, s7, 1 ); \
+   s6 = mm128_alignr_64( s7, s6, 1 ); \
+   s7 = t; \
+   t = mm128_alignr_64( s3, s2, 1 ); \
+   s2 =  mm128_alignr_64( s2, s3, 1 ); \
+   s3 = t; \
+} 
 
 #define LYRA_12_ROUNDS_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \
    LYRA_ROUND_AVX(s0,s1,s2,s3,s4,s5,s6,s7) \

algo/skein/skein.c

@@ -31,18 +31,19 @@ int scanhash_skein( struct work *work, uint32_t max_nonce,
 	const uint32_t Htarg = ptarget[7];
 	const uint32_t first_nonce = pdata[19];
 	uint32_t n = first_nonce;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   int thr_id = mythr->id;
 
    swab32_array( endiandata, pdata, 20 );
 
 	do {
 		be32enc(&endiandata[19], n); 
 		skeinhash(hash64, endiandata);
-		if (hash64[7] < Htarg && fulltest(hash64, ptarget)) {
-			*hashes_done = n - first_nonce + 1;
-			pdata[19] = n;
-			return true;
-		}
+      if (hash64[7] <= Htarg )
+      if ( fulltest(hash64, ptarget) && !opt_benchmark )
+      {
+         pdata[19] = n;
+         submit_solution( work, hash64, mythr );
+      }
 		n++;
 
 	} while (n < max_nonce && !work_restart[thr_id].restart);

algo/skein/skein2.c

@@ -34,31 +34,31 @@ void skein2hash(void *output, const void *input)
 	sph_skein512_close(&ctx_skein, hash);
 
 	memcpy(output, hash, 32);
-
 }
 
 int scanhash_skein2( struct work *work,	uint32_t max_nonce,
                      uint64_t *hashes_done, struct thr_info *mythr )
 {
-        uint32_t *pdata = work->data;
-        uint32_t *ptarget = work->target;
+   uint32_t *pdata = work->data;
+   uint32_t *ptarget = work->target;
 	uint32_t hash64[8] __attribute__ ((aligned (64)));
 	uint32_t endiandata[20] __attribute__ ((aligned (64)));
 	const uint32_t Htarg = ptarget[7];
 	const uint32_t first_nonce = pdata[19];
 	uint32_t n = first_nonce;
-   int thr_id = mythr->id;  // thr_id arg is deprecated
+   int thr_id = mythr->id; 
 
-        swab32_array( endiandata, pdata, 20 );
+   swab32_array( endiandata, pdata, 20 );
 
 	do {
 		be32enc(&endiandata[19], n);
 		skein2hash(hash64, endiandata);
-		if (hash64[7] < Htarg && fulltest(hash64, ptarget)) {
-			*hashes_done = n - first_nonce + 1;
-			pdata[19] = n;
-			return true;
-		}
+      if (hash64[7] <= Htarg )
+      if ( fulltest(hash64, ptarget) && !opt_benchmark )
+      {
+         pdata[19] = n;
+         submit_solution( work, hash64, mythr );
+      }
 		n++;
 
 	} while (n < max_nonce && !work_restart[thr_id].restart);