v3.23.3

Makefile.am

@@ -36,21 +36,17 @@ cpuminer_SOURCES = \
   algo/argon2/argon2d/argon2d/argon2d_thread.c \
   algo/argon2/argon2d/argon2d/encoding.c \
   algo/blake/sph_blake.c \
-  algo/blake/blake256-hash-4way.c \
-  algo/blake/blake512-hash-4way.c \
+  algo/blake/blake256-hash.c \
+  algo/blake/blake512-hash.c \
   algo/blake/blake-gate.c \
   algo/blake/blake.c \
   algo/blake/blake-4way.c \
   algo/blake/sph_blake2b.c \
   algo/blake/sph-blake2s.c \
-  algo/blake/blake2s-hash-4way.c \
+  algo/blake/blake2s-hash.c \
   algo/blake/blake2s.c \
-  algo/blake/blake2s-gate.c \
-  algo/blake/blake2s-4way.c \
-  algo/blake/blake2b-hash-4way.c \
+  algo/blake/blake2b-hash.c \
   algo/blake/blake2b.c \
-  algo/blake/blake2b-gate.c \
-  algo/blake/blake2b-4way.c \
   algo/blake/blakecoin-gate.c \
   algo/blake/mod_blakecoin.c \
   algo/blake/blakecoin.c \

RELEASE_NOTES

@@ -65,6 +65,14 @@ If not what makes it happen or not happen?
 Change Log
 ----------
 
+v3.23.3
+
+#400: Removed excessive thread restarts when mining solo.
+Fixed build_msys2.sh for gcc-13 by removing unsupported option "--param=evrp-mode=legacy" from CFLAGS.
+Added CPUID detection and reporting of CPUs and SW builds supporting SHA512 extension.
+Added prototype of sha-512 using SHA512 intrinsics, untested.
+Other improvements and code cleanup.
+
 v3.23.2
 
 sha256dt, sha256t & sha256d +10% with SHA, small improvement with AVX2.

algo-gate-api.h

@@ -97,6 +97,7 @@ typedef  uint32_t set_t;
 #define SHA_OPT       0x20   // Zen1, Icelake (deprecated)
 #define AVX512_OPT    0x40   // Skylake-X, Zen4 (AVX512[F,VL,DQ,BW])
 #define VAES_OPT      0x80   // Icelake, Zen3
+#define SHA512_OPT   0x100   // Lunar Lake, Arrow Lake
 
 // AVX10 does not have explicit algo features:
 //  AVX10_512 is compatible with AVX512 + VAES

algo/blake/blake-4way.c

@@ -1,5 +1,5 @@
 #include "blake-gate.h"
-#include "blake-hash-4way.h"
+#include "blake256-hash.h"
 #include <string.h>
 #include <stdint.h>
 #include <memory.h>

algo/blake/blake-hash-4way.h → algo/blake/blake256-hash.h

@@ -1,5 +1,5 @@
-#ifndef BLAKE_HASH_4WAY__
-#define BLAKE_HASH_4WAY__ 1
+#ifndef BLAKE256_HASH__
+#define BLAKE256_HASH__ 1
 
 #include <stddef.h>
 #include "simd-utils.h"
@@ -11,13 +11,6 @@
 void  blake256_transform_le( uint32_t *H, const uint32_t *buf,
                              const uint32_t T0, const uint32_t T1, int rounds );
 
-/////////////////////////
-//
-//  Blake-512 1 way SSE2
-
-void  blake512_transform_le( uint64_t *H, const uint64_t *buf,
-                             const uint64_t T0, const uint64_t T1 );
-
 //////////////////////////
 //
 //   Blake-256 4 way SSE2
@@ -86,30 +79,6 @@ void blake256r8_8way_init(void *cc);
 void blake256r8_8way_update(void *cc, const void *data, size_t len);
 void blake256r8_8way_close(void *cc, void *dst);
 
-// Blake-512 4 way AVX2
-
-typedef struct {
-   __m256i buf[16];
-   __m256i H[8];
-   __m256i S[4];   
-   size_t ptr;
-   uint64_t T0, T1;
-} blake_4way_big_context __attribute__ ((aligned (128)));
-
-typedef blake_4way_big_context blake512_4way_context;
-
-void blake512_4way_init( blake_4way_big_context *sc );
-void blake512_4way_update( void *cc, const void *data, size_t len );
-void blake512_4way_close( void *cc, void *dst );
-void blake512_4way_full( blake_4way_big_context *sc, void * dst,
-                         const void *data, size_t len );
-void blake512_4way_full_le( blake_4way_big_context *sc, void * dst,
-                            const void *data, size_t len );
-void blake512_4way_prehash_le( blake_4way_big_context *sc, __m256i *midstate,
-                               const void *data );
-void blake512_4way_final_le( blake_4way_big_context *sc, void *hash,
-                             const __m256i nonce, const __m256i *midstate );
-
 #if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
 
 ////////////////////////////
@@ -137,7 +106,6 @@ void blake256_16way_round0_prehash_le( void *midstate, const void *midhash,
 void blake256_16way_final_rounds_le( void *final_hash, const void *midstate,
                      const void *midhash, const void *data, const int rounds );
 
-
 // 14 rounds, blake, decred
 typedef blake_16way_small_context blake256r14_16way_context;
 void blake256r14_16way_init(void *cc);
@@ -150,33 +118,7 @@ void blake256r8_16way_init(void *cc);
 void blake256r8_16way_update(void *cc, const void *data, size_t len);
 void blake256r8_16way_close(void *cc, void *dst);
 
-////////////////////////////
-//
-//// Blake-512 8 way AVX512
-
-typedef struct {
-   __m512i buf[16];
-   __m512i H[8];
-   __m512i S[4];
-   size_t ptr;
-   uint64_t T0, T1;
-} blake_8way_big_context __attribute__ ((aligned (128)));
-
-typedef blake_8way_big_context blake512_8way_context;
-
-void blake512_8way_init( blake_8way_big_context *sc );
-void blake512_8way_update( void *cc, const void *data, size_t len );
-void blake512_8way_close( void *cc, void *dst );
-void blake512_8way_full( blake_8way_big_context *sc, void * dst,
-                        const void *data, size_t len );
-void blake512_8way_full_le( blake_8way_big_context *sc, void * dst,
-                            const void *data, size_t len );
-void blake512_8way_prehash_le( blake_8way_big_context *sc, __m512i *midstate,
-                               const void *data );
-void blake512_8way_final_le( blake_8way_big_context *sc, void *hash,
-                             const __m512i nonce, const __m512i *midstate );
-
 #endif  // AVX512
 #endif  // AVX2
 
-#endif  // BLAKE_HASH_4WAY_H__
+#endif  // BLAKE256_HASH_H__