Slightly improve entropy performance

Author: yoniko

lib/common/bits.h

@@ -172,4 +172,29 @@ MEM_STATIC unsigned ZSTD_highbit32(U32 val)   /* compress, dictBuilder, decodeCo
     return 31 - ZSTD_countLeadingZeros32(val);
 }
 
+/* ZSTD_rotateRight_*():
+ * Rotates a bitfield to the right by "count" bits.
+ * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts
+ */
+FORCE_INLINE_TEMPLATE
+U64 ZSTD_rotateRight_U64(U64 const value, U32 count) {
+    assert(count < 64);
+    count &= 0x3F; /* for fickle pattern recognition */
+    return (value >> count) | (U64)(value << ((0U - count) & 0x3F));
+}
+
+FORCE_INLINE_TEMPLATE
+U32 ZSTD_rotateRight_U32(U32 const value, U32 count) {
+    assert(count < 32);
+    count &= 0x1F; /* for fickle pattern recognition */
+    return (value >> count) | (U32)(value << ((0U - count) & 0x1F));
+}
+
+FORCE_INLINE_TEMPLATE
+U16 ZSTD_rotateRight_U16(U16 const value, U32 count) {
+    assert(count < 16);
+    count &= 0x0F; /* for fickle pattern recognition */
+    return (value >> count) | (U16)(value << ((0U - count) & 0x0F));
+}
+
 #endif /* ZSTD_BITS_H */

lib/compress/zstd_compress.c

@@ -27,7 +27,7 @@
 #include "zstd_opt.h"
 #include "zstd_ldm.h"
 #include "zstd_compress_superblock.h"
-#include  "../common/bits.h"      /* ZSTD_highbit32 */
+#include  "../common/bits.h"      /* ZSTD_highbit32, ZSTD_rotateRight_U64 */
 
 /* ***************************************************************
 *  Tuning parameters
@@ -1884,6 +1884,19 @@ typedef enum {
     ZSTD_resetTarget_CCtx
 } ZSTD_resetTarget_e;
 
+/* Mixes bits in a 64 bits in a value, based on XXH3_rrmxmx */
+static U64 ZSTD_bitmix(U64 val, U64 len) {
+    val ^= ZSTD_rotateRight_U64(val, 49) ^ ZSTD_rotateRight_U64(val, 24);
+    val *= 0x9FB21C651E98DF25ULL;
+    val ^= (val >> 35) + len ;
+    val *= 0x9FB21C651E98DF25ULL;
+    return val ^ (val >> 28);
+}
+
+/* Mixes in the hashSalt and hashSaltEntropy to create a new hashSalt */
+static void ZSTD_advanceHashSalt(ZSTD_matchState_t* ms) {
+    ms->hashSalt = ZSTD_bitmix(ms->hashSalt, 8) ^ ZSTD_bitmix((U64) ms->hashSaltEntropy, 4);
+}
 
 static size_t
 ZSTD_reset_matchState(ZSTD_matchState_t* ms,
@@ -1939,8 +1952,7 @@ ZSTD_reset_matchState(ZSTD_matchState_t* ms,
          * 0 when we reset a Cdict */
         if(forWho == ZSTD_resetTarget_CCtx) {
             ms->tagTable = (U16 *) ZSTD_cwksp_reserve_aligned_init_once(ws, tagTableSize);
-            ms->hashSalt = (U64) ZSTD_hashPtr(&ms->hashSalt, 32, sizeof(ms->hashSalt)) << 32 |
-                           ZSTD_hashPtr(&ms->hashSaltEntropy, 32, sizeof(ms->hashSaltEntropy));
+            ZSTD_advanceHashSalt(ms);
         } else {
             /* When we are not salting we want to always memset the memory */
             ms->tagTable = (U16 *) ZSTD_cwksp_reserve_aligned(ws, tagTableSize);

lib/compress/zstd_lazy.c

@@ -774,31 +774,6 @@ MEM_STATIC U32 ZSTD_VecMask_next(ZSTD_VecMask val) {
     return ZSTD_countTrailingZeros64(val);
 }
 
-/* ZSTD_rotateRight_*():
- * Rotates a bitfield to the right by "count" bits.
- * https://en.wikipedia.org/w/index.php?title=Circular_shift&oldid=991635599#Implementing_circular_shifts
- */
-FORCE_INLINE_TEMPLATE
-U64 ZSTD_rotateRight_U64(U64 const value, U32 count) {
-    assert(count < 64);
-    count &= 0x3F; /* for fickle pattern recognition */
-    return (value >> count) | (U64)(value << ((0U - count) & 0x3F));
-}
-
-FORCE_INLINE_TEMPLATE
-U32 ZSTD_rotateRight_U32(U32 const value, U32 count) {
-    assert(count < 32);
-    count &= 0x1F; /* for fickle pattern recognition */
-    return (value >> count) | (U32)(value << ((0U - count) & 0x1F));
-}
-
-FORCE_INLINE_TEMPLATE
-U16 ZSTD_rotateRight_U16(U16 const value, U32 count) {
-    assert(count < 16);
-    count &= 0x0F; /* for fickle pattern recognition */
-    return (value >> count) | (U16)(value << ((0U - count) & 0x0F));
-}
-
 /* ZSTD_row_nextIndex():
  * Returns the next index to insert at within a tagTable row, and updates the "head"
  * value to reflect the update. Essentially cycles backwards from [0, {entries per row})
@@ -891,6 +866,7 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
     U32* const hashTable = ms->hashTable;
     U16* const tagTable = ms->tagTable;
     U32 const hashLog = ms->rowHashLog;
+    U32 hashSaltEntropyCollected = 0;
     const BYTE* const base = ms->window.base;
 
     DEBUGLOG(6, "ZSTD_row_update_internalImpl(): updateStartIdx=%u, updateEndIdx=%u", updateStartIdx, updateEndIdx);
@@ -906,8 +882,9 @@ FORCE_INLINE_TEMPLATE void ZSTD_row_update_internalImpl(ZSTD_matchState_t* ms,
         assert(hash == ZSTD_hashPtrSalted(base + updateStartIdx, hashLog + ZSTD_ROW_HASH_TAG_BITS, mls, ms->hashSalt));
         ((BYTE*)tagRow)[pos + ZSTD_ROW_HASH_TAG_OFFSET] = hash & ZSTD_ROW_HASH_TAG_MASK;
         row[pos] = updateStartIdx;
-        ms->hashSaltEntropy += hash;
+        hashSaltEntropyCollected = hash;
     }
+    ms->hashSaltEntropy += hashSaltEntropyCollected; /* collect salt entropy */
 }
 
 /* ZSTD_row_update_internal():