datagen refactoring

diff --git a/lib/zstd.c b/lib/zstd.c
index b85df00a2e43c570bea8adff31204f8434364aac..38e7634a2e48843c2adb9c70a68ab1ef32998b80 100644 (file)
--- a/lib/zstd.c
+++ b/lib/zstd.c
@@ -861,13 +861,13 @@ static size_t ZSTD_compressSequences(BYTE* dst, size_t maxDstSize,
 }
 
 
-static void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* anchor, size_t offset, size_t matchLength)
+static void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const BYTE* literals, size_t offset, size_t matchLength)
 {
     BYTE* op_lit = seqStorePtr->lit;
     BYTE* const l_end = op_lit + litLength;
 
     /* copy Literals */
-    while (op_lit<l_end) COPY8(op_lit, anchor);
+    while (op_lit<l_end) COPY8(op_lit, literals);
     seqStorePtr->lit += litLength;
 
     /* literal Length */

diff --git a/programs/datagen.c b/programs/datagen.c
index 1ac4313f7416ea8e329fe00466ae27ff2db9bdd0..e5471727ef27ee7eeceb56ca3b529921c403458e 100644 (file)
--- a/programs/datagen.c
+++ b/programs/datagen.c
@@ -122,84 +122,85 @@ static char RDG_genChar(U32* seed, const void* ltctx)
     return lt[id];
 }
 
+#define RDG_DICTSIZE    (32 KB)
 #define RDG_RAND15BITS  ((RDG_rand(seed) >> 3) & 32767)
 #define RDG_RANDLENGTH  ( ((RDG_rand(seed) >> 7) & 7) ? (RDG_rand(seed) & 15) : (RDG_rand(seed) & 511) + 15)
-#define RDG_DICTSIZE    (32 KB)
-void RDG_generate(U64 size, U32 seedInit, double matchProba, double litProba)
+void RDG_genBlock(void* buffer, size_t buffSize, size_t prefixSize, double matchProba, void* litTable, unsigned* seedPtr)
 {
-    BYTE fullbuff[RDG_DICTSIZE + 128 KB + 1];
-    BYTE* buff = fullbuff + RDG_DICTSIZE;
-    U64 total=0;
-    U32 P32 = (U32)(32768 * matchProba);
-    U32 pos=1;
-    U32 genBlockSize = 128 KB;
-    void* ldctx = RDG_createLiteralDistrib(litProba);
-    FILE* fout = stdout;
-    U32* seed = &seedInit;;
+    BYTE* buffPtr = ((BYTE*)buffer) - prefixSize;
+    const U32 matchProba32 = (U32)(32768 * matchProba);
+    size_t pos = prefixSize;
+    void* ldctx = litTable;
+    U32* seed = seedPtr;
 
     /* init */
-    SET_BINARY_MODE(stdout);
-    fullbuff[0] = RDG_genChar(seed, ldctx);
-    while (pos<32 KB)
+    if (pos==0) buffPtr[0] = RDG_genChar(seed, ldctx), pos=1;
+
+    /* Generate compressible data */
+    while (pos < buffSize)
     {
         /* Select : Literal (char) or Match (within 32K) */
-        if (RDG_RAND15BITS < P32)
+        if (RDG_RAND15BITS < matchProba32)
         {
-            /* Copy (within 64K) */
+            /* Copy (within 32K) */
+            int match;
             U32 d;
-            int ref;
             int length = RDG_RANDLENGTH + 4;
             U32 offset = RDG_RAND15BITS + 1;
             if (offset > pos) offset = pos;
-            ref = pos - offset;
+            if (pos + length > buffSize) length = buffSize - pos;
+            match = pos - offset;
             d = pos + length;
-            while (pos < d) fullbuff[pos++] = fullbuff[ref++];
+            while (pos < d) buffPtr[pos++] = buffPtr[match++];
         }
         else
         {
             /* Literal (noise) */
-            U32 d = pos + RDG_RANDLENGTH;
-            while (pos < d) fullbuff[pos++] = RDG_genChar(seed, ldctx);
+            U32 d;
+            int length = RDG_RANDLENGTH;
+            if (pos + length > buffSize) length = buffSize - pos;
+            d = pos + length;
+            while (pos < d) buffPtr[pos++] = RDG_genChar(seed, ldctx);
         }
     }
+}
+
+
+void RDG_genBuffer(void* buffer, size_t size, double matchProba, double litProba, unsigned seed)
+{
+    void* ldctx;
+    if (litProba==0.0) litProba = matchProba / 3.6;
+    ldctx = RDG_createLiteralDistrib(litProba);
+    RDG_genBlock(buffer, size, 0, matchProba, ldctx, &seed);
+}
+
+
+#define RDG_BLOCKSIZE (128 KB)
+void RDG_genOut(unsigned long long size, double matchProba, double litProba, unsigned seed)
+{
+    BYTE fullbuff[RDG_DICTSIZE + RDG_BLOCKSIZE + 1];
+    BYTE* buff = fullbuff + RDG_DICTSIZE;
+    U64 total = 0;
+    U32 genBlockSize = RDG_BLOCKSIZE;
+    void* ldctx;
+
+    /* init */
+    if (litProba==0.0) litProba = matchProba / 3.6;
+    ldctx = RDG_createLiteralDistrib(litProba);
+    SET_BINARY_MODE(stdout);
+
+    /* Generate dict */
+    RDG_genBlock(fullbuff, RDG_DICTSIZE, 0, matchProba, ldctx, &seed);
 
     /* Generate compressible data */
-    pos = 0;
     while (total < size)
     {
-        if (size-total < 128 KB) genBlockSize = (U32)(size-total);
+        RDG_genBlock(buff, RDG_BLOCKSIZE, RDG_DICTSIZE, matchProba, ldctx, &seed);
+        if (size-total < RDG_BLOCKSIZE) genBlockSize = (U32)(size-total);
         total += genBlockSize;
         buff[genBlockSize] = 0;
-        pos = 0;
-        while (pos<genBlockSize)
-        {
-            /* Select : Literal (char) or Match (within 32K) */
-            if (RDG_RAND15BITS < P32)
-            {
-                /* Copy (within 64K) */
-                int ref;
-                U32 d;
-                int length = RDG_RANDLENGTH + 4;
-                U32 offset = RDG_RAND15BITS + 1;
-                if (pos + length > genBlockSize ) length = genBlockSize - pos;
-                ref = pos - offset;
-                d = pos + length;
-                while (pos < d) buff[pos++] = buff[ref++];
-            }
-            else
-            {
-                /* Literal (noise) */
-                U32 d;
-                int length = RDG_RANDLENGTH;
-                if (pos + length > genBlockSize) length = genBlockSize - pos;
-                d = pos + length;
-                while (pos < d) buff[pos++] = RDG_genChar(seed, ldctx);
-            }
-        }
-
-        /* output generated data */
-        fwrite(buff, 1, genBlockSize, fout);
-        /* Regenerate prefix */
-        memcpy(fullbuff, buff + 96 KB, 32 KB);
+        fwrite(buff, 1, genBlockSize, stdout);
+        /* update dict */
+        memcpy(fullbuff, buff + (RDG_BLOCKSIZE - RDG_DICTSIZE), RDG_DICTSIZE);
     }
 }

diff --git a/programs/datagen.h b/programs/datagen.h
index 6d07f2046d7760341c1177b72e1eeceab85a0815..631d1463db1aeacc6b7b806bdde5b51a3271697a 100644 (file)
--- a/programs/datagen.h
+++ b/programs/datagen.h
@@ -24,4 +24,17 @@
 */
 
 
-void RDG_generate(unsigned long long size, unsigned seed, double matchProba, double litProba);
+#include <stddef.h>   /* size_t */
+
+void RDG_genOut(unsigned long long size, double matchProba, double litProba, unsigned seed);
+void RDG_genBuffer(void* buffer, size_t size, double matchProba, double litProba, unsigned seed);
+/* RDG_genOut
+   Generate 'size' bytes of compressible data into stdout.
+   Compressibility can be controlled using 'matchProba'.
+   'LitProba' is optional, and affect variability of bytes. If litProba==0.0, default value is used.
+   Generated data can be selected using 'seed'.
+   If (matchProba, litProba and seed) are equal, the function always generate the same content.
+
+   RDG_genBuffer
+   Same as RDG_genOut, but generate data into provided buffer
+*/

diff --git a/programs/datagencli.c b/programs/datagencli.c
index b6ca508b57046bda6350ced142c4e62b916f6c6b..801e1980b03006f2787f2cdef7818eb54a0eb595 100644 (file)
--- a/programs/datagencli.c
+++ b/programs/datagencli.c
@@ -96,7 +96,7 @@ int main(int argc, char** argv)
 {
     int argNb;
     double proba = (double)COMPRESSIBILITY_DEFAULT / 100;
-    double litProba = proba / 3.6;
+    double litProba = 0.0;
     U64 size = SIZE_DEFAULT;
     U32 seed = SEED_DEFAULT;
     char* programName;
@@ -154,7 +154,6 @@ int main(int argc, char** argv)
                     }
                     if (proba>100.) proba=100.;
                     proba /= 100.;
-                    litProba = proba / 3.6;
                     break;
                 case 'L':   /* hidden argument : Literal distribution probability */
                     argument++;
@@ -184,7 +183,7 @@ int main(int argc, char** argv)
     DISPLAYLEVEL(3, "Seed = %u \n", seed);
     if (proba!=COMPRESSIBILITY_DEFAULT) DISPLAYLEVEL(3, "Compressibility : %i%%\n", (U32)(proba*100));
 
-    RDG_generate(size, seed, proba, litProba);
+    RDG_genOut(size, proba, litProba, seed);
     DISPLAYLEVEL(1, "\n");
 
     return 0;

diff --git a/programs/fullbench.c b/programs/fullbench.c
index 1992df6e0ed30283b55bb301945fdba5844ae5f9..99ac0b9e15b53c42146625d4dd4e4164056f3ee8 100644 (file)
--- a/programs/fullbench.c
+++ b/programs/fullbench.c
@@ -62,6 +62,7 @@
 
 #include "zstd.h"
 #include "fse_static.h"
+#include "datagen.h"
 
 
 /**************************************
@@ -215,6 +216,7 @@ static U64 BMK_GetFileSize(char* infilename)
 }
 
 
+#if 1
 static U32 BMK_rotl32(unsigned val32, unsigned nbBits) { return((val32 << nbBits) | (val32 >> (32 - nbBits))); }
 
 static U32 BMK_rand(U32* src)
@@ -265,7 +267,7 @@ static void BMK_datagen(void* buffer, size_t bufferSize, double proba, U32 seed)
         }
     }
 }
-
+#endif
 
 /*********************************************************
 *  Benchmark wrappers
@@ -329,8 +331,8 @@ size_t local_conditionalNull(void* dst, size_t dstSize, void* buff2, const void*
         if (b==0) total = 0;   // 825
         //if (!b) total = 0;     // 825
         //total = b ? total : 0; // 622
-        //total *= !!b;          // 465
         //total &= -!b;          // 622
+        //total *= !!b;          // 465
     }
     return total;
 }
@@ -357,7 +359,7 @@ size_t benchMem(void* src, size_t srcSize, U32 benchNb)
     double bestTime = 100000000.;
     size_t errorCode = 0;
 
-    // Declaration
+    /* Selection */
     switch(benchNb)
     {
     case 1:
@@ -399,7 +401,7 @@ size_t benchMem(void* src, size_t srcSize, U32 benchNb)
     case 11:
         g_cSize = ZSTD_compress(buff2, dstBuffSize, src, srcSize);
         break;
-    case 31:  // ZSTD_decodeLiteralsBlock
+    case 31:  /* ZSTD_decodeLiteralsBlock */
         {
             blockProperties_t bp;
             ZSTD_compress(dstBuff, dstBuffSize, src, srcSize);
@@ -417,7 +419,7 @@ size_t benchMem(void* src, size_t srcSize, U32 benchNb)
             srcSize = srcSize > 128 KB ? 128 KB : srcSize;   // relative to block
             break;
         }
-    case 32:   // ZSTD_decodeSeqHeaders
+    case 32:   /* ZSTD_decodeSeqHeaders */
         {
             blockProperties_t bp;
             const BYTE* ip = dstBuff;
@@ -444,15 +446,14 @@ size_t benchMem(void* src, size_t srcSize, U32 benchNb)
 
     /* test functions */
 
-    case 101:   // conditionalNull
+    case 101:   /* conditionalNull */
         {
             size_t i;
-            U32 seed = (U32)srcSize;
             for (i=0; i<srcSize; i++)
-                buff2[i] = (BYTE)(BMK_rand(&seed) & 15);
+                buff2[i] = i & 15;
             break;
         }
-    case 102:   //
+    case 102:   /* local_decodeLiteralsForward */
         {
             blockProperties_t bp;
             ZSTD_compress(dstBuff, dstBuffSize, src, srcSize);
@@ -515,6 +516,7 @@ int benchSample(U32 benchNb)
 
     /* Fill buffer */
     BMK_datagen(origBuff, benchedSize, g_compressibilityDefault, 0);
+    //RDG_generate(benchedSize, 0, g_compressibilityDefault, g_compressibilityDefault / 3.6);
 
     /* bench */
     DISPLAY("\r%79s\r", "");