zstd-native-ruby 1.0.0

data/vendor/zstd/programs/benchzstd.h

@@ -0,0 +1,191 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /* benchzstd :
+  * benchmark Zstandard compression / decompression
+  * over a set of files or buffers
+  * and display progress result and final summary
+  */
+
+#ifndef BENCH_ZSTD_H_3242387
+#define BENCH_ZSTD_H_3242387
+
+/* ===  Dependencies  === */
+#include <stddef.h>   /* size_t */
+#define ZSTD_STATIC_LINKING_ONLY   /* ZSTD_compressionParameters */
+#include "../lib/zstd.h"     /* ZSTD_compressionParameters */
+
+/* ===  Constants  === */
+
+#define MB_UNIT 1000000
+
+
+/* ===  Benchmark functions  === */
+
+/* Creates a variant `typeName`, able to express "error or valid result".
+ * Functions with return type `typeName`
+ * must first check if result is valid, using BMK_isSuccessful_*(),
+ * and only then can extract `baseType`.
+ */
+#define VARIANT_ERROR_RESULT(baseType, variantName)  \
+                                             \
+typedef struct {                             \
+    baseType internal_never_use_directly;    \
+    int tag;                                 \
+} variantName
+
+
+typedef struct {
+    size_t cSize;
+    unsigned long long cSpeed;   /* bytes / sec */
+    unsigned long long dSpeed;
+    size_t cMem;                 /* memory usage during compression */
+} BMK_benchResult_t;
+
+VARIANT_ERROR_RESULT(BMK_benchResult_t, BMK_benchOutcome_t);
+
+/* check first if the return structure represents an error or a valid result */
+int BMK_isSuccessful_benchOutcome(BMK_benchOutcome_t outcome);
+
+/* extract result from variant type.
+ * note : this function will abort() program execution if result is not valid
+ *        check result validity first, by using BMK_isSuccessful_benchOutcome()
+ */
+BMK_benchResult_t BMK_extract_benchResult(BMK_benchOutcome_t outcome);
+
+
+/*! BMK_benchFiles() -- called by zstdcli */
+/*  Loads files from fileNamesTable into memory,
+ *  and an optional dictionary from dictFileName (can be NULL),
+ *  then uses benchMem().
+ *  fileNamesTable - name of files to benchmark.
+ *  nbFiles - number of files (size of fileNamesTable), must be > 0.
+ *  dictFileName - name of dictionary file to load.
+ *  cLevel - compression level to benchmark, errors if invalid.
+ *  compressionParams - advanced compression Parameters.
+ *  displayLevel - what gets printed:
+ *      0 : no display;
+ *      1 : errors;
+ *      2 : + result + interaction + warnings;
+ *      3 : + information;
+ *      4 : + debug
+ * @return: 0 on success, !0 on error
+ */
+int BMK_benchFiles(
+            const char* const * fileNamesTable, unsigned nbFiles,
+            const char* dictFileName,
+            int cLevel, const ZSTD_compressionParameters* compressionParams,
+            int displayLevel);
+
+
+typedef enum {
+    BMK_both = 0,
+    BMK_decodeOnly = 1,
+    BMK_compressOnly = 2
+} BMK_mode_t;
+
+typedef struct {
+    BMK_mode_t mode;        /* 0: both, 1: compress only 2: decode only */
+    unsigned nbSeconds;     /* default timing is in nbSeconds */
+    size_t chunkSizeMax;    /* Maximum size of each independent chunk */
+    size_t targetCBlockSize;/* Approximative size of compressed blocks */
+    int nbWorkers;          /* multithreading */
+    unsigned realTime;      /* real time priority */
+    int additionalParam;    /* used by python speed benchmark */
+    int ldmFlag;            /* enables long distance matching */
+    int ldmMinMatch;        /* below: parameters for long distance matching, see zstd.1.md */
+    int ldmHashLog;
+    int ldmBucketSizeLog;
+    int ldmHashRateLog;
+    ZSTD_ParamSwitch_e literalCompressionMode;
+    int useRowMatchFinder;  /* use row-based matchfinder if possible */
+} BMK_advancedParams_t;
+
+/* returns default parameters used by nonAdvanced functions */
+BMK_advancedParams_t BMK_initAdvancedParams(void);
+
+/*! BMK_benchFilesAdvanced():
+ *  Same as BMK_benchFiles(),
+ *  with more controls, provided through advancedParams_t structure */
+int BMK_benchFilesAdvanced(
+               const char* const * fileNamesTable, unsigned nbFiles,
+               const char* dictFileName,
+               int startCLevel, int endCLevel,
+               const ZSTD_compressionParameters* compressionParams,
+               int displayLevel, const BMK_advancedParams_t* adv);
+
+/*! BMK_syntheticTest() -- called from zstdcli */
+/*  Generates a sample with datagen, using @compressibility argument
+ * @cLevel - compression level to benchmark, errors if invalid
+ * @compressibility - determines compressibility of sample, range [0.0 - 1.0]
+ *        if @compressibility < 0.0, uses the lorem ipsum generator
+ * @compressionParams - basic compression Parameters
+ * @displayLevel - see benchFiles
+ * @adv - see advanced_Params_t
+ * @return: 0 on success, !0 on error
+ */
+int BMK_syntheticTest(double compressibility,
+                      int startingCLevel, int endCLevel,
+                      const ZSTD_compressionParameters* compressionParams,
+                      int displayLevel, const BMK_advancedParams_t* adv);
+
+
+
+/* ===  Benchmark Zstandard in a memory-to-memory scenario  === */
+
+/** BMK_benchMem() -- core benchmarking function, called in paramgrill
+ *  applies ZSTD_compress_generic() and ZSTD_decompress_generic() on data in srcBuffer
+ *  with specific compression parameters provided by other arguments using benchFunction
+ *  (cLevel, comprParams + adv in advanced Mode) */
+/*  srcBuffer - data source, expected to be valid compressed data if in Decode Only Mode
+ *  srcSize - size of data in srcBuffer
+ *  fileSizes - srcBuffer is considered cut into 1+ segments, to compress separately.
+ *              note : sum(fileSizes) must be == srcSize.  (<== ensure it's properly checked)
+ *  nbFiles - nb of segments
+ *  cLevel - compression level
+ *  comprParams - basic compression parameters
+ *  dictBuffer - a dictionary if used, null otherwise
+ *  dictBufferSize - size of dictBuffer, 0 otherwise
+ *  displayLevel - see BMK_benchFiles
+ *  displayName - name used by display
+ * @return:
+ *      a variant, which expresses either an error, or a valid result.
+ *      Use BMK_isSuccessful_benchOutcome() to check if function was successful.
+ *      If yes, extract the valid result with BMK_extract_benchResult(),
+ *      it will contain :
+ *          .cSpeed: compression speed in bytes per second,
+ *          .dSpeed: decompression speed in bytes per second,
+ *          .cSize : compressed size, in bytes
+ *          .cMem  : memory budget required for the compression context
+ */
+BMK_benchOutcome_t BMK_benchMem(const void* srcBuffer, size_t srcSize,
+                        const size_t* fileSizes, unsigned nbFiles,
+                        int cLevel, const ZSTD_compressionParameters* comprParams,
+                        const void* dictBuffer, size_t dictBufferSize,
+                        int displayLevel, const char* displayName);
+
+
+/* BMK_benchMemAdvanced() : used by Paramgrill
+ * same as BMK_benchMem() with following additional options :
+ * dstBuffer - destination buffer to write compressed output in, NULL if none provided.
+ * dstCapacity - capacity of destination buffer, give 0 if dstBuffer = NULL
+ * adv = see advancedParams_t
+ */
+BMK_benchOutcome_t BMK_benchMemAdvanced(const void* srcBuffer, size_t srcSize,
+                        void* dstBuffer, size_t dstCapacity,
+                        const size_t* fileSizes, unsigned nbFiles,
+                        int cLevel, const ZSTD_compressionParameters* comprParams,
+                        const void* dictBuffer, size_t dictBufferSize,
+                        int displayLevel, const char* displayName,
+                        const BMK_advancedParams_t* adv);
+
+
+
+#endif   /* BENCH_ZSTD_H_3242387 */

data/vendor/zstd/programs/datagen.c

@@ -0,0 +1,186 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+
+/*-************************************
+*  Dependencies
+**************************************/
+#include "datagen.h"
+#include "platform.h"  /* SET_BINARY_MODE */
+#include <stdlib.h>    /* malloc, free */
+#include <stdio.h>     /* FILE, fwrite, fprintf */
+#include <string.h>    /* memcpy */
+#include "../lib/common/mem.h"  /* U32 */
+
+
+/*-************************************
+*  Macros
+**************************************/
+#define KB *(1 <<10)
+#define MIN(a,b)  ( (a) < (b) ? (a) : (b) )
+
+#define RDG_DEBUG 0
+#define TRACE(...)   if (RDG_DEBUG) fprintf(stderr, __VA_ARGS__ )
+
+
+/*-************************************
+*  Local constants
+**************************************/
+#define LTLOG 13
+#define LTSIZE (1<<LTLOG)
+#define LTMASK (LTSIZE-1)
+
+
+/*-*******************************************************
+*  Local Functions
+*********************************************************/
+#define RDG_rotl32(x,r) ((x << r) | (x >> (32 - r)))
+static U32 RDG_rand(U32* src)
+{
+    static const U32 prime1 = 2654435761U;
+    static const U32 prime2 = 2246822519U;
+    U32 rand32 = *src;
+    rand32 *= prime1;
+    rand32 ^= prime2;
+    rand32  = RDG_rotl32(rand32, 13);
+    *src = rand32;
+    return rand32 >> 5;
+}
+
+typedef U32 fixedPoint_24_8;
+
+static void RDG_fillLiteralDistrib(BYTE* ldt, fixedPoint_24_8 ld)
+{
+    BYTE const firstChar = (ld<=0.0) ?   0 : '(';
+    BYTE const lastChar  = (ld<=0.0) ? 255 : '}';
+    BYTE character = (ld<=0.0) ? 0 : '0';
+    U32 u;
+
+    if (ld<=0) ld = 0;
+    for (u=0; u<LTSIZE; ) {
+        U32 const weight = (((LTSIZE - u) * ld) >> 8) + 1;
+        U32 const end = MIN ( u + weight , LTSIZE);
+        while (u < end) ldt[u++] = character;
+        character++;
+        if (character > lastChar) character = firstChar;
+    }
+}
+
+
+static BYTE RDG_genChar(U32* seed, const BYTE* ldt)
+{
+    U32 const id = RDG_rand(seed) & LTMASK;
+    return ldt[id];  /* memory-sanitizer fails here, stating "uninitialized value" when table initialized with P==0.0. Checked : table is fully initialized */
+}
+
+
+static U32 RDG_rand15Bits (U32* seedPtr)
+{
+    return RDG_rand(seedPtr) & 0x7FFF;
+}
+
+static U32 RDG_randLength(U32* seedPtr)
+{
+    if (RDG_rand(seedPtr) & 7) return (RDG_rand(seedPtr) & 0xF);   /* small length */
+    return (RDG_rand(seedPtr) & 0x1FF) + 0xF;
+}
+
+static void RDG_genBlock(void* buffer, size_t buffSize, size_t prefixSize,
+                         double matchProba, const BYTE* ldt, U32* seedPtr)
+{
+    BYTE* const buffPtr = (BYTE*)buffer;
+    U32 const matchProba32 = (U32)(32768 * matchProba);
+    size_t pos = prefixSize;
+    U32 prevOffset = 1;
+
+    /* special case : sparse content */
+    while (matchProba >= 1.0) {
+        size_t size0 = RDG_rand(seedPtr) & 3;
+        size0  = (size_t)1 << (16 + size0 * 2);
+        size0 += RDG_rand(seedPtr) & (size0-1);   /* because size0 is power of 2*/
+        if (buffSize < pos + size0) {
+            memset(buffPtr+pos, 0, buffSize-pos);
+            return;
+        }
+        memset(buffPtr+pos, 0, size0);
+        pos += size0;
+        buffPtr[pos-1] = RDG_genChar(seedPtr, ldt);
+        continue;
+    }
+
+    /* init */
+    if (pos==0) buffPtr[0] = RDG_genChar(seedPtr, ldt), pos=1;
+
+    /* Generate compressible data */
+    while (pos < buffSize) {
+        /* Select : Literal (char) or Match (within 32K) */
+        if (RDG_rand15Bits(seedPtr) < matchProba32) {
+            /* Copy (within 32K) */
+            U32 const length = RDG_randLength(seedPtr) + 4;
+            U32 const d = (U32) MIN(pos + length , buffSize);
+            U32 const repeatOffset = (RDG_rand(seedPtr) & 15) == 2;
+            U32 const randOffset = RDG_rand15Bits(seedPtr) + 1;
+            U32 const offset = repeatOffset ? prevOffset : (U32) MIN(randOffset , pos);
+            size_t match = pos - offset;
+            while (pos < d) { buffPtr[pos++] = buffPtr[match++];   /* correctly manages overlaps */ }
+            prevOffset = offset;
+        } else {
+            /* Literal (noise) */
+            U32 const length = RDG_randLength(seedPtr);
+            U32 const d = (U32) MIN(pos + length, buffSize);
+            while (pos < d) { buffPtr[pos++] = RDG_genChar(seedPtr, ldt); }
+    }   }
+}
+
+
+void RDG_genBuffer(void* buffer, size_t size, double matchProba, double litProba, unsigned seed)
+{
+    U32 seed32 = seed;
+    BYTE ldt[LTSIZE];
+    memset(ldt, '0', sizeof(ldt));  /* yes, character '0', this is intentional */
+    if (litProba<=0.0) litProba = matchProba / 4.5;
+    RDG_fillLiteralDistrib(ldt, (fixedPoint_24_8)(litProba * 256 + 0.001));
+    RDG_genBlock(buffer, size, 0, matchProba, ldt, &seed32);
+}
+
+
+void RDG_genStdout(unsigned long long size, double matchProba, double litProba, unsigned seed)
+{
+    U32 seed32 = seed;
+    size_t const stdBlockSize = 128 KB;
+    size_t const stdDictSize = 32 KB;
+    BYTE* const buff = (BYTE*)malloc(stdDictSize + stdBlockSize);
+    U64 total = 0;
+    BYTE ldt[LTSIZE];   /* literals distribution table */
+
+    /* init */
+    if (buff==NULL) { perror("datagen"); exit(1); }
+    if (litProba<=0.0) litProba = matchProba / 4.5;
+    memset(ldt, '0', sizeof(ldt));   /* yes, character '0', this is intentional */
+    RDG_fillLiteralDistrib(ldt, (fixedPoint_24_8)(litProba * 256 + 0.001));
+    SET_BINARY_MODE(stdout);
+
+    /* Generate initial dict */
+    RDG_genBlock(buff, stdDictSize, 0, matchProba, ldt, &seed32);
+
+    /* Generate compressible data */
+    while (total < size) {
+        size_t const genBlockSize = (size_t) (MIN (stdBlockSize, size-total));
+        RDG_genBlock(buff, stdDictSize+stdBlockSize, stdDictSize, matchProba, ldt, &seed32);
+        total += genBlockSize;
+        { size_t const unused = fwrite(buff, 1, genBlockSize, stdout); (void)unused; }
+        /* update dict */
+        memcpy(buff, buff + stdBlockSize, stdDictSize);
+    }
+
+    /* cleanup */
+    free(buff);
+}

data/vendor/zstd/programs/datagen.h

@@ -0,0 +1,38 @@
+/*
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+
+#ifndef DATAGEN_H
+#define DATAGEN_H
+
+#include <stddef.h>   /* size_t */
+
+#if defined (__cplusplus)
+extern "C" {
+#endif
+
+void RDG_genStdout(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_genBuffer
+   Generate 'size' bytes of compressible data into 'buffer'.
+   Compressibility can be controlled using 'matchProba', which is floating point value between 0 and 1.
+   'LitProba' is optional, it affect variability of individual bytes. If litProba==0.0, default value will be used.
+   Generated data pattern can be modified using different 'seed'.
+   For a triplet (matchProba, litProba, seed), the function always generate the same content.
+
+   RDG_genStdout
+   Same as RDG_genBuffer, but generates data into stdout
+*/
+
+#if defined (__cplusplus)
+} /* extern "C" */
+#endif
+
+#endif