Merge branch 'dev' into fix1241
This commit is contained in:
cyan4973
commit
aade1e5904
@ -25,7 +25,12 @@ matrix:
|
||||
- env: Cmd='make valgrindinstall && make -C tests clean valgrindTest'
|
||||
|
||||
- env: Cmd='make arminstall && make armfuzz'
|
||||
- env: Cmd='make arminstall && make aarch64fuzz'
|
||||
|
||||
# Following test is disabled, as there is a bug in Travis' ld
|
||||
# preventing aarch64 compilation to complete.
|
||||
# > collect2: error: ld terminated with signal 11 [Segmentation fault], core dumped
|
||||
# to be re-enabled in a few commit, as it's possible that a random code change circumvent the ld bug
|
||||
# - env: Cmd='make arminstall && make aarch64fuzz'
|
||||
|
||||
- env: Cmd='make ppcinstall && make ppcfuzz'
|
||||
- env: Cmd='make ppcinstall && make ppc64fuzz'
|
||||
|
||||
@ -0,0 +1,48 @@
|
||||
ARG :=
|
||||
|
||||
CC ?= gcc
|
||||
CFLAGS ?= -O3
|
||||
INCLUDES := -I ../randomDictBuilder -I ../fastCover -I ../../../programs -I ../../../lib/common -I ../../../lib -I ../../../lib/dictBuilder
|
||||
|
||||
RANDOM_FILE := ../randomDictBuilder/random.c
|
||||
FAST_FILE := ../fastCover/fastCover.c
|
||||
IO_FILE := ../randomDictBuilder/io.c
|
||||
|
||||
all: run clean
|
||||
|
||||
.PHONY: run
|
||||
run: benchmark
|
||||
echo "Benchmarking with $(ARG)"
|
||||
./benchmark $(ARG)
|
||||
|
||||
.PHONY: test
|
||||
test: benchmarkTest clean
|
||||
|
||||
.PHONY: benchmarkTest
|
||||
benchmarkTest: benchmark test.sh
|
||||
sh test.sh
|
||||
|
||||
benchmark: benchmark.o io.o random.o fastCover.o libzstd.a
|
||||
$(CC) $(CFLAGS) benchmark.o io.o random.o fastCover.o libzstd.a -o benchmark
|
||||
|
||||
benchmark.o: benchmark.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c benchmark.c
|
||||
|
||||
random.o: $(RANDOM_FILE)
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c $(RANDOM_FILE)
|
||||
|
||||
fastCover.o: $(FAST_FILE)
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c $(FAST_FILE)
|
||||
|
||||
io.o: $(IO_FILE)
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c $(IO_FILE)
|
||||
|
||||
libzstd.a:
|
||||
$(MAKE) -C ../../../lib libzstd.a
|
||||
mv ../../../lib/libzstd.a .
|
||||
|
||||
.PHONY: clean
|
||||
clean:
|
||||
rm -f *.o benchmark libzstd.a
|
||||
$(MAKE) -C ../../../lib clean
|
||||
echo "Cleaning is completed"
|
||||
@ -0,0 +1,126 @@
|
||||
Benchmarking Dictionary Builder
|
||||
|
||||
### Permitted Argument:
|
||||
Input File/Directory (in=fileName): required; file/directory used to build dictionary; if directory, will operate recursively for files inside directory; can include multiple files/directories, each following "in="
|
||||
|
||||
###Running Test:
|
||||
make test
|
||||
|
||||
###Usage:
|
||||
Benchmark given input files: make ARG= followed by permitted arguments
|
||||
|
||||
### Examples:
|
||||
make ARG="in=../../../lib/dictBuilder in=../../../lib/compress"
|
||||
|
||||
###Benchmarking Result:
|
||||
- First Cover is optimize cover, second Cover uses optimized d and k from first one.
|
||||
- For every f value of fastCover, the first one is optimize fastCover and the second one uses optimized d and k from first one.
|
||||
- Fourth column is chosen d and fifth column is chosen k
|
||||
|
||||
github:
|
||||
NODICT 0.000005 2.999642
|
||||
RANDOM 0.141553 8.786957
|
||||
LEGACY 0.904340 8.989482
|
||||
COVER 53.621302 10.641263 8 1298
|
||||
COVER 4.085037 10.641263 8 1298
|
||||
FAST15 17.636211 10.586461 8 1778
|
||||
FAST15 0.221236 10.586461 8 1778
|
||||
FAST16 18.716259 10.492503 6 1778
|
||||
FAST16 0.251522 10.492503 6 1778
|
||||
FAST17 17.614391 10.611737 8 1778
|
||||
FAST17 0.241011 10.611737 8 1778
|
||||
FAST18 19.926270 10.621586 8 1778
|
||||
FAST18 0.287195 10.621586 8 1778
|
||||
FAST19 19.626808 10.629626 8 1778
|
||||
FAST19 0.340191 10.629626 8 1778
|
||||
FAST20 18.918657 10.610308 8 1778
|
||||
FAST20 0.463307 10.610308 8 1778
|
||||
FAST21 20.502362 10.625733 8 1778
|
||||
FAST21 0.638202 10.625733 8 1778
|
||||
FAST22 22.702695 10.625281 8 1778
|
||||
FAST22 1.353399 10.625281 8 1778
|
||||
FAST23 28.041990 10.602342 8 1778
|
||||
FAST23 3.029502 10.602342 8 1778
|
||||
FAST24 35.662961 10.603379 8 1778
|
||||
FAST24 6.524258 10.603379 8 1778
|
||||
|
||||
hg-commands:
|
||||
NODICT 0.000005 2.425291
|
||||
RANDOM 0.080469 3.489515
|
||||
LEGACY 0.794417 3.911896
|
||||
COVER 54.198788 4.131136 8 386
|
||||
COVER 2.191729 4.131136 8 386
|
||||
FAST15 11.852793 3.903719 6 1106
|
||||
FAST15 0.175406 3.903719 6 1106
|
||||
FAST16 12.863315 4.005077 8 530
|
||||
FAST16 0.158410 4.005077 8 530
|
||||
FAST17 11.977917 4.097811 8 818
|
||||
FAST17 0.162381 4.097811 8 818
|
||||
FAST18 11.749304 4.136081 8 770
|
||||
FAST18 0.173242 4.136081 8 770
|
||||
FAST19 11.905785 4.166021 8 530
|
||||
FAST19 0.186403 4.166021 8 530
|
||||
FAST20 13.293999 4.163740 8 482
|
||||
FAST20 0.241508 4.163740 8 482
|
||||
FAST21 16.623177 4.157057 8 434
|
||||
FAST21 0.372647 4.157057 8 434
|
||||
FAST22 20.918409 4.158195 8 290
|
||||
FAST22 0.570431 4.158195 8 290
|
||||
FAST23 21.762805 4.161450 8 434
|
||||
FAST23 1.162206 4.161450 8 434
|
||||
FAST24 29.133745 4.159658 8 338
|
||||
FAST24 3.054376 4.159658 8 338
|
||||
|
||||
hg-changelog:
|
||||
NODICT 0.000006 1.377613
|
||||
RANDOM 0.601346 2.096785
|
||||
LEGACY 2.544973 2.058273
|
||||
COVER 222.639708 2.188654 8 98
|
||||
COVER 6.072892 2.188654 8 98
|
||||
FAST15 70.394523 2.127194 8 866
|
||||
FAST15 0.899766 2.127194 8 866
|
||||
FAST16 69.845529 2.145401 8 338
|
||||
FAST16 0.881569 2.145401 8 338
|
||||
FAST17 69.382431 2.157544 8 194
|
||||
FAST17 0.943291 2.157544 8 194
|
||||
FAST18 71.348283 2.173127 8 98
|
||||
FAST18 1.034765 2.173127 8 98
|
||||
FAST19 71.380923 2.179527 8 98
|
||||
FAST19 1.254700 2.179527 8 98
|
||||
FAST20 72.802714 2.183233 6 98
|
||||
FAST20 1.368704 2.183233 6 98
|
||||
FAST21 82.042339 2.180920 8 98
|
||||
FAST21 2.213864 2.180920 8 98
|
||||
FAST22 90.666200 2.184297 8 98
|
||||
FAST22 3.590399 2.184297 8 98
|
||||
FAST23 108.926377 2.187666 6 98
|
||||
FAST23 8.723759 2.187666 6 98
|
||||
FAST24 134.296232 2.189889 6 98
|
||||
FAST24 19.396532 2.189889 6 98
|
||||
|
||||
hg-manifest:
|
||||
NODICT 0.000005 1.866385
|
||||
RANDOM 0.982192 2.309485
|
||||
LEGACY 9.507729 2.506775
|
||||
COVER 922.742066 2.582597 8 434
|
||||
COVER 36.500276 2.582597 8 434
|
||||
FAST15 163.886717 2.377689 8 1682
|
||||
FAST15 2.107328 2.377689 8 1682
|
||||
FAST16 152.684592 2.464814 8 1538
|
||||
FAST16 2.157789 2.464814 8 1538
|
||||
FAST17 154.463459 2.539834 6 1826
|
||||
FAST17 2.282455 2.539834 6 1826
|
||||
FAST18 155.540044 2.576924 8 1922
|
||||
FAST18 2.101807 2.576924 8 1922
|
||||
FAST19 152.650343 2.592479 6 290
|
||||
FAST19 2.359461 2.592479 6 290
|
||||
FAST20 174.623634 2.594551 8 194
|
||||
FAST20 2.870022 2.594551 8 194
|
||||
FAST21 219.876653 2.597128 6 194
|
||||
FAST21 4.386269 2.597128 6 194
|
||||
FAST22 247.986803 2.596971 6 386
|
||||
FAST22 6.201144 2.596971 6 386
|
||||
FAST23 276.051806 2.601416 8 194
|
||||
FAST23 11.613477 2.601416 8 194
|
||||
FAST24 328.234024 2.602830 6 194
|
||||
FAST24 26.710364 2.602830 6 194
|
||||
@ -0,0 +1,423 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* strcmp, strlen */
|
||||
#include <errno.h> /* errno */
|
||||
#include <ctype.h>
|
||||
#include <time.h>
|
||||
#include "random.h"
|
||||
#include "fastCover.h"
|
||||
#include "dictBuilder.h"
|
||||
#include "zstd_internal.h" /* includes zstd.h */
|
||||
#include "io.h"
|
||||
#include "util.h"
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
|
||||
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); }
|
||||
|
||||
static const U64 g_refreshRate = SEC_TO_MICRO / 6;
|
||||
static UTIL_time_t g_displayClock = UTIL_TIME_INITIALIZER;
|
||||
|
||||
#define DISPLAYUPDATE(l, ...) { if (displayLevel>=l) { \
|
||||
if ((UTIL_clockSpanMicro(g_displayClock) > g_refreshRate) || (displayLevel>=4)) \
|
||||
{ g_displayClock = UTIL_getTime(); DISPLAY(__VA_ARGS__); \
|
||||
if (displayLevel>=4) fflush(stderr); } } }
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Exceptions
|
||||
***************************************/
|
||||
#ifndef DEBUG
|
||||
# define DEBUG 0
|
||||
#endif
|
||||
#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
|
||||
#define EXM_THROW(error, ...) \
|
||||
{ \
|
||||
DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
|
||||
DISPLAY("Error %i : ", error); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
DISPLAY("\n"); \
|
||||
exit(error); \
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Constants
|
||||
***************************************/
|
||||
static const unsigned g_defaultMaxDictSize = 110 KB;
|
||||
#define DEFAULT_CLEVEL 3
|
||||
#define DEFAULT_DISPLAYLEVEL 2
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Struct
|
||||
***************************************/
|
||||
typedef struct {
|
||||
const void* dictBuffer;
|
||||
size_t dictSize;
|
||||
} dictInfo;
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Dictionary related operations
|
||||
***************************************/
|
||||
/** createDictFromFiles() :
|
||||
* Based on type of param given, train dictionary using the corresponding algorithm
|
||||
* @return dictInfo containing dictionary buffer and dictionary size
|
||||
*/
|
||||
dictInfo* createDictFromFiles(sampleInfo *info, unsigned maxDictSize,
|
||||
ZDICT_random_params_t *randomParams, ZDICT_cover_params_t *coverParams,
|
||||
ZDICT_legacy_params_t *legacyParams, ZDICT_fastCover_params_t *fastParams) {
|
||||
unsigned const displayLevel = randomParams ? randomParams->zParams.notificationLevel :
|
||||
coverParams ? coverParams->zParams.notificationLevel :
|
||||
legacyParams ? legacyParams->zParams.notificationLevel :
|
||||
fastParams ? fastParams->zParams.notificationLevel :
|
||||
DEFAULT_DISPLAYLEVEL; /* no dict */
|
||||
void* const dictBuffer = malloc(maxDictSize);
|
||||
|
||||
dictInfo* dInfo = NULL;
|
||||
|
||||
/* Checks */
|
||||
if (!dictBuffer)
|
||||
EXM_THROW(12, "not enough memory for trainFromFiles"); /* should not happen */
|
||||
|
||||
{ size_t dictSize;
|
||||
if(randomParams) {
|
||||
dictSize = ZDICT_trainFromBuffer_random(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, *randomParams);
|
||||
}else if(coverParams) {
|
||||
dictSize = ZDICT_optimizeTrainFromBuffer_cover(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, coverParams);
|
||||
} else if(legacyParams) {
|
||||
dictSize = ZDICT_trainFromBuffer_legacy(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, *legacyParams);
|
||||
} else if(fastParams) {
|
||||
dictSize = ZDICT_optimizeTrainFromBuffer_fastCover(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, fastParams);
|
||||
} else {
|
||||
dictSize = 0;
|
||||
}
|
||||
if (ZDICT_isError(dictSize)) {
|
||||
DISPLAYLEVEL(1, "dictionary training failed : %s \n", ZDICT_getErrorName(dictSize)); /* should not happen */
|
||||
free(dictBuffer);
|
||||
return dInfo;
|
||||
}
|
||||
dInfo = (dictInfo *)malloc(sizeof(dictInfo));
|
||||
dInfo->dictBuffer = dictBuffer;
|
||||
dInfo->dictSize = dictSize;
|
||||
}
|
||||
return dInfo;
|
||||
}
|
||||
|
||||
|
||||
/** compressWithDict() :
|
||||
* Compress samples from sample buffer given dicionary stored on dictionary buffer and compression level
|
||||
* @return compression ratio
|
||||
*/
|
||||
double compressWithDict(sampleInfo *srcInfo, dictInfo* dInfo, int compressionLevel, int displayLevel) {
|
||||
/* Local variables */
|
||||
size_t totalCompressedSize = 0;
|
||||
size_t totalOriginalSize = 0;
|
||||
const unsigned hasDict = dInfo->dictSize > 0 ? 1 : 0;
|
||||
double cRatio;
|
||||
size_t dstCapacity;
|
||||
int i;
|
||||
|
||||
/* Pointers */
|
||||
ZSTD_CDict *cdict = NULL;
|
||||
ZSTD_CCtx* cctx = NULL;
|
||||
size_t *offsets = NULL;
|
||||
void* dst = NULL;
|
||||
|
||||
/* Allocate dst with enough space to compress the maximum sized sample */
|
||||
{
|
||||
size_t maxSampleSize = 0;
|
||||
for (int i = 0; i < srcInfo->nbSamples; i++) {
|
||||
maxSampleSize = MAX(srcInfo->samplesSizes[i], maxSampleSize);
|
||||
}
|
||||
dstCapacity = ZSTD_compressBound(maxSampleSize);
|
||||
dst = malloc(dstCapacity);
|
||||
}
|
||||
|
||||
/* Calculate offset for each sample */
|
||||
offsets = (size_t *)malloc((srcInfo->nbSamples + 1) * sizeof(size_t));
|
||||
offsets[0] = 0;
|
||||
for (i = 1; i <= srcInfo->nbSamples; i++) {
|
||||
offsets[i] = offsets[i - 1] + srcInfo->samplesSizes[i - 1];
|
||||
}
|
||||
|
||||
/* Create the cctx */
|
||||
cctx = ZSTD_createCCtx();
|
||||
if(!cctx || !dst) {
|
||||
cRatio = -1;
|
||||
goto _cleanup;
|
||||
}
|
||||
|
||||
/* Create CDict if there's a dictionary stored on buffer */
|
||||
if (hasDict) {
|
||||
cdict = ZSTD_createCDict(dInfo->dictBuffer, dInfo->dictSize, compressionLevel);
|
||||
if(!cdict) {
|
||||
cRatio = -1;
|
||||
goto _cleanup;
|
||||
}
|
||||
}
|
||||
|
||||
/* Compress each sample and sum their sizes*/
|
||||
const BYTE *const samples = (const BYTE *)srcInfo->srcBuffer;
|
||||
for (i = 0; i < srcInfo->nbSamples; i++) {
|
||||
size_t compressedSize;
|
||||
if(hasDict) {
|
||||
compressedSize = ZSTD_compress_usingCDict(cctx, dst, dstCapacity, samples + offsets[i], srcInfo->samplesSizes[i], cdict);
|
||||
} else {
|
||||
compressedSize = ZSTD_compressCCtx(cctx, dst, dstCapacity,samples + offsets[i], srcInfo->samplesSizes[i], compressionLevel);
|
||||
}
|
||||
if (ZSTD_isError(compressedSize)) {
|
||||
cRatio = -1;
|
||||
goto _cleanup;
|
||||
}
|
||||
totalCompressedSize += compressedSize;
|
||||
}
|
||||
|
||||
/* Sum orignal sizes */
|
||||
for (i = 0; i<srcInfo->nbSamples; i++) {
|
||||
totalOriginalSize += srcInfo->samplesSizes[i];
|
||||
}
|
||||
|
||||
/* Calculate compression ratio */
|
||||
DISPLAYLEVEL(2, "original size is %lu\n", totalOriginalSize);
|
||||
DISPLAYLEVEL(2, "compressed size is %lu\n", totalCompressedSize);
|
||||
cRatio = (double)totalOriginalSize/(double)totalCompressedSize;
|
||||
|
||||
_cleanup:
|
||||
free(dst);
|
||||
free(offsets);
|
||||
ZSTD_freeCCtx(cctx);
|
||||
ZSTD_freeCDict(cdict);
|
||||
return cRatio;
|
||||
}
|
||||
|
||||
|
||||
/** FreeDictInfo() :
|
||||
* Free memory allocated for dictInfo
|
||||
*/
|
||||
void freeDictInfo(dictInfo* info) {
|
||||
if (!info) return;
|
||||
if (info->dictBuffer) free((void*)(info->dictBuffer));
|
||||
free(info);
|
||||
}
|
||||
|
||||
|
||||
|
||||
/*-********************************************************
|
||||
* Benchmarking functions
|
||||
**********************************************************/
|
||||
/** benchmarkDictBuilder() :
|
||||
* Measure how long a dictionary builder takes and compression ratio with the dictionary built
|
||||
* @return 0 if benchmark successfully, 1 otherwise
|
||||
*/
|
||||
int benchmarkDictBuilder(sampleInfo *srcInfo, unsigned maxDictSize, ZDICT_random_params_t *randomParam,
|
||||
ZDICT_cover_params_t *coverParam, ZDICT_legacy_params_t *legacyParam,
|
||||
ZDICT_fastCover_params_t *fastParam) {
|
||||
/* Local variables */
|
||||
const unsigned displayLevel = randomParam ? randomParam->zParams.notificationLevel :
|
||||
coverParam ? coverParam->zParams.notificationLevel :
|
||||
legacyParam ? legacyParam->zParams.notificationLevel :
|
||||
fastParam ? fastParam->zParams.notificationLevel:
|
||||
DEFAULT_DISPLAYLEVEL; /* no dict */
|
||||
const char* name = randomParam ? "RANDOM" :
|
||||
coverParam ? "COVER" :
|
||||
legacyParam ? "LEGACY" :
|
||||
fastParam ? "FAST":
|
||||
"NODICT"; /* no dict */
|
||||
const unsigned cLevel = randomParam ? randomParam->zParams.compressionLevel :
|
||||
coverParam ? coverParam->zParams.compressionLevel :
|
||||
legacyParam ? legacyParam->zParams.compressionLevel :
|
||||
fastParam ? fastParam->zParams.compressionLevel:
|
||||
DEFAULT_CLEVEL; /* no dict */
|
||||
int result = 0;
|
||||
|
||||
/* Calculate speed */
|
||||
const UTIL_time_t begin = UTIL_getTime();
|
||||
dictInfo* dInfo = createDictFromFiles(srcInfo, maxDictSize, randomParam, coverParam, legacyParam, fastParam);
|
||||
const U64 timeMicro = UTIL_clockSpanMicro(begin);
|
||||
const double timeSec = timeMicro / (double)SEC_TO_MICRO;
|
||||
if (!dInfo) {
|
||||
DISPLAYLEVEL(1, "%s does not train successfully\n", name);
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
DISPLAYLEVEL(1, "%s took %f seconds to execute \n", name, timeSec);
|
||||
|
||||
/* Calculate compression ratio */
|
||||
const double cRatio = compressWithDict(srcInfo, dInfo, cLevel, displayLevel);
|
||||
if (cRatio < 0) {
|
||||
DISPLAYLEVEL(1, "Compressing with %s dictionary does not work\n", name);
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
|
||||
}
|
||||
DISPLAYLEVEL(1, "Compression ratio with %s dictionary is %f\n", name, cRatio);
|
||||
|
||||
_cleanup:
|
||||
freeDictInfo(dInfo);
|
||||
return result;
|
||||
}
|
||||
|
||||
|
||||
|
||||
int main(int argCount, const char* argv[])
|
||||
{
|
||||
const int displayLevel = DEFAULT_DISPLAYLEVEL;
|
||||
const char* programName = argv[0];
|
||||
int result = 0;
|
||||
|
||||
/* Initialize arguments to default values */
|
||||
unsigned k = 200;
|
||||
unsigned d = 8;
|
||||
const unsigned cLevel = DEFAULT_CLEVEL;
|
||||
const unsigned dictID = 0;
|
||||
const unsigned maxDictSize = g_defaultMaxDictSize;
|
||||
|
||||
/* Initialize table to store input files */
|
||||
const char** filenameTable = (const char**)malloc(argCount * sizeof(const char*));
|
||||
unsigned filenameIdx = 0;
|
||||
|
||||
char* fileNamesBuf = NULL;
|
||||
unsigned fileNamesNb = filenameIdx;
|
||||
const int followLinks = 0;
|
||||
const char** extendedFileList = NULL;
|
||||
|
||||
/* Parse arguments */
|
||||
for (int i = 1; i < argCount; i++) {
|
||||
const char* argument = argv[i];
|
||||
if (longCommandWArg(&argument, "in=")) {
|
||||
filenameTable[filenameIdx] = argument;
|
||||
filenameIdx++;
|
||||
continue;
|
||||
}
|
||||
DISPLAYLEVEL(1, "benchmark: Incorrect parameters\n");
|
||||
return 1;
|
||||
}
|
||||
|
||||
/* Get the list of all files recursively (because followLinks==0)*/
|
||||
extendedFileList = UTIL_createFileList(filenameTable, filenameIdx, &fileNamesBuf,
|
||||
&fileNamesNb, followLinks);
|
||||
if (extendedFileList) {
|
||||
unsigned u;
|
||||
for (u=0; u<fileNamesNb; u++) DISPLAYLEVEL(4, "%u %s\n", u, extendedFileList[u]);
|
||||
free((void*)filenameTable);
|
||||
filenameTable = extendedFileList;
|
||||
filenameIdx = fileNamesNb;
|
||||
}
|
||||
|
||||
/* get sampleInfo */
|
||||
size_t blockSize = 0;
|
||||
sampleInfo* srcInfo= getSampleInfo(filenameTable,
|
||||
filenameIdx, blockSize, maxDictSize, displayLevel);
|
||||
|
||||
/* set up zParams */
|
||||
ZDICT_params_t zParams;
|
||||
zParams.compressionLevel = cLevel;
|
||||
zParams.notificationLevel = displayLevel;
|
||||
zParams.dictID = dictID;
|
||||
|
||||
/* with no dict */
|
||||
{
|
||||
const int noDictResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, NULL, NULL, NULL);
|
||||
if(noDictResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
}
|
||||
|
||||
/* for random */
|
||||
{
|
||||
ZDICT_random_params_t randomParam;
|
||||
randomParam.zParams = zParams;
|
||||
randomParam.k = k;
|
||||
const int randomResult = benchmarkDictBuilder(srcInfo, maxDictSize, &randomParam, NULL, NULL, NULL);
|
||||
DISPLAYLEVEL(2, "k=%u\n", randomParam.k);
|
||||
if(randomResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
}
|
||||
|
||||
/* for legacy */
|
||||
{
|
||||
ZDICT_legacy_params_t legacyParam;
|
||||
legacyParam.zParams = zParams;
|
||||
legacyParam.selectivityLevel = 9;
|
||||
const int legacyResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, NULL, &legacyParam, NULL);
|
||||
DISPLAYLEVEL(2, "selectivityLevel=%u\n", legacyParam.selectivityLevel);
|
||||
if(legacyResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
}
|
||||
|
||||
/* for cover */
|
||||
{
|
||||
ZDICT_cover_params_t coverParam;
|
||||
memset(&coverParam, 0, sizeof(coverParam));
|
||||
coverParam.zParams = zParams;
|
||||
coverParam.splitPoint = 1.0;
|
||||
coverParam.steps = 40;
|
||||
coverParam.nbThreads = 1;
|
||||
const int coverOptResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, &coverParam, NULL, NULL);
|
||||
DISPLAYLEVEL(2, "k=%u\nd=%u\nsteps=%u\nsplit=%u\n", coverParam.k, coverParam.d, coverParam.steps, (unsigned)(coverParam.splitPoint * 100));
|
||||
if(coverOptResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
|
||||
const int coverResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, &coverParam, NULL, NULL);
|
||||
DISPLAYLEVEL(2, "k=%u\nd=%u\nsteps=%u\nsplit=%u\n", coverParam.k, coverParam.d, coverParam.steps, (unsigned)(coverParam.splitPoint * 100));
|
||||
if(coverResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/* for fastCover */
|
||||
for (unsigned f = 15; f < 25; f++){
|
||||
DISPLAYLEVEL(2, "current f is %u\n", f);
|
||||
/* for fastCover (optimizing k and d) */
|
||||
ZDICT_fastCover_params_t fastParam;
|
||||
memset(&fastParam, 0, sizeof(fastParam));
|
||||
fastParam.zParams = zParams;
|
||||
fastParam.splitPoint = 1.0;
|
||||
fastParam.f = f;
|
||||
fastParam.steps = 40;
|
||||
fastParam.nbThreads = 1;
|
||||
const int fastOptResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, NULL, NULL, &fastParam);
|
||||
DISPLAYLEVEL(2, "k=%u\nd=%u\nf=%u\nsteps=%u\nsplit=%u\n", fastParam.k, fastParam.d, fastParam.f, fastParam.steps, (unsigned)(fastParam.splitPoint * 100));
|
||||
if(fastOptResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
|
||||
|
||||
/* for fastCover (with k and d provided) */
|
||||
const int fastResult = benchmarkDictBuilder(srcInfo, maxDictSize, NULL, NULL, NULL, &fastParam);
|
||||
DISPLAYLEVEL(2, "k=%u\nd=%u\nf=%u\nsteps=%u\nsplit=%u\n", fastParam.k, fastParam.d, fastParam.f, fastParam.steps, (unsigned)(fastParam.splitPoint * 100));
|
||||
if(fastResult) {
|
||||
result = 1;
|
||||
goto _cleanup;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
|
||||
/* Free allocated memory */
|
||||
_cleanup:
|
||||
UTIL_freeFileList(extendedFileList, fileNamesBuf);
|
||||
freeSampleInfo(srcInfo);
|
||||
return result;
|
||||
}
|
||||
@ -0,0 +1,6 @@
|
||||
/* ZDICT_trainFromBuffer_legacy() :
|
||||
* issue : samplesBuffer need to be followed by a noisy guard band.
|
||||
* work around : duplicate the buffer, and add the noise */
|
||||
size_t ZDICT_trainFromBuffer_legacy(void* dictBuffer, size_t dictBufferCapacity,
|
||||
const void* samplesBuffer, const size_t* samplesSizes, unsigned nbSamples,
|
||||
ZDICT_legacy_params_t params);
|
||||
@ -0,0 +1,2 @@
|
||||
echo "Benchmark with in=../../lib/common"
|
||||
./benchmark in=../../../lib/common
|
||||
54
contrib/experimental_dict_builders/fastCover/Makefile
Normal file
54
contrib/experimental_dict_builders/fastCover/Makefile
Normal file
@ -0,0 +1,54 @@
|
||||
ARG :=
|
||||
|
||||
CC ?= gcc
|
||||
CFLAGS ?= -O3 -g
|
||||
INCLUDES := -I ../../../programs -I ../randomDictBuilder -I ../../../lib/common -I ../../../lib -I ../../../lib/dictBuilder
|
||||
|
||||
IO_FILE := ../randomDictBuilder/io.c
|
||||
|
||||
TEST_INPUT := ../../../lib
|
||||
TEST_OUTPUT := fastCoverDict
|
||||
|
||||
all: main run clean
|
||||
|
||||
.PHONY: test
|
||||
test: main testrun testshell clean
|
||||
|
||||
.PHONY: run
|
||||
run:
|
||||
echo "Building a fastCover dictionary with given arguments"
|
||||
./main $(ARG)
|
||||
|
||||
main: main.o io.o fastCover.o libzstd.a
|
||||
$(CC) $(CFLAGS) main.o io.o fastCover.o libzstd.a -o main
|
||||
|
||||
main.o: main.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c main.c
|
||||
|
||||
fastCover.o: fastCover.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c fastCover.c
|
||||
|
||||
io.o: $(IO_FILE)
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c $(IO_FILE)
|
||||
|
||||
libzstd.a:
|
||||
$(MAKE) MOREFLAGS=-g -C ../../../lib libzstd.a
|
||||
mv ../../../lib/libzstd.a .
|
||||
|
||||
.PHONY: testrun
|
||||
testrun: main
|
||||
echo "Run with $(TEST_INPUT) and $(TEST_OUTPUT) "
|
||||
./main in=$(TEST_INPUT) out=$(TEST_OUTPUT)
|
||||
zstd -be3 -D $(TEST_OUTPUT) -r $(TEST_INPUT) -q
|
||||
rm -f $(TEST_OUTPUT)
|
||||
|
||||
.PHONY: testshell
|
||||
testshell: test.sh
|
||||
sh test.sh
|
||||
echo "Finish running test.sh"
|
||||
|
||||
.PHONY: clean
|
||||
clean:
|
||||
rm -f *.o main libzstd.a
|
||||
$(MAKE) -C ../../../lib clean
|
||||
echo "Cleaning is completed"
|
||||
24
contrib/experimental_dict_builders/fastCover/README.md
Normal file
24
contrib/experimental_dict_builders/fastCover/README.md
Normal file
@ -0,0 +1,24 @@
|
||||
FastCover Dictionary Builder
|
||||
|
||||
### Permitted Arguments:
|
||||
Input File/Directory (in=fileName): required; file/directory used to build dictionary; if directory, will operate recursively for files inside directory; can include multiple files/directories, each following "in="
|
||||
Output Dictionary (out=dictName): if not provided, default to fastCoverDict
|
||||
Dictionary ID (dictID=#): nonnegative number; if not provided, default to 0
|
||||
Maximum Dictionary Size (maxdict=#): positive number; in bytes, if not provided, default to 110KB
|
||||
Size of Selected Segment (k=#): positive number; in bytes; if not provided, default to 200
|
||||
Size of Dmer (d=#): either 6 or 8; if not provided, default to 8
|
||||
Number of steps (steps=#): positive number, if not provided, default to 32
|
||||
Percentage of samples used for training(split=#): positive number; if not provided, default to 100
|
||||
|
||||
|
||||
###Running Test:
|
||||
make test
|
||||
|
||||
|
||||
###Usage:
|
||||
To build a random dictionary with the provided arguments: make ARG= followed by arguments
|
||||
|
||||
|
||||
### Examples:
|
||||
make ARG="in=../../../lib/dictBuilder out=dict100 dictID=520"
|
||||
make ARG="in=../../../lib/dictBuilder in=../../../lib/compress"
|
||||
768
contrib/experimental_dict_builders/fastCover/fastCover.c
Normal file
768
contrib/experimental_dict_builders/fastCover/fastCover.c
Normal file
@ -0,0 +1,768 @@
|
||||
/*-*************************************
|
||||
* Dependencies
|
||||
***************************************/
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* memset */
|
||||
#include <time.h> /* clock */
|
||||
#include "mem.h" /* read */
|
||||
#include "pool.h"
|
||||
#include "threading.h"
|
||||
#include "fastCover.h"
|
||||
#include "zstd_internal.h" /* includes zstd.h */
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Constants
|
||||
***************************************/
|
||||
#define FASTCOVER_MAX_SAMPLES_SIZE (sizeof(size_t) == 8 ? ((U32)-1) : ((U32)1 GB))
|
||||
#define FASTCOVER_MAX_F 32
|
||||
#define DEFAULT_SPLITPOINT 1.0
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
static int g_displayLevel = 2;
|
||||
#define DISPLAY(...) \
|
||||
{ \
|
||||
fprintf(stderr, __VA_ARGS__); \
|
||||
fflush(stderr); \
|
||||
}
|
||||
#define LOCALDISPLAYLEVEL(displayLevel, l, ...) \
|
||||
if (displayLevel >= l) { \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
} /* 0 : no display; 1: errors; 2: default; 3: details; 4: debug */
|
||||
#define DISPLAYLEVEL(l, ...) LOCALDISPLAYLEVEL(g_displayLevel, l, __VA_ARGS__)
|
||||
|
||||
#define LOCALDISPLAYUPDATE(displayLevel, l, ...) \
|
||||
if (displayLevel >= l) { \
|
||||
if ((clock() - g_time > refreshRate) || (displayLevel >= 4)) { \
|
||||
g_time = clock(); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
} \
|
||||
}
|
||||
#define DISPLAYUPDATE(l, ...) LOCALDISPLAYUPDATE(g_displayLevel, l, __VA_ARGS__)
|
||||
static const clock_t refreshRate = CLOCKS_PER_SEC * 15 / 100;
|
||||
static clock_t g_time = 0;
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Hash Functions
|
||||
***************************************/
|
||||
static const U64 prime6bytes = 227718039650203ULL;
|
||||
static size_t ZSTD_hash6(U64 u, U32 h) { return (size_t)(((u << (64-48)) * prime6bytes) >> (64-h)) ; }
|
||||
static size_t ZSTD_hash6Ptr(const void* p, U32 h) { return ZSTD_hash6(MEM_readLE64(p), h); }
|
||||
|
||||
static const U64 prime8bytes = 0xCF1BBCDCB7A56463ULL;
|
||||
static size_t ZSTD_hash8(U64 u, U32 h) { return (size_t)(((u) * prime8bytes) >> (64-h)) ; }
|
||||
static size_t ZSTD_hash8Ptr(const void* p, U32 h) { return ZSTD_hash8(MEM_readLE64(p), h); }
|
||||
|
||||
|
||||
/**
|
||||
* Hash the d-byte value pointed to by p and mod 2^f
|
||||
*/
|
||||
static size_t FASTCOVER_hashPtrToIndex(const void* p, U32 h, unsigned d) {
|
||||
if (d == 6) {
|
||||
return ZSTD_hash6Ptr(p, h) & ((1 << h) - 1);
|
||||
}
|
||||
return ZSTD_hash8Ptr(p, h) & ((1 << h) - 1);
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Context
|
||||
***************************************/
|
||||
typedef struct {
|
||||
const BYTE *samples;
|
||||
size_t *offsets;
|
||||
const size_t *samplesSizes;
|
||||
size_t nbSamples;
|
||||
size_t nbTrainSamples;
|
||||
size_t nbTestSamples;
|
||||
size_t nbDmers;
|
||||
U32 *freqs;
|
||||
U16 *segmentFreqs;
|
||||
unsigned d;
|
||||
} FASTCOVER_ctx_t;
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Helper functions
|
||||
***************************************/
|
||||
/**
|
||||
* Returns the sum of the sample sizes.
|
||||
*/
|
||||
static size_t FASTCOVER_sum(const size_t *samplesSizes, unsigned nbSamples) {
|
||||
size_t sum = 0;
|
||||
unsigned i;
|
||||
for (i = 0; i < nbSamples; ++i) {
|
||||
sum += samplesSizes[i];
|
||||
}
|
||||
return sum;
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* fast functions
|
||||
***************************************/
|
||||
/**
|
||||
* A segment is a range in the source as well as the score of the segment.
|
||||
*/
|
||||
typedef struct {
|
||||
U32 begin;
|
||||
U32 end;
|
||||
U32 score;
|
||||
} FASTCOVER_segment_t;
|
||||
|
||||
|
||||
/**
|
||||
* Selects the best segment in an epoch.
|
||||
* Segments of are scored according to the function:
|
||||
*
|
||||
* Let F(d) be the frequency of all dmers with hash value d.
|
||||
* Let S_i be hash value of the dmer at position i of segment S which has length k.
|
||||
*
|
||||
* Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})
|
||||
*
|
||||
* Once the dmer with hash value d is in the dictionay we set F(d) = F(d)/2.
|
||||
*/
|
||||
static FASTCOVER_segment_t FASTCOVER_selectSegment(const FASTCOVER_ctx_t *ctx,
|
||||
U32 *freqs, U32 begin,U32 end,
|
||||
ZDICT_fastCover_params_t parameters) {
|
||||
/* Constants */
|
||||
const U32 k = parameters.k;
|
||||
const U32 d = parameters.d;
|
||||
const U32 dmersInK = k - d + 1;
|
||||
/* Try each segment (activeSegment) and save the best (bestSegment) */
|
||||
FASTCOVER_segment_t bestSegment = {0, 0, 0};
|
||||
FASTCOVER_segment_t activeSegment;
|
||||
/* Reset the activeDmers in the segment */
|
||||
/* The activeSegment starts at the beginning of the epoch. */
|
||||
activeSegment.begin = begin;
|
||||
activeSegment.end = begin;
|
||||
activeSegment.score = 0;
|
||||
{
|
||||
/* Slide the activeSegment through the whole epoch.
|
||||
* Save the best segment in bestSegment.
|
||||
*/
|
||||
while (activeSegment.end < end) {
|
||||
/* Get hash value of current dmer */
|
||||
const size_t index = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.end, parameters.f, ctx->d);
|
||||
/* Add frequency of this index to score if this is the first occurence of index in active segment */
|
||||
if (ctx->segmentFreqs[index] == 0) {
|
||||
activeSegment.score += freqs[index];
|
||||
}
|
||||
ctx->segmentFreqs[index] += 1;
|
||||
/* Increment end of segment */
|
||||
activeSegment.end += 1;
|
||||
/* If the window is now too large, drop the first position */
|
||||
if (activeSegment.end - activeSegment.begin == dmersInK + 1) {
|
||||
/* Get hash value of the dmer to be eliminated from active segment */
|
||||
const size_t delIndex = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.begin, parameters.f, ctx->d);
|
||||
ctx->segmentFreqs[delIndex] -= 1;
|
||||
/* Subtract frequency of this index from score if this is the last occurrence of this index in active segment */
|
||||
if (ctx->segmentFreqs[delIndex] == 0) {
|
||||
activeSegment.score -= freqs[delIndex];
|
||||
}
|
||||
/* Increment start of segment */
|
||||
activeSegment.begin += 1;
|
||||
}
|
||||
/* If this segment is the best so far save it */
|
||||
if (activeSegment.score > bestSegment.score) {
|
||||
bestSegment = activeSegment;
|
||||
}
|
||||
}
|
||||
/* Zero out rest of segmentFreqs array */
|
||||
while (activeSegment.begin < end) {
|
||||
const size_t delIndex = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.begin, parameters.f, ctx->d);
|
||||
ctx->segmentFreqs[delIndex] -= 1;
|
||||
activeSegment.begin += 1;
|
||||
}
|
||||
}
|
||||
{
|
||||
/* Trim off the zero frequency head and tail from the segment. */
|
||||
U32 newBegin = bestSegment.end;
|
||||
U32 newEnd = bestSegment.begin;
|
||||
U32 pos;
|
||||
for (pos = bestSegment.begin; pos != bestSegment.end; ++pos) {
|
||||
const size_t index = FASTCOVER_hashPtrToIndex(ctx->samples + pos, parameters.f, ctx->d);
|
||||
U32 freq = freqs[index];
|
||||
if (freq != 0) {
|
||||
newBegin = MIN(newBegin, pos);
|
||||
newEnd = pos + 1;
|
||||
}
|
||||
}
|
||||
bestSegment.begin = newBegin;
|
||||
bestSegment.end = newEnd;
|
||||
}
|
||||
{
|
||||
/* Half the frequency of hash value of each dmer covered by the chosen segment. */
|
||||
U32 pos;
|
||||
for (pos = bestSegment.begin; pos != bestSegment.end; ++pos) {
|
||||
const size_t i = FASTCOVER_hashPtrToIndex(ctx->samples + pos, parameters.f, ctx->d);
|
||||
freqs[i] = 0;
|
||||
}
|
||||
}
|
||||
return bestSegment;
|
||||
}
|
||||
|
||||
/**
|
||||
* Check the validity of the parameters.
|
||||
* Returns non-zero if the parameters are valid and 0 otherwise.
|
||||
*/
|
||||
static int FASTCOVER_checkParameters(ZDICT_fastCover_params_t parameters,
|
||||
size_t maxDictSize) {
|
||||
/* k, d, and f are required parameters */
|
||||
if (parameters.d == 0 || parameters.k == 0 || parameters.f == 0) {
|
||||
return 0;
|
||||
}
|
||||
/* d has to be 6 or 8 */
|
||||
if (parameters.d != 6 && parameters.d != 8) {
|
||||
return 0;
|
||||
}
|
||||
/* 0 < f <= FASTCOVER_MAX_F */
|
||||
if (parameters.f > FASTCOVER_MAX_F) {
|
||||
return 0;
|
||||
}
|
||||
/* k <= maxDictSize */
|
||||
if (parameters.k > maxDictSize) {
|
||||
return 0;
|
||||
}
|
||||
/* d <= k */
|
||||
if (parameters.d > parameters.k) {
|
||||
return 0;
|
||||
}
|
||||
/* 0 < splitPoint <= 1 */
|
||||
if (parameters.splitPoint <= 0 || parameters.splitPoint > 1) {
|
||||
return 0;
|
||||
}
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Clean up a context initialized with `FASTCOVER_ctx_init()`.
|
||||
*/
|
||||
static void FASTCOVER_ctx_destroy(FASTCOVER_ctx_t *ctx) {
|
||||
if (!ctx) {
|
||||
return;
|
||||
}
|
||||
if (ctx->segmentFreqs) {
|
||||
free(ctx->segmentFreqs);
|
||||
ctx->segmentFreqs = NULL;
|
||||
}
|
||||
if (ctx->freqs) {
|
||||
free(ctx->freqs);
|
||||
ctx->freqs = NULL;
|
||||
}
|
||||
if (ctx->offsets) {
|
||||
free(ctx->offsets);
|
||||
ctx->offsets = NULL;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Calculate for frequency of hash value of each dmer in ctx->samples
|
||||
*/
|
||||
static void FASTCOVER_computeFrequency(U32 *freqs, unsigned f, FASTCOVER_ctx_t *ctx){
|
||||
/* inCurrSample keeps track of this hash value has already be seen in previous dmers in the same sample*/
|
||||
BYTE* inCurrSample = (BYTE *)malloc((1 << f) * sizeof(BYTE));
|
||||
size_t start; /* start of current dmer */
|
||||
for (unsigned i = 0; i < ctx->nbTrainSamples; i++) {
|
||||
memset(inCurrSample, 0, (1 << f) * sizeof(*inCurrSample)); /* Reset inCurrSample for each sample */
|
||||
size_t currSampleStart = ctx->offsets[i];
|
||||
size_t currSampleEnd = ctx->offsets[i+1];
|
||||
start = currSampleStart;
|
||||
while (start + ctx->d <= currSampleEnd) {
|
||||
const size_t dmerIndex = FASTCOVER_hashPtrToIndex(ctx->samples + start, f, ctx->d);
|
||||
/* if no dmer with same hash value has been seen in current sample */
|
||||
if (inCurrSample[dmerIndex] == 0) {
|
||||
inCurrSample[dmerIndex]++;
|
||||
freqs[dmerIndex]++;
|
||||
}
|
||||
start++;
|
||||
}
|
||||
}
|
||||
free(inCurrSample);
|
||||
}
|
||||
|
||||
/**
|
||||
* Prepare a context for dictionary building.
|
||||
* The context is only dependent on the parameter `d` and can used multiple
|
||||
* times.
|
||||
* Returns 1 on success or zero on error.
|
||||
* The context must be destroyed with `FASTCOVER_ctx_destroy()`.
|
||||
*/
|
||||
static int FASTCOVER_ctx_init(FASTCOVER_ctx_t *ctx, const void *samplesBuffer,
|
||||
const size_t *samplesSizes, unsigned nbSamples,
|
||||
unsigned d, double splitPoint, unsigned f) {
|
||||
const BYTE *const samples = (const BYTE *)samplesBuffer;
|
||||
const size_t totalSamplesSize = FASTCOVER_sum(samplesSizes, nbSamples);
|
||||
/* Split samples into testing and training sets */
|
||||
const unsigned nbTrainSamples = splitPoint < 1.0 ? (unsigned)((double)nbSamples * splitPoint) : nbSamples;
|
||||
const unsigned nbTestSamples = splitPoint < 1.0 ? nbSamples - nbTrainSamples : nbSamples;
|
||||
const size_t trainingSamplesSize = splitPoint < 1.0 ? FASTCOVER_sum(samplesSizes, nbTrainSamples) : totalSamplesSize;
|
||||
const size_t testSamplesSize = splitPoint < 1.0 ? FASTCOVER_sum(samplesSizes + nbTrainSamples, nbTestSamples) : totalSamplesSize;
|
||||
/* Checks */
|
||||
if (totalSamplesSize < MAX(d, sizeof(U64)) ||
|
||||
totalSamplesSize >= (size_t)FASTCOVER_MAX_SAMPLES_SIZE) {
|
||||
DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",
|
||||
(U32)(totalSamplesSize>>20), (FASTCOVER_MAX_SAMPLES_SIZE >> 20));
|
||||
return 0;
|
||||
}
|
||||
/* Check if there are at least 5 training samples */
|
||||
if (nbTrainSamples < 5) {
|
||||
DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid.", nbTrainSamples);
|
||||
return 0;
|
||||
}
|
||||
/* Check if there's testing sample */
|
||||
if (nbTestSamples < 1) {
|
||||
DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.", nbTestSamples);
|
||||
return 0;
|
||||
}
|
||||
/* Zero the context */
|
||||
memset(ctx, 0, sizeof(*ctx));
|
||||
DISPLAYLEVEL(2, "Training on %u samples of total size %u\n", nbTrainSamples,
|
||||
(U32)trainingSamplesSize);
|
||||
DISPLAYLEVEL(2, "Testing on %u samples of total size %u\n", nbTestSamples,
|
||||
(U32)testSamplesSize);
|
||||
|
||||
ctx->samples = samples;
|
||||
ctx->samplesSizes = samplesSizes;
|
||||
ctx->nbSamples = nbSamples;
|
||||
ctx->nbTrainSamples = nbTrainSamples;
|
||||
ctx->nbTestSamples = nbTestSamples;
|
||||
ctx->nbDmers = trainingSamplesSize - d + 1;
|
||||
ctx->d = d;
|
||||
|
||||
/* The offsets of each file */
|
||||
ctx->offsets = (size_t *)malloc((nbSamples + 1) * sizeof(size_t));
|
||||
if (!ctx->offsets) {
|
||||
DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n");
|
||||
FASTCOVER_ctx_destroy(ctx);
|
||||
return 0;
|
||||
}
|
||||
|
||||
/* Fill offsets from the samplesSizes */
|
||||
{
|
||||
U32 i;
|
||||
ctx->offsets[0] = 0;
|
||||
for (i = 1; i <= nbSamples; ++i) {
|
||||
ctx->offsets[i] = ctx->offsets[i - 1] + samplesSizes[i - 1];
|
||||
}
|
||||
}
|
||||
|
||||
/* Initialize frequency array of size 2^f */
|
||||
ctx->freqs = (U32 *)calloc((1 << f), sizeof(U32));
|
||||
ctx->segmentFreqs = (U16 *)calloc((1 << f), sizeof(U16));
|
||||
DISPLAYLEVEL(2, "Computing frequencies\n");
|
||||
FASTCOVER_computeFrequency(ctx->freqs, f, ctx);
|
||||
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Given the prepared context build the dictionary.
|
||||
*/
|
||||
static size_t FASTCOVER_buildDictionary(const FASTCOVER_ctx_t *ctx, U32 *freqs,
|
||||
void *dictBuffer,
|
||||
size_t dictBufferCapacity,
|
||||
ZDICT_fastCover_params_t parameters){
|
||||
BYTE *const dict = (BYTE *)dictBuffer;
|
||||
size_t tail = dictBufferCapacity;
|
||||
/* Divide the data up into epochs of equal size.
|
||||
* We will select at least one segment from each epoch.
|
||||
*/
|
||||
const U32 epochs = MAX(1, (U32)(dictBufferCapacity / parameters.k));
|
||||
const U32 epochSize = (U32)(ctx->nbDmers / epochs);
|
||||
size_t epoch;
|
||||
DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n", epochs,
|
||||
epochSize);
|
||||
/* Loop through the epochs until there are no more segments or the dictionary
|
||||
* is full.
|
||||
*/
|
||||
for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) {
|
||||
const U32 epochBegin = (U32)(epoch * epochSize);
|
||||
const U32 epochEnd = epochBegin + epochSize;
|
||||
size_t segmentSize;
|
||||
/* Select a segment */
|
||||
FASTCOVER_segment_t segment = FASTCOVER_selectSegment(
|
||||
ctx, freqs, epochBegin, epochEnd, parameters);
|
||||
|
||||
/* If the segment covers no dmers, then we are out of content */
|
||||
if (segment.score == 0) {
|
||||
break;
|
||||
}
|
||||
|
||||
/* Trim the segment if necessary and if it is too small then we are done */
|
||||
segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
|
||||
if (segmentSize < parameters.d) {
|
||||
break;
|
||||
}
|
||||
|
||||
/* We fill the dictionary from the back to allow the best segments to be
|
||||
* referenced with the smallest offsets.
|
||||
*/
|
||||
tail -= segmentSize;
|
||||
memcpy(dict + tail, ctx->samples + segment.begin, segmentSize);
|
||||
DISPLAYUPDATE(
|
||||
2, "\r%u%% ",
|
||||
(U32)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));
|
||||
}
|
||||
DISPLAYLEVEL(2, "\r%79s\r", "");
|
||||
return tail;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* FASTCOVER_best_t is used for two purposes:
|
||||
* 1. Synchronizing threads.
|
||||
* 2. Saving the best parameters and dictionary.
|
||||
*
|
||||
* All of the methods except FASTCOVER_best_init() are thread safe if zstd is
|
||||
* compiled with multithreaded support.
|
||||
*/
|
||||
typedef struct fast_best_s {
|
||||
ZSTD_pthread_mutex_t mutex;
|
||||
ZSTD_pthread_cond_t cond;
|
||||
size_t liveJobs;
|
||||
void *dict;
|
||||
size_t dictSize;
|
||||
ZDICT_fastCover_params_t parameters;
|
||||
size_t compressedSize;
|
||||
} FASTCOVER_best_t;
|
||||
|
||||
/**
|
||||
* Initialize the `FASTCOVER_best_t`.
|
||||
*/
|
||||
static void FASTCOVER_best_init(FASTCOVER_best_t *best) {
|
||||
if (best==NULL) return; /* compatible with init on NULL */
|
||||
(void)ZSTD_pthread_mutex_init(&best->mutex, NULL);
|
||||
(void)ZSTD_pthread_cond_init(&best->cond, NULL);
|
||||
best->liveJobs = 0;
|
||||
best->dict = NULL;
|
||||
best->dictSize = 0;
|
||||
best->compressedSize = (size_t)-1;
|
||||
memset(&best->parameters, 0, sizeof(best->parameters));
|
||||
}
|
||||
|
||||
/**
|
||||
* Wait until liveJobs == 0.
|
||||
*/
|
||||
static void FASTCOVER_best_wait(FASTCOVER_best_t *best) {
|
||||
if (!best) {
|
||||
return;
|
||||
}
|
||||
ZSTD_pthread_mutex_lock(&best->mutex);
|
||||
while (best->liveJobs != 0) {
|
||||
ZSTD_pthread_cond_wait(&best->cond, &best->mutex);
|
||||
}
|
||||
ZSTD_pthread_mutex_unlock(&best->mutex);
|
||||
}
|
||||
|
||||
/**
|
||||
* Call FASTCOVER_best_wait() and then destroy the FASTCOVER_best_t.
|
||||
*/
|
||||
static void FASTCOVER_best_destroy(FASTCOVER_best_t *best) {
|
||||
if (!best) {
|
||||
return;
|
||||
}
|
||||
FASTCOVER_best_wait(best);
|
||||
if (best->dict) {
|
||||
free(best->dict);
|
||||
}
|
||||
ZSTD_pthread_mutex_destroy(&best->mutex);
|
||||
ZSTD_pthread_cond_destroy(&best->cond);
|
||||
}
|
||||
|
||||
/**
|
||||
* Called when a thread is about to be launched.
|
||||
* Increments liveJobs.
|
||||
*/
|
||||
static void FASTCOVER_best_start(FASTCOVER_best_t *best) {
|
||||
if (!best) {
|
||||
return;
|
||||
}
|
||||
ZSTD_pthread_mutex_lock(&best->mutex);
|
||||
++best->liveJobs;
|
||||
ZSTD_pthread_mutex_unlock(&best->mutex);
|
||||
}
|
||||
|
||||
/**
|
||||
* Called when a thread finishes executing, both on error or success.
|
||||
* Decrements liveJobs and signals any waiting threads if liveJobs == 0.
|
||||
* If this dictionary is the best so far save it and its parameters.
|
||||
*/
|
||||
static void FASTCOVER_best_finish(FASTCOVER_best_t *best, size_t compressedSize,
|
||||
ZDICT_fastCover_params_t parameters, void *dict,
|
||||
size_t dictSize) {
|
||||
if (!best) {
|
||||
return;
|
||||
}
|
||||
{
|
||||
size_t liveJobs;
|
||||
ZSTD_pthread_mutex_lock(&best->mutex);
|
||||
--best->liveJobs;
|
||||
liveJobs = best->liveJobs;
|
||||
/* If the new dictionary is better */
|
||||
if (compressedSize < best->compressedSize) {
|
||||
/* Allocate space if necessary */
|
||||
if (!best->dict || best->dictSize < dictSize) {
|
||||
if (best->dict) {
|
||||
free(best->dict);
|
||||
}
|
||||
best->dict = malloc(dictSize);
|
||||
if (!best->dict) {
|
||||
best->compressedSize = ERROR(GENERIC);
|
||||
best->dictSize = 0;
|
||||
return;
|
||||
}
|
||||
}
|
||||
/* Save the dictionary, parameters, and size */
|
||||
memcpy(best->dict, dict, dictSize);
|
||||
best->dictSize = dictSize;
|
||||
best->parameters = parameters;
|
||||
best->compressedSize = compressedSize;
|
||||
}
|
||||
ZSTD_pthread_mutex_unlock(&best->mutex);
|
||||
if (liveJobs == 0) {
|
||||
ZSTD_pthread_cond_broadcast(&best->cond);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* Parameters for FASTCOVER_tryParameters().
|
||||
*/
|
||||
typedef struct FASTCOVER_tryParameters_data_s {
|
||||
const FASTCOVER_ctx_t *ctx;
|
||||
FASTCOVER_best_t *best;
|
||||
size_t dictBufferCapacity;
|
||||
ZDICT_fastCover_params_t parameters;
|
||||
} FASTCOVER_tryParameters_data_t;
|
||||
|
||||
/**
|
||||
* Tries a set of parameters and updates the FASTCOVER_best_t with the results.
|
||||
* This function is thread safe if zstd is compiled with multithreaded support.
|
||||
* It takes its parameters as an *OWNING* opaque pointer to support threading.
|
||||
*/
|
||||
static void FASTCOVER_tryParameters(void *opaque) {
|
||||
/* Save parameters as local variables */
|
||||
FASTCOVER_tryParameters_data_t *const data = (FASTCOVER_tryParameters_data_t *)opaque;
|
||||
const FASTCOVER_ctx_t *const ctx = data->ctx;
|
||||
const ZDICT_fastCover_params_t parameters = data->parameters;
|
||||
size_t dictBufferCapacity = data->dictBufferCapacity;
|
||||
size_t totalCompressedSize = ERROR(GENERIC);
|
||||
/* Allocate space for hash table, dict, and freqs */
|
||||
BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
|
||||
U32 *freqs = (U32*) malloc((1 << parameters.f) * sizeof(U32));
|
||||
if (!dict || !freqs) {
|
||||
DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");
|
||||
goto _cleanup;
|
||||
}
|
||||
/* Copy the frequencies because we need to modify them */
|
||||
memcpy(freqs, ctx->freqs, (1 << parameters.f) * sizeof(U32));
|
||||
/* Build the dictionary */
|
||||
{
|
||||
const size_t tail = FASTCOVER_buildDictionary(ctx, freqs, dict,
|
||||
dictBufferCapacity, parameters);
|
||||
|
||||
dictBufferCapacity = ZDICT_finalizeDictionary(
|
||||
dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
|
||||
ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples,
|
||||
parameters.zParams);
|
||||
if (ZDICT_isError(dictBufferCapacity)) {
|
||||
DISPLAYLEVEL(1, "Failed to finalize dictionary\n");
|
||||
goto _cleanup;
|
||||
}
|
||||
}
|
||||
/* Check total compressed size */
|
||||
{
|
||||
/* Pointers */
|
||||
ZSTD_CCtx *cctx;
|
||||
ZSTD_CDict *cdict;
|
||||
void *dst;
|
||||
/* Local variables */
|
||||
size_t dstCapacity;
|
||||
size_t i;
|
||||
/* Allocate dst with enough space to compress the maximum sized sample */
|
||||
{
|
||||
size_t maxSampleSize = 0;
|
||||
i = parameters.splitPoint < 1.0 ? ctx->nbTrainSamples : 0;
|
||||
for (; i < ctx->nbSamples; ++i) {
|
||||
maxSampleSize = MAX(ctx->samplesSizes[i], maxSampleSize);
|
||||
}
|
||||
dstCapacity = ZSTD_compressBound(maxSampleSize);
|
||||
dst = malloc(dstCapacity);
|
||||
}
|
||||
/* Create the cctx and cdict */
|
||||
cctx = ZSTD_createCCtx();
|
||||
cdict = ZSTD_createCDict(dict, dictBufferCapacity,
|
||||
parameters.zParams.compressionLevel);
|
||||
if (!dst || !cctx || !cdict) {
|
||||
goto _compressCleanup;
|
||||
}
|
||||
/* Compress each sample and sum their sizes (or error) */
|
||||
totalCompressedSize = dictBufferCapacity;
|
||||
i = parameters.splitPoint < 1.0 ? ctx->nbTrainSamples : 0;
|
||||
for (; i < ctx->nbSamples; ++i) {
|
||||
const size_t size = ZSTD_compress_usingCDict(
|
||||
cctx, dst, dstCapacity, ctx->samples + ctx->offsets[i],
|
||||
ctx->samplesSizes[i], cdict);
|
||||
if (ZSTD_isError(size)) {
|
||||
totalCompressedSize = ERROR(GENERIC);
|
||||
goto _compressCleanup;
|
||||
}
|
||||
totalCompressedSize += size;
|
||||
}
|
||||
_compressCleanup:
|
||||
ZSTD_freeCCtx(cctx);
|
||||
ZSTD_freeCDict(cdict);
|
||||
if (dst) {
|
||||
free(dst);
|
||||
}
|
||||
}
|
||||
|
||||
_cleanup:
|
||||
FASTCOVER_best_finish(data->best, totalCompressedSize, parameters, dict,
|
||||
dictBufferCapacity);
|
||||
free(data);
|
||||
if (dict) {
|
||||
free(dict);
|
||||
}
|
||||
if (freqs) {
|
||||
free(freqs);
|
||||
}
|
||||
}
|
||||
|
||||
ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(
|
||||
void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,
|
||||
const size_t *samplesSizes, unsigned nbSamples,
|
||||
ZDICT_fastCover_params_t *parameters) {
|
||||
/* constants */
|
||||
const unsigned nbThreads = parameters->nbThreads;
|
||||
const double splitPoint =
|
||||
parameters->splitPoint <= 0.0 ? DEFAULT_SPLITPOINT : parameters->splitPoint;
|
||||
const unsigned kMinD = parameters->d == 0 ? 6 : parameters->d;
|
||||
const unsigned kMaxD = parameters->d == 0 ? 8 : parameters->d;
|
||||
const unsigned kMinK = parameters->k == 0 ? 50 : parameters->k;
|
||||
const unsigned kMaxK = parameters->k == 0 ? 2000 : parameters->k;
|
||||
const unsigned kSteps = parameters->steps == 0 ? 40 : parameters->steps;
|
||||
const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);
|
||||
const unsigned kIterations =
|
||||
(1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
|
||||
const unsigned f = parameters->f == 0 ? 23 : parameters->f;
|
||||
|
||||
/* Local variables */
|
||||
const int displayLevel = parameters->zParams.notificationLevel;
|
||||
unsigned iteration = 1;
|
||||
unsigned d;
|
||||
unsigned k;
|
||||
FASTCOVER_best_t best;
|
||||
POOL_ctx *pool = NULL;
|
||||
|
||||
/* Checks */
|
||||
if (splitPoint <= 0 || splitPoint > 1) {
|
||||
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
if (kMinK < kMaxD || kMaxK < kMinK) {
|
||||
LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
if (nbSamples == 0) {
|
||||
DISPLAYLEVEL(1, "fast must have at least one input file\n");
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
|
||||
DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
|
||||
ZDICT_DICTSIZE_MIN);
|
||||
return ERROR(dstSize_tooSmall);
|
||||
}
|
||||
if (nbThreads > 1) {
|
||||
pool = POOL_create(nbThreads, 1);
|
||||
if (!pool) {
|
||||
return ERROR(memory_allocation);
|
||||
}
|
||||
}
|
||||
/* Initialization */
|
||||
FASTCOVER_best_init(&best);
|
||||
/* Turn down global display level to clean up display at level 2 and below */
|
||||
g_displayLevel = displayLevel == 0 ? 0 : displayLevel - 1;
|
||||
/* Loop through d first because each new value needs a new context */
|
||||
LOCALDISPLAYLEVEL(displayLevel, 2, "Trying %u different sets of parameters\n",
|
||||
kIterations);
|
||||
for (d = kMinD; d <= kMaxD; d += 2) {
|
||||
/* Initialize the context for this value of d */
|
||||
FASTCOVER_ctx_t ctx;
|
||||
LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
|
||||
if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f)) {
|
||||
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
|
||||
FASTCOVER_best_destroy(&best);
|
||||
POOL_free(pool);
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
/* Loop through k reusing the same context */
|
||||
for (k = kMinK; k <= kMaxK; k += kStepSize) {
|
||||
/* Prepare the arguments */
|
||||
FASTCOVER_tryParameters_data_t *data = (FASTCOVER_tryParameters_data_t *)malloc(
|
||||
sizeof(FASTCOVER_tryParameters_data_t));
|
||||
LOCALDISPLAYLEVEL(displayLevel, 3, "k=%u\n", k);
|
||||
if (!data) {
|
||||
LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to allocate parameters\n");
|
||||
FASTCOVER_best_destroy(&best);
|
||||
FASTCOVER_ctx_destroy(&ctx);
|
||||
POOL_free(pool);
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
data->ctx = &ctx;
|
||||
data->best = &best;
|
||||
data->dictBufferCapacity = dictBufferCapacity;
|
||||
data->parameters = *parameters;
|
||||
data->parameters.k = k;
|
||||
data->parameters.d = d;
|
||||
data->parameters.f = f;
|
||||
data->parameters.splitPoint = splitPoint;
|
||||
data->parameters.steps = kSteps;
|
||||
data->parameters.zParams.notificationLevel = g_displayLevel;
|
||||
/* Check the parameters */
|
||||
if (!FASTCOVER_checkParameters(data->parameters, dictBufferCapacity)) {
|
||||
DISPLAYLEVEL(1, "fastCover parameters incorrect\n");
|
||||
free(data);
|
||||
continue;
|
||||
}
|
||||
/* Call the function and pass ownership of data to it */
|
||||
FASTCOVER_best_start(&best);
|
||||
if (pool) {
|
||||
POOL_add(pool, &FASTCOVER_tryParameters, data);
|
||||
} else {
|
||||
FASTCOVER_tryParameters(data);
|
||||
}
|
||||
/* Print status */
|
||||
LOCALDISPLAYUPDATE(displayLevel, 2, "\r%u%% ",
|
||||
(U32)((iteration * 100) / kIterations));
|
||||
++iteration;
|
||||
}
|
||||
FASTCOVER_best_wait(&best);
|
||||
FASTCOVER_ctx_destroy(&ctx);
|
||||
}
|
||||
LOCALDISPLAYLEVEL(displayLevel, 2, "\r%79s\r", "");
|
||||
/* Fill the output buffer and parameters with output of the best parameters */
|
||||
{
|
||||
const size_t dictSize = best.dictSize;
|
||||
if (ZSTD_isError(best.compressedSize)) {
|
||||
const size_t compressedSize = best.compressedSize;
|
||||
FASTCOVER_best_destroy(&best);
|
||||
POOL_free(pool);
|
||||
return compressedSize;
|
||||
}
|
||||
*parameters = best.parameters;
|
||||
memcpy(dictBuffer, best.dict, dictSize);
|
||||
FASTCOVER_best_destroy(&best);
|
||||
POOL_free(pool);
|
||||
return dictSize;
|
||||
}
|
||||
|
||||
}
|
||||
47
contrib/experimental_dict_builders/fastCover/fastCover.h
Normal file
47
contrib/experimental_dict_builders/fastCover/fastCover.h
Normal file
@ -0,0 +1,47 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* memset */
|
||||
#include <time.h> /* clock */
|
||||
#include "mem.h" /* read */
|
||||
#include "pool.h"
|
||||
#include "threading.h"
|
||||
#include "zstd_internal.h" /* includes zstd.h */
|
||||
#ifndef ZDICT_STATIC_LINKING_ONLY
|
||||
#define ZDICT_STATIC_LINKING_ONLY
|
||||
#endif
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
typedef struct {
|
||||
unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+] */
|
||||
unsigned d; /* dmer size : constraint: 0 < d <= k : Reasonable range [6, 16] */
|
||||
unsigned f; /* log of size of frequency array */
|
||||
unsigned steps; /* Number of steps : Only used for optimization : 0 means default (32) : Higher means more parameters checked */
|
||||
unsigned nbThreads; /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
|
||||
double splitPoint; /* Percentage of samples used for training: the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (1.0), 1.0 when all samples are used for both training and testing */
|
||||
ZDICT_params_t zParams;
|
||||
} ZDICT_fastCover_params_t;
|
||||
|
||||
|
||||
|
||||
/*! ZDICT_optimizeTrainFromBuffer_fastCover():
|
||||
* Train a dictionary from an array of samples using a modified version of the COVER algorithm.
|
||||
* Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
|
||||
* supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
|
||||
* The resulting dictionary will be saved into `dictBuffer`.
|
||||
* All of the parameters except for f are optional.
|
||||
* If d is non-zero then we don't check multiple values of d, otherwise we check d = {6, 8, 10, 12, 14, 16}.
|
||||
* if steps is zero it defaults to its default value.
|
||||
* If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [16, 2048].
|
||||
*
|
||||
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
|
||||
* or an error code, which can be tested with ZDICT_isError().
|
||||
* On success `*parameters` contains the parameters selected.
|
||||
*/
|
||||
ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(
|
||||
void *dictBuffer, size_t dictBufferCapacity, const void *samplesBuffer,
|
||||
const size_t *samplesSizes, unsigned nbSamples,
|
||||
ZDICT_fastCover_params_t *parameters);
|
||||
177
contrib/experimental_dict_builders/fastCover/main.c
Normal file
177
contrib/experimental_dict_builders/fastCover/main.c
Normal file
@ -0,0 +1,177 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* strcmp, strlen */
|
||||
#include <errno.h> /* errno */
|
||||
#include <ctype.h>
|
||||
#include "fastCover.h"
|
||||
#include "io.h"
|
||||
#include "util.h"
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
|
||||
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); }
|
||||
|
||||
static const U64 g_refreshRate = SEC_TO_MICRO / 6;
|
||||
static UTIL_time_t g_displayClock = UTIL_TIME_INITIALIZER;
|
||||
|
||||
#define DISPLAYUPDATE(l, ...) { if (displayLevel>=l) { \
|
||||
if ((UTIL_clockSpanMicro(g_displayClock) > g_refreshRate) || (displayLevel>=4)) \
|
||||
{ g_displayClock = UTIL_getTime(); DISPLAY(__VA_ARGS__); \
|
||||
if (displayLevel>=4) fflush(stderr); } } }
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Exceptions
|
||||
***************************************/
|
||||
#ifndef DEBUG
|
||||
# define DEBUG 0
|
||||
#endif
|
||||
#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
|
||||
#define EXM_THROW(error, ...) \
|
||||
{ \
|
||||
DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
|
||||
DISPLAY("Error %i : ", error); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
DISPLAY("\n"); \
|
||||
exit(error); \
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Constants
|
||||
***************************************/
|
||||
static const unsigned g_defaultMaxDictSize = 110 KB;
|
||||
#define DEFAULT_CLEVEL 3
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* FASTCOVER
|
||||
***************************************/
|
||||
int FASTCOVER_trainFromFiles(const char* dictFileName, sampleInfo *info,
|
||||
unsigned maxDictSize,
|
||||
ZDICT_fastCover_params_t *params) {
|
||||
unsigned const displayLevel = params->zParams.notificationLevel;
|
||||
void* const dictBuffer = malloc(maxDictSize);
|
||||
|
||||
int result = 0;
|
||||
|
||||
/* Checks */
|
||||
if (!dictBuffer)
|
||||
EXM_THROW(12, "not enough memory for trainFromFiles"); /* should not happen */
|
||||
|
||||
{ size_t dictSize;
|
||||
dictSize = ZDICT_optimizeTrainFromBuffer_fastCover(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, params);
|
||||
DISPLAYLEVEL(2, "k=%u\nd=%u\nf=%u\nsteps=%u\nsplit=%u\n", params->k, params->d, params->f, params->steps, (unsigned)(params->splitPoint*100));
|
||||
if (ZDICT_isError(dictSize)) {
|
||||
DISPLAYLEVEL(1, "dictionary training failed : %s \n", ZDICT_getErrorName(dictSize)); /* should not happen */
|
||||
result = 1;
|
||||
goto _done;
|
||||
}
|
||||
/* save dict */
|
||||
DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName);
|
||||
saveDict(dictFileName, dictBuffer, dictSize);
|
||||
}
|
||||
|
||||
/* clean up */
|
||||
_done:
|
||||
free(dictBuffer);
|
||||
return result;
|
||||
}
|
||||
|
||||
|
||||
|
||||
int main(int argCount, const char* argv[])
|
||||
{
|
||||
int displayLevel = 2;
|
||||
const char* programName = argv[0];
|
||||
int operationResult = 0;
|
||||
|
||||
/* Initialize arguments to default values */
|
||||
unsigned k = 200;
|
||||
unsigned d = 8;
|
||||
unsigned f = 23;
|
||||
unsigned steps = 32;
|
||||
unsigned nbThreads = 1;
|
||||
unsigned split = 100;
|
||||
const char* outputFile = "fastCoverDict";
|
||||
unsigned dictID = 0;
|
||||
unsigned maxDictSize = g_defaultMaxDictSize;
|
||||
|
||||
/* Initialize table to store input files */
|
||||
const char** filenameTable = (const char**)malloc(argCount * sizeof(const char*));
|
||||
unsigned filenameIdx = 0;
|
||||
|
||||
char* fileNamesBuf = NULL;
|
||||
unsigned fileNamesNb = filenameIdx;
|
||||
int followLinks = 0; /* follow directory recursively */
|
||||
const char** extendedFileList = NULL;
|
||||
|
||||
/* Parse arguments */
|
||||
for (int i = 1; i < argCount; i++) {
|
||||
const char* argument = argv[i];
|
||||
if (longCommandWArg(&argument, "k=")) { k = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "d=")) { d = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "f=")) { f = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "steps=")) { steps = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "split=")) { split = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "dictID=")) { dictID = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "maxdict=")) { maxDictSize = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "in=")) {
|
||||
filenameTable[filenameIdx] = argument;
|
||||
filenameIdx++;
|
||||
continue;
|
||||
}
|
||||
if (longCommandWArg(&argument, "out=")) {
|
||||
outputFile = argument;
|
||||
continue;
|
||||
}
|
||||
DISPLAYLEVEL(1, "Incorrect parameters\n");
|
||||
operationResult = 1;
|
||||
return operationResult;
|
||||
}
|
||||
|
||||
/* Get the list of all files recursively (because followLinks==0)*/
|
||||
extendedFileList = UTIL_createFileList(filenameTable, filenameIdx, &fileNamesBuf,
|
||||
&fileNamesNb, followLinks);
|
||||
if (extendedFileList) {
|
||||
unsigned u;
|
||||
for (u=0; u<fileNamesNb; u++) DISPLAYLEVEL(4, "%u %s\n", u, extendedFileList[u]);
|
||||
free((void*)filenameTable);
|
||||
filenameTable = extendedFileList;
|
||||
filenameIdx = fileNamesNb;
|
||||
}
|
||||
|
||||
size_t blockSize = 0;
|
||||
|
||||
/* Set up zParams */
|
||||
ZDICT_params_t zParams;
|
||||
zParams.compressionLevel = DEFAULT_CLEVEL;
|
||||
zParams.notificationLevel = displayLevel;
|
||||
zParams.dictID = dictID;
|
||||
|
||||
/* Set up fastCover params */
|
||||
ZDICT_fastCover_params_t params;
|
||||
params.zParams = zParams;
|
||||
params.k = k;
|
||||
params.d = d;
|
||||
params.f = f;
|
||||
params.steps = steps;
|
||||
params.nbThreads = nbThreads;
|
||||
params.splitPoint = (double)split/100;
|
||||
|
||||
/* Build dictionary */
|
||||
sampleInfo* info = getSampleInfo(filenameTable,
|
||||
filenameIdx, blockSize, maxDictSize, zParams.notificationLevel);
|
||||
operationResult = FASTCOVER_trainFromFiles(outputFile, info, maxDictSize, ¶ms);
|
||||
|
||||
/* Free allocated memory */
|
||||
UTIL_freeFileList(extendedFileList, fileNamesBuf);
|
||||
freeSampleInfo(info);
|
||||
|
||||
return operationResult;
|
||||
}
|
||||
15
contrib/experimental_dict_builders/fastCover/test.sh
Normal file
15
contrib/experimental_dict_builders/fastCover/test.sh
Normal file
@ -0,0 +1,15 @@
|
||||
echo "Building fastCover dictionary with in=../../lib/common k=200 f=20 out=dict1"
|
||||
./main in=../../../lib/common k=200 f=20 out=dict1
|
||||
zstd -be3 -D dict1 -r ../../../lib/common -q
|
||||
echo "Building fastCover dictionary with in=../../lib/common k=500 f=24 out=dict2 dictID=100 maxdict=140000"
|
||||
./main in=../../../lib/common k=500 f=24 out=dict2 dictID=100 maxdict=140000
|
||||
zstd -be3 -D dict2 -r ../../../lib/common -q
|
||||
echo "Building fastCover dictionary with 2 sample sources"
|
||||
./main in=../../../lib/common in=../../../lib/compress out=dict3
|
||||
zstd -be3 -D dict3 -r ../../../lib/common -q
|
||||
echo "Removing dict1 dict2 dict3"
|
||||
rm -f dict1 dict2 dict3
|
||||
|
||||
echo "Testing with invalid parameters, should fail"
|
||||
! ./main in=../../../lib/common r=10
|
||||
! ./main in=../../../lib/common d=10
|
||||
@ -0,0 +1,52 @@
|
||||
ARG :=
|
||||
|
||||
CC ?= gcc
|
||||
CFLAGS ?= -O3
|
||||
INCLUDES := -I ../../../programs -I ../../../lib/common -I ../../../lib -I ../../../lib/dictBuilder
|
||||
|
||||
TEST_INPUT := ../../../lib
|
||||
TEST_OUTPUT := randomDict
|
||||
|
||||
all: main run clean
|
||||
|
||||
.PHONY: test
|
||||
test: main testrun testshell clean
|
||||
|
||||
.PHONY: run
|
||||
run:
|
||||
echo "Building a random dictionary with given arguments"
|
||||
./main $(ARG)
|
||||
|
||||
main: main.o io.o random.o libzstd.a
|
||||
$(CC) $(CFLAGS) main.o io.o random.o libzstd.a -o main
|
||||
|
||||
main.o: main.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c main.c
|
||||
|
||||
random.o: random.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c random.c
|
||||
|
||||
io.o: io.c
|
||||
$(CC) $(CFLAGS) $(INCLUDES) -c io.c
|
||||
|
||||
libzstd.a:
|
||||
$(MAKE) -C ../../../lib libzstd.a
|
||||
mv ../../../lib/libzstd.a .
|
||||
|
||||
.PHONY: testrun
|
||||
testrun: main
|
||||
echo "Run with $(TEST_INPUT) and $(TEST_OUTPUT) "
|
||||
./main in=$(TEST_INPUT) out=$(TEST_OUTPUT)
|
||||
zstd -be3 -D $(TEST_OUTPUT) -r $(TEST_INPUT) -q
|
||||
rm -f $(TEST_OUTPUT)
|
||||
|
||||
.PHONY: testshell
|
||||
testshell: test.sh
|
||||
sh test.sh
|
||||
echo "Finish running test.sh"
|
||||
|
||||
.PHONY: clean
|
||||
clean:
|
||||
rm -f *.o main libzstd.a
|
||||
$(MAKE) -C ../../../lib clean
|
||||
echo "Cleaning is completed"
|
||||
@ -0,0 +1,20 @@
|
||||
Random Dictionary Builder
|
||||
|
||||
### Permitted Arguments:
|
||||
Input File/Directory (in=fileName): required; file/directory used to build dictionary; if directory, will operate recursively for files inside directory; can include multiple files/directories, each following "in="
|
||||
Output Dictionary (out=dictName): if not provided, default to defaultDict
|
||||
Dictionary ID (dictID=#): nonnegative number; if not provided, default to 0
|
||||
Maximum Dictionary Size (maxdict=#): positive number; in bytes, if not provided, default to 110KB
|
||||
Size of Randomly Selected Segment (k=#): positive number; in bytes; if not provided, default to 200
|
||||
|
||||
###Running Test:
|
||||
make test
|
||||
|
||||
|
||||
###Usage:
|
||||
To build a random dictionary with the provided arguments: make ARG= followed by arguments
|
||||
|
||||
|
||||
### Examples:
|
||||
make ARG="in=../../../lib/dictBuilder out=dict100 dictID=520"
|
||||
make ARG="in=../../../lib/dictBuilder in=../../../lib/compress"
|
||||
284
contrib/experimental_dict_builders/randomDictBuilder/io.c
Normal file
284
contrib/experimental_dict_builders/randomDictBuilder/io.c
Normal file
@ -0,0 +1,284 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* strcmp, strlen */
|
||||
#include <errno.h> /* errno */
|
||||
#include <ctype.h>
|
||||
#include "io.h"
|
||||
#include "fileio.h" /* stdinmark, stdoutmark, ZSTD_EXTENSION */
|
||||
#include "platform.h" /* Large Files support */
|
||||
#include "util.h"
|
||||
#include "zdict.h"
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
|
||||
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); }
|
||||
|
||||
static const U64 g_refreshRate = SEC_TO_MICRO / 6;
|
||||
static UTIL_time_t g_displayClock = UTIL_TIME_INITIALIZER;
|
||||
|
||||
#define DISPLAYUPDATE(l, ...) { if (displayLevel>=l) { \
|
||||
if ((UTIL_clockSpanMicro(g_displayClock) > g_refreshRate) || (displayLevel>=4)) \
|
||||
{ g_displayClock = UTIL_getTime(); DISPLAY(__VA_ARGS__); \
|
||||
if (displayLevel>=4) fflush(stderr); } } }
|
||||
|
||||
/*-*************************************
|
||||
* Exceptions
|
||||
***************************************/
|
||||
#ifndef DEBUG
|
||||
# define DEBUG 0
|
||||
#endif
|
||||
#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
|
||||
#define EXM_THROW(error, ...) \
|
||||
{ \
|
||||
DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
|
||||
DISPLAY("Error %i : ", error); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
DISPLAY("\n"); \
|
||||
exit(error); \
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Constants
|
||||
***************************************/
|
||||
|
||||
#define SAMPLESIZE_MAX (128 KB)
|
||||
#define RANDOM_MAX_SAMPLES_SIZE (sizeof(size_t) == 8 ? ((U32)-1) : ((U32)1 GB))
|
||||
#define RANDOM_MEMMULT 9
|
||||
static const size_t g_maxMemory = (sizeof(size_t) == 4) ?
|
||||
(2 GB - 64 MB) : ((size_t)(512 MB) << sizeof(size_t));
|
||||
|
||||
#define NOISELENGTH 32
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Commandline related functions
|
||||
***************************************/
|
||||
unsigned readU32FromChar(const char** stringPtr){
|
||||
const char errorMsg[] = "error: numeric value too large";
|
||||
unsigned result = 0;
|
||||
while ((**stringPtr >='0') && (**stringPtr <='9')) {
|
||||
unsigned const max = (((unsigned)(-1)) / 10) - 1;
|
||||
if (result > max) exit(1);
|
||||
result *= 10, result += **stringPtr - '0', (*stringPtr)++ ;
|
||||
}
|
||||
if ((**stringPtr=='K') || (**stringPtr=='M')) {
|
||||
unsigned const maxK = ((unsigned)(-1)) >> 10;
|
||||
if (result > maxK) exit(1);
|
||||
result <<= 10;
|
||||
if (**stringPtr=='M') {
|
||||
if (result > maxK) exit(1);
|
||||
result <<= 10;
|
||||
}
|
||||
(*stringPtr)++; /* skip `K` or `M` */
|
||||
if (**stringPtr=='i') (*stringPtr)++;
|
||||
if (**stringPtr=='B') (*stringPtr)++;
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
unsigned longCommandWArg(const char** stringPtr, const char* longCommand){
|
||||
size_t const comSize = strlen(longCommand);
|
||||
int const result = !strncmp(*stringPtr, longCommand, comSize);
|
||||
if (result) *stringPtr += comSize;
|
||||
return result;
|
||||
}
|
||||
|
||||
|
||||
/* ********************************************************
|
||||
* File related operations
|
||||
**********************************************************/
|
||||
/** loadFiles() :
|
||||
* load samples from files listed in fileNamesTable into buffer.
|
||||
* works even if buffer is too small to load all samples.
|
||||
* Also provides the size of each sample into sampleSizes table
|
||||
* which must be sized correctly, using DiB_fileStats().
|
||||
* @return : nb of samples effectively loaded into `buffer`
|
||||
* *bufferSizePtr is modified, it provides the amount data loaded within buffer.
|
||||
* sampleSizes is filled with the size of each sample.
|
||||
*/
|
||||
static unsigned loadFiles(void* buffer, size_t* bufferSizePtr, size_t* sampleSizes,
|
||||
unsigned sstSize, const char** fileNamesTable, unsigned nbFiles,
|
||||
size_t targetChunkSize, unsigned displayLevel) {
|
||||
char* const buff = (char*)buffer;
|
||||
size_t pos = 0;
|
||||
unsigned nbLoadedChunks = 0, fileIndex;
|
||||
|
||||
for (fileIndex=0; fileIndex<nbFiles; fileIndex++) {
|
||||
const char* const fileName = fileNamesTable[fileIndex];
|
||||
unsigned long long const fs64 = UTIL_getFileSize(fileName);
|
||||
unsigned long long remainingToLoad = (fs64 == UTIL_FILESIZE_UNKNOWN) ? 0 : fs64;
|
||||
U32 const nbChunks = targetChunkSize ? (U32)((fs64 + (targetChunkSize-1)) / targetChunkSize) : 1;
|
||||
U64 const chunkSize = targetChunkSize ? MIN(targetChunkSize, fs64) : fs64;
|
||||
size_t const maxChunkSize = (size_t)MIN(chunkSize, SAMPLESIZE_MAX);
|
||||
U32 cnb;
|
||||
FILE* const f = fopen(fileName, "rb");
|
||||
if (f==NULL) EXM_THROW(10, "zstd: dictBuilder: %s %s ", fileName, strerror(errno));
|
||||
DISPLAYUPDATE(2, "Loading %s... \r", fileName);
|
||||
for (cnb=0; cnb<nbChunks; cnb++) {
|
||||
size_t const toLoad = (size_t)MIN(maxChunkSize, remainingToLoad);
|
||||
if (toLoad > *bufferSizePtr-pos) break;
|
||||
{ size_t const readSize = fread(buff+pos, 1, toLoad, f);
|
||||
if (readSize != toLoad) EXM_THROW(11, "Pb reading %s", fileName);
|
||||
pos += readSize;
|
||||
sampleSizes[nbLoadedChunks++] = toLoad;
|
||||
remainingToLoad -= targetChunkSize;
|
||||
if (nbLoadedChunks == sstSize) { /* no more space left in sampleSizes table */
|
||||
fileIndex = nbFiles; /* stop there */
|
||||
break;
|
||||
}
|
||||
if (toLoad < targetChunkSize) {
|
||||
fseek(f, (long)(targetChunkSize - toLoad), SEEK_CUR);
|
||||
} } }
|
||||
fclose(f);
|
||||
}
|
||||
DISPLAYLEVEL(2, "\r%79s\r", "");
|
||||
*bufferSizePtr = pos;
|
||||
DISPLAYLEVEL(4, "loaded : %u KB \n", (U32)(pos >> 10))
|
||||
return nbLoadedChunks;
|
||||
}
|
||||
|
||||
#define rotl32(x,r) ((x << r) | (x >> (32 - r)))
|
||||
static U32 getRand(U32* src)
|
||||
{
|
||||
static const U32 prime1 = 2654435761U;
|
||||
static const U32 prime2 = 2246822519U;
|
||||
U32 rand32 = *src;
|
||||
rand32 *= prime1;
|
||||
rand32 ^= prime2;
|
||||
rand32 = rotl32(rand32, 13);
|
||||
*src = rand32;
|
||||
return rand32 >> 5;
|
||||
}
|
||||
|
||||
/* shuffle() :
|
||||
* shuffle a table of file names in a semi-random way
|
||||
* It improves dictionary quality by reducing "locality" impact, so if sample set is very large,
|
||||
* it will load random elements from it, instead of just the first ones. */
|
||||
static void shuffle(const char** fileNamesTable, unsigned nbFiles) {
|
||||
U32 seed = 0xFD2FB528;
|
||||
unsigned i;
|
||||
for (i = nbFiles - 1; i > 0; --i) {
|
||||
unsigned const j = getRand(&seed) % (i + 1);
|
||||
const char* const tmp = fileNamesTable[j];
|
||||
fileNamesTable[j] = fileNamesTable[i];
|
||||
fileNamesTable[i] = tmp;
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
/*-********************************************************
|
||||
* Dictionary training functions
|
||||
**********************************************************/
|
||||
size_t findMaxMem(unsigned long long requiredMem) {
|
||||
size_t const step = 8 MB;
|
||||
void* testmem = NULL;
|
||||
|
||||
requiredMem = (((requiredMem >> 23) + 1) << 23);
|
||||
requiredMem += step;
|
||||
if (requiredMem > g_maxMemory) requiredMem = g_maxMemory;
|
||||
|
||||
while (!testmem) {
|
||||
testmem = malloc((size_t)requiredMem);
|
||||
requiredMem -= step;
|
||||
}
|
||||
|
||||
free(testmem);
|
||||
return (size_t)requiredMem;
|
||||
}
|
||||
|
||||
void saveDict(const char* dictFileName,
|
||||
const void* buff, size_t buffSize) {
|
||||
FILE* const f = fopen(dictFileName, "wb");
|
||||
if (f==NULL) EXM_THROW(3, "cannot open %s ", dictFileName);
|
||||
|
||||
{ size_t const n = fwrite(buff, 1, buffSize, f);
|
||||
if (n!=buffSize) EXM_THROW(4, "%s : write error", dictFileName) }
|
||||
|
||||
{ size_t const n = (size_t)fclose(f);
|
||||
if (n!=0) EXM_THROW(5, "%s : flush error", dictFileName) }
|
||||
}
|
||||
|
||||
/*! getFileStats() :
|
||||
* Given a list of files, and a chunkSize (0 == no chunk, whole files)
|
||||
* provides the amount of data to be loaded and the resulting nb of samples.
|
||||
* This is useful primarily for allocation purpose => sample buffer, and sample sizes table.
|
||||
*/
|
||||
static fileStats getFileStats(const char** fileNamesTable, unsigned nbFiles,
|
||||
size_t chunkSize, unsigned displayLevel) {
|
||||
fileStats fs;
|
||||
unsigned n;
|
||||
memset(&fs, 0, sizeof(fs));
|
||||
for (n=0; n<nbFiles; n++) {
|
||||
U64 const fileSize = UTIL_getFileSize(fileNamesTable[n]);
|
||||
U64 const srcSize = (fileSize == UTIL_FILESIZE_UNKNOWN) ? 0 : fileSize;
|
||||
U32 const nbSamples = (U32)(chunkSize ? (srcSize + (chunkSize-1)) / chunkSize : 1);
|
||||
U64 const chunkToLoad = chunkSize ? MIN(chunkSize, srcSize) : srcSize;
|
||||
size_t const cappedChunkSize = (size_t)MIN(chunkToLoad, SAMPLESIZE_MAX);
|
||||
fs.totalSizeToLoad += cappedChunkSize * nbSamples;
|
||||
fs.oneSampleTooLarge |= (chunkSize > 2*SAMPLESIZE_MAX);
|
||||
fs.nbSamples += nbSamples;
|
||||
}
|
||||
DISPLAYLEVEL(4, "Preparing to load : %u KB \n", (U32)(fs.totalSizeToLoad >> 10));
|
||||
return fs;
|
||||
}
|
||||
|
||||
|
||||
|
||||
|
||||
sampleInfo* getSampleInfo(const char** fileNamesTable, unsigned nbFiles, size_t chunkSize,
|
||||
unsigned maxDictSize, const unsigned displayLevel) {
|
||||
fileStats const fs = getFileStats(fileNamesTable, nbFiles, chunkSize, displayLevel);
|
||||
size_t* const sampleSizes = (size_t*)malloc(fs.nbSamples * sizeof(size_t));
|
||||
size_t const memMult = RANDOM_MEMMULT;
|
||||
size_t const maxMem = findMaxMem(fs.totalSizeToLoad * memMult) / memMult;
|
||||
size_t loadedSize = (size_t) MIN ((unsigned long long)maxMem, fs.totalSizeToLoad);
|
||||
void* const srcBuffer = malloc(loadedSize+NOISELENGTH);
|
||||
|
||||
/* Checks */
|
||||
if ((!sampleSizes) || (!srcBuffer))
|
||||
EXM_THROW(12, "not enough memory for trainFromFiles"); /* should not happen */
|
||||
if (fs.oneSampleTooLarge) {
|
||||
DISPLAYLEVEL(2, "! Warning : some sample(s) are very large \n");
|
||||
DISPLAYLEVEL(2, "! Note that dictionary is only useful for small samples. \n");
|
||||
DISPLAYLEVEL(2, "! As a consequence, only the first %u bytes of each sample are loaded \n", SAMPLESIZE_MAX);
|
||||
}
|
||||
if (fs.nbSamples < 5) {
|
||||
DISPLAYLEVEL(2, "! Warning : nb of samples too low for proper processing ! \n");
|
||||
DISPLAYLEVEL(2, "! Please provide _one file per sample_. \n");
|
||||
DISPLAYLEVEL(2, "! Alternatively, split files into fixed-size blocks representative of samples, with -B# \n");
|
||||
EXM_THROW(14, "nb of samples too low"); /* we now clearly forbid this case */
|
||||
}
|
||||
if (fs.totalSizeToLoad < (unsigned long long)(8 * maxDictSize)) {
|
||||
DISPLAYLEVEL(2, "! Warning : data size of samples too small for target dictionary size \n");
|
||||
DISPLAYLEVEL(2, "! Samples should be about 100x larger than target dictionary size \n");
|
||||
}
|
||||
|
||||
/* init */
|
||||
if (loadedSize < fs.totalSizeToLoad)
|
||||
DISPLAYLEVEL(1, "Not enough memory; training on %u MB only...\n", (unsigned)(loadedSize >> 20));
|
||||
|
||||
/* Load input buffer */
|
||||
DISPLAYLEVEL(3, "Shuffling input files\n");
|
||||
shuffle(fileNamesTable, nbFiles);
|
||||
nbFiles = loadFiles(srcBuffer, &loadedSize, sampleSizes, fs.nbSamples,
|
||||
fileNamesTable, nbFiles, chunkSize, displayLevel);
|
||||
|
||||
sampleInfo *info = (sampleInfo *)malloc(sizeof(sampleInfo));
|
||||
|
||||
info->nbSamples = fs.nbSamples;
|
||||
info->samplesSizes = sampleSizes;
|
||||
info->srcBuffer = srcBuffer;
|
||||
|
||||
return info;
|
||||
}
|
||||
|
||||
|
||||
void freeSampleInfo(sampleInfo *info) {
|
||||
if (!info) return;
|
||||
if (info->samplesSizes) free((void*)(info->samplesSizes));
|
||||
if (info->srcBuffer) free((void*)(info->srcBuffer));
|
||||
free(info);
|
||||
}
|
||||
60
contrib/experimental_dict_builders/randomDictBuilder/io.h
Normal file
60
contrib/experimental_dict_builders/randomDictBuilder/io.h
Normal file
@ -0,0 +1,60 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* strcmp, strlen */
|
||||
#include <errno.h> /* errno */
|
||||
#include <ctype.h>
|
||||
#include "zstd_internal.h" /* includes zstd.h */
|
||||
#include "fileio.h" /* stdinmark, stdoutmark, ZSTD_EXTENSION */
|
||||
#include "platform.h" /* Large Files support */
|
||||
#include "util.h"
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Structs
|
||||
***************************************/
|
||||
typedef struct {
|
||||
U64 totalSizeToLoad;
|
||||
unsigned oneSampleTooLarge;
|
||||
unsigned nbSamples;
|
||||
} fileStats;
|
||||
|
||||
typedef struct {
|
||||
const void* srcBuffer;
|
||||
const size_t *samplesSizes;
|
||||
size_t nbSamples;
|
||||
}sampleInfo;
|
||||
|
||||
|
||||
|
||||
/*! getSampleInfo():
|
||||
* Load from input files and add samples to buffer
|
||||
* @return: a sampleInfo struct containing infomation about buffer where samples are stored,
|
||||
* size of each sample, and total number of samples
|
||||
*/
|
||||
sampleInfo* getSampleInfo(const char** fileNamesTable, unsigned nbFiles, size_t chunkSize,
|
||||
unsigned maxDictSize, const unsigned displayLevel);
|
||||
|
||||
|
||||
|
||||
/*! freeSampleInfo():
|
||||
* Free memory allocated for info
|
||||
*/
|
||||
void freeSampleInfo(sampleInfo *info);
|
||||
|
||||
|
||||
|
||||
/*! saveDict():
|
||||
* Save data stored on buff to dictFileName
|
||||
*/
|
||||
void saveDict(const char* dictFileName, const void* buff, size_t buffSize);
|
||||
|
||||
|
||||
unsigned readU32FromChar(const char** stringPtr);
|
||||
|
||||
/** longCommandWArg() :
|
||||
* check if *stringPtr is the same as longCommand.
|
||||
* If yes, @return 1 and advances *stringPtr to the position which immediately follows longCommand.
|
||||
* @return 0 and doesn't modify *stringPtr otherwise.
|
||||
*/
|
||||
unsigned longCommandWArg(const char** stringPtr, const char* longCommand);
|
||||
161
contrib/experimental_dict_builders/randomDictBuilder/main.c
Normal file
161
contrib/experimental_dict_builders/randomDictBuilder/main.c
Normal file
@ -0,0 +1,161 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* strcmp, strlen */
|
||||
#include <errno.h> /* errno */
|
||||
#include <ctype.h>
|
||||
#include "random.h"
|
||||
#include "io.h"
|
||||
#include "util.h"
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
|
||||
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); }
|
||||
|
||||
static const U64 g_refreshRate = SEC_TO_MICRO / 6;
|
||||
static UTIL_time_t g_displayClock = UTIL_TIME_INITIALIZER;
|
||||
|
||||
#define DISPLAYUPDATE(l, ...) { if (displayLevel>=l) { \
|
||||
if ((UTIL_clockSpanMicro(g_displayClock) > g_refreshRate) || (displayLevel>=4)) \
|
||||
{ g_displayClock = UTIL_getTime(); DISPLAY(__VA_ARGS__); \
|
||||
if (displayLevel>=4) fflush(stderr); } } }
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Exceptions
|
||||
***************************************/
|
||||
#ifndef DEBUG
|
||||
# define DEBUG 0
|
||||
#endif
|
||||
#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
|
||||
#define EXM_THROW(error, ...) \
|
||||
{ \
|
||||
DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
|
||||
DISPLAY("Error %i : ", error); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
DISPLAY("\n"); \
|
||||
exit(error); \
|
||||
}
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* Constants
|
||||
***************************************/
|
||||
static const unsigned g_defaultMaxDictSize = 110 KB;
|
||||
#define DEFAULT_CLEVEL 3
|
||||
#define DEFAULT_k 200
|
||||
#define DEFAULT_OUTPUTFILE "defaultDict"
|
||||
#define DEFAULT_DICTID 0
|
||||
|
||||
|
||||
|
||||
/*-*************************************
|
||||
* RANDOM
|
||||
***************************************/
|
||||
int RANDOM_trainFromFiles(const char* dictFileName, sampleInfo *info,
|
||||
unsigned maxDictSize,
|
||||
ZDICT_random_params_t *params) {
|
||||
unsigned const displayLevel = params->zParams.notificationLevel;
|
||||
void* const dictBuffer = malloc(maxDictSize);
|
||||
|
||||
int result = 0;
|
||||
|
||||
/* Checks */
|
||||
if (!dictBuffer)
|
||||
EXM_THROW(12, "not enough memory for trainFromFiles"); /* should not happen */
|
||||
|
||||
{ size_t dictSize;
|
||||
dictSize = ZDICT_trainFromBuffer_random(dictBuffer, maxDictSize, info->srcBuffer,
|
||||
info->samplesSizes, info->nbSamples, *params);
|
||||
DISPLAYLEVEL(2, "k=%u\n", params->k);
|
||||
if (ZDICT_isError(dictSize)) {
|
||||
DISPLAYLEVEL(1, "dictionary training failed : %s \n", ZDICT_getErrorName(dictSize)); /* should not happen */
|
||||
result = 1;
|
||||
goto _done;
|
||||
}
|
||||
/* save dict */
|
||||
DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName);
|
||||
saveDict(dictFileName, dictBuffer, dictSize);
|
||||
}
|
||||
|
||||
/* clean up */
|
||||
_done:
|
||||
free(dictBuffer);
|
||||
return result;
|
||||
}
|
||||
|
||||
|
||||
|
||||
int main(int argCount, const char* argv[])
|
||||
{
|
||||
int displayLevel = 2;
|
||||
const char* programName = argv[0];
|
||||
int operationResult = 0;
|
||||
|
||||
/* Initialize arguments to default values */
|
||||
unsigned k = DEFAULT_k;
|
||||
const char* outputFile = DEFAULT_OUTPUTFILE;
|
||||
unsigned dictID = DEFAULT_DICTID;
|
||||
unsigned maxDictSize = g_defaultMaxDictSize;
|
||||
|
||||
/* Initialize table to store input files */
|
||||
const char** filenameTable = (const char**)malloc(argCount * sizeof(const char*));
|
||||
unsigned filenameIdx = 0;
|
||||
|
||||
/* Parse arguments */
|
||||
for (int i = 1; i < argCount; i++) {
|
||||
const char* argument = argv[i];
|
||||
if (longCommandWArg(&argument, "k=")) { k = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "dictID=")) { dictID = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "maxdict=")) { maxDictSize = readU32FromChar(&argument); continue; }
|
||||
if (longCommandWArg(&argument, "in=")) {
|
||||
filenameTable[filenameIdx] = argument;
|
||||
filenameIdx++;
|
||||
continue;
|
||||
}
|
||||
if (longCommandWArg(&argument, "out=")) {
|
||||
outputFile = argument;
|
||||
continue;
|
||||
}
|
||||
DISPLAYLEVEL(1, "Incorrect parameters\n");
|
||||
operationResult = 1;
|
||||
return operationResult;
|
||||
}
|
||||
|
||||
char* fileNamesBuf = NULL;
|
||||
unsigned fileNamesNb = filenameIdx;
|
||||
int followLinks = 0; /* follow directory recursively */
|
||||
const char** extendedFileList = NULL;
|
||||
extendedFileList = UTIL_createFileList(filenameTable, filenameIdx, &fileNamesBuf,
|
||||
&fileNamesNb, followLinks);
|
||||
if (extendedFileList) {
|
||||
unsigned u;
|
||||
for (u=0; u<fileNamesNb; u++) DISPLAYLEVEL(4, "%u %s\n", u, extendedFileList[u]);
|
||||
free((void*)filenameTable);
|
||||
filenameTable = extendedFileList;
|
||||
filenameIdx = fileNamesNb;
|
||||
}
|
||||
|
||||
size_t blockSize = 0;
|
||||
|
||||
ZDICT_random_params_t params;
|
||||
ZDICT_params_t zParams;
|
||||
zParams.compressionLevel = DEFAULT_CLEVEL;
|
||||
zParams.notificationLevel = displayLevel;
|
||||
zParams.dictID = dictID;
|
||||
params.zParams = zParams;
|
||||
params.k = k;
|
||||
|
||||
sampleInfo* info = getSampleInfo(filenameTable,
|
||||
filenameIdx, blockSize, maxDictSize, zParams.notificationLevel);
|
||||
operationResult = RANDOM_trainFromFiles(outputFile, info, maxDictSize, ¶ms);
|
||||
|
||||
/* Free allocated memory */
|
||||
UTIL_freeFileList(extendedFileList, fileNamesBuf);
|
||||
freeSampleInfo(info);
|
||||
|
||||
return operationResult;
|
||||
}
|
||||
163
contrib/experimental_dict_builders/randomDictBuilder/random.c
Normal file
163
contrib/experimental_dict_builders/randomDictBuilder/random.c
Normal file
@ -0,0 +1,163 @@
|
||||
/*-*************************************
|
||||
* Dependencies
|
||||
***************************************/
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* memset */
|
||||
#include <time.h> /* clock */
|
||||
#include "random.h"
|
||||
#include "util.h" /* UTIL_getFileSize, UTIL_getTotalFileSize */
|
||||
#ifndef ZDICT_STATIC_LINKING_ONLY
|
||||
#define ZDICT_STATIC_LINKING_ONLY
|
||||
#endif
|
||||
#include "zdict.h"
|
||||
|
||||
/*-*************************************
|
||||
* Console display
|
||||
***************************************/
|
||||
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
|
||||
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); }
|
||||
|
||||
#define LOCALDISPLAYUPDATE(displayLevel, l, ...) \
|
||||
if (displayLevel >= l) { \
|
||||
if ((clock() - g_time > refreshRate) || (displayLevel >= 4)) { \
|
||||
g_time = clock(); \
|
||||
DISPLAY(__VA_ARGS__); \
|
||||
} \
|
||||
}
|
||||
#define DISPLAYUPDATE(l, ...) LOCALDISPLAYUPDATE(displayLevel, l, __VA_ARGS__)
|
||||
static const clock_t refreshRate = CLOCKS_PER_SEC * 15 / 100;
|
||||
static clock_t g_time = 0;
|
||||
|
||||
|
||||
|
||||
/* ********************************************************
|
||||
* Random Dictionary Builder
|
||||
**********************************************************/
|
||||
/**
|
||||
* Returns the sum of the sample sizes.
|
||||
*/
|
||||
static size_t RANDOM_sum(const size_t *samplesSizes, unsigned nbSamples) {
|
||||
size_t sum = 0;
|
||||
unsigned i;
|
||||
for (i = 0; i < nbSamples; ++i) {
|
||||
sum += samplesSizes[i];
|
||||
}
|
||||
return sum;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* A segment is an inclusive range in the source.
|
||||
*/
|
||||
typedef struct {
|
||||
U32 begin;
|
||||
U32 end;
|
||||
} RANDOM_segment_t;
|
||||
|
||||
|
||||
/**
|
||||
* Selects a random segment from totalSamplesSize - k + 1 possible segments
|
||||
*/
|
||||
static RANDOM_segment_t RANDOM_selectSegment(const size_t totalSamplesSize,
|
||||
ZDICT_random_params_t parameters) {
|
||||
const U32 k = parameters.k;
|
||||
RANDOM_segment_t segment;
|
||||
unsigned index;
|
||||
|
||||
/* Randomly generate a number from 0 to sampleSizes - k */
|
||||
index = rand()%(totalSamplesSize - k + 1);
|
||||
|
||||
/* inclusive */
|
||||
segment.begin = index;
|
||||
segment.end = index + k - 1;
|
||||
|
||||
return segment;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Check the validity of the parameters.
|
||||
* Returns non-zero if the parameters are valid and 0 otherwise.
|
||||
*/
|
||||
static int RANDOM_checkParameters(ZDICT_random_params_t parameters,
|
||||
size_t maxDictSize) {
|
||||
/* k is a required parameter */
|
||||
if (parameters.k == 0) {
|
||||
return 0;
|
||||
}
|
||||
/* k <= maxDictSize */
|
||||
if (parameters.k > maxDictSize) {
|
||||
return 0;
|
||||
}
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Given the prepared context build the dictionary.
|
||||
*/
|
||||
static size_t RANDOM_buildDictionary(const size_t totalSamplesSize, const BYTE *samples,
|
||||
void *dictBuffer, size_t dictBufferCapacity,
|
||||
ZDICT_random_params_t parameters) {
|
||||
BYTE *const dict = (BYTE *)dictBuffer;
|
||||
size_t tail = dictBufferCapacity;
|
||||
const int displayLevel = parameters.zParams.notificationLevel;
|
||||
while (tail > 0) {
|
||||
|
||||
/* Select a segment */
|
||||
RANDOM_segment_t segment = RANDOM_selectSegment(totalSamplesSize, parameters);
|
||||
|
||||
size_t segmentSize;
|
||||
segmentSize = MIN(segment.end - segment.begin + 1, tail);
|
||||
|
||||
tail -= segmentSize;
|
||||
memcpy(dict + tail, samples + segment.begin, segmentSize);
|
||||
DISPLAYUPDATE(
|
||||
2, "\r%u%% ",
|
||||
(U32)(((dictBufferCapacity - tail) * 100) / dictBufferCapacity));
|
||||
}
|
||||
|
||||
return tail;
|
||||
}
|
||||
|
||||
|
||||
|
||||
|
||||
ZDICTLIB_API size_t ZDICT_trainFromBuffer_random(
|
||||
void *dictBuffer, size_t dictBufferCapacity,
|
||||
const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
|
||||
ZDICT_random_params_t parameters) {
|
||||
const int displayLevel = parameters.zParams.notificationLevel;
|
||||
BYTE* const dict = (BYTE*)dictBuffer;
|
||||
/* Checks */
|
||||
if (!RANDOM_checkParameters(parameters, dictBufferCapacity)) {
|
||||
DISPLAYLEVEL(1, "k is incorrect\n");
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
if (nbSamples == 0) {
|
||||
DISPLAYLEVEL(1, "Random must have at least one input file\n");
|
||||
return ERROR(GENERIC);
|
||||
}
|
||||
if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
|
||||
DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
|
||||
ZDICT_DICTSIZE_MIN);
|
||||
return ERROR(dstSize_tooSmall);
|
||||
}
|
||||
const size_t totalSamplesSize = RANDOM_sum(samplesSizes, nbSamples);
|
||||
const BYTE *const samples = (const BYTE *)samplesBuffer;
|
||||
|
||||
DISPLAYLEVEL(2, "Building dictionary\n");
|
||||
{
|
||||
const size_t tail = RANDOM_buildDictionary(totalSamplesSize, samples,
|
||||
dictBuffer, dictBufferCapacity, parameters);
|
||||
const size_t dictSize = ZDICT_finalizeDictionary(
|
||||
dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
|
||||
samplesBuffer, samplesSizes, nbSamples, parameters.zParams);
|
||||
if (!ZSTD_isError(dictSize)) {
|
||||
DISPLAYLEVEL(2, "Constructed dictionary of size %u\n",
|
||||
(U32)dictSize);
|
||||
}
|
||||
return dictSize;
|
||||
}
|
||||
}
|
||||
@ -0,0 +1,29 @@
|
||||
#include <stdio.h> /* fprintf */
|
||||
#include <stdlib.h> /* malloc, free, qsort */
|
||||
#include <string.h> /* memset */
|
||||
#include <time.h> /* clock */
|
||||
#include "zstd_internal.h" /* includes zstd.h */
|
||||
#ifndef ZDICT_STATIC_LINKING_ONLY
|
||||
#define ZDICT_STATIC_LINKING_ONLY
|
||||
#endif
|
||||
#include "zdict.h"
|
||||
|
||||
|
||||
|
||||
typedef struct {
|
||||
unsigned k; /* Segment size : constraint: 0 < k : Reasonable range [16, 2048+]; Default to 200 */
|
||||
ZDICT_params_t zParams;
|
||||
} ZDICT_random_params_t;
|
||||
|
||||
|
||||
/*! ZDICT_trainFromBuffer_random():
|
||||
* Train a dictionary from an array of samples.
|
||||
* Samples must be stored concatenated in a single flat buffer `samplesBuffer`,
|
||||
* supplied with an array of sizes `samplesSizes`, providing the size of each sample, in order.
|
||||
* The resulting dictionary will be saved into `dictBuffer`.
|
||||
* @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
|
||||
* or an error code, which can be tested with ZDICT_isError().
|
||||
*/
|
||||
ZDICTLIB_API size_t ZDICT_trainFromBuffer_random( void *dictBuffer, size_t dictBufferCapacity,
|
||||
const void *samplesBuffer, const size_t *samplesSizes, unsigned nbSamples,
|
||||
ZDICT_random_params_t parameters);
|
||||
14
contrib/experimental_dict_builders/randomDictBuilder/test.sh
Normal file
14
contrib/experimental_dict_builders/randomDictBuilder/test.sh
Normal file
@ -0,0 +1,14 @@
|
||||
echo "Building random dictionary with in=../../lib/common k=200 out=dict1"
|
||||
./main in=../../../lib/common k=200 out=dict1
|
||||
zstd -be3 -D dict1 -r ../../../lib/common -q
|
||||
echo "Building random dictionary with in=../../lib/common k=500 out=dict2 dictID=100 maxdict=140000"
|
||||
./main in=../../../lib/common k=500 out=dict2 dictID=100 maxdict=140000
|
||||
zstd -be3 -D dict2 -r ../../../lib/common -q
|
||||
echo "Building random dictionary with 2 sample sources"
|
||||
./main in=../../../lib/common in=../../../lib/compress out=dict3
|
||||
zstd -be3 -D dict3 -r ../../../lib/common -q
|
||||
echo "Removing dict1 dict2 dict3"
|
||||
rm -f dict1 dict2 dict3
|
||||
|
||||
echo "Testing with invalid parameters, should fail"
|
||||
! ./main r=10
|
||||
@ -3335,9 +3335,11 @@ size_t ZSTD_CStreamOutSize(void)
|
||||
static size_t ZSTD_resetCStream_internal(ZSTD_CStream* cctx,
|
||||
const void* const dict, size_t const dictSize, ZSTD_dictContentType_e const dictContentType,
|
||||
const ZSTD_CDict* const cdict,
|
||||
ZSTD_CCtx_params const params, unsigned long long const pledgedSrcSize)
|
||||
ZSTD_CCtx_params params, unsigned long long const pledgedSrcSize)
|
||||
{
|
||||
DEBUGLOG(4, "ZSTD_resetCStream_internal");
|
||||
/* Finalize the compression parameters */
|
||||
params.cParams = ZSTD_getCParamsFromCCtxParams(¶ms, pledgedSrcSize, dictSize);
|
||||
/* params are supposed to be fully validated at this point */
|
||||
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
|
||||
assert(!((dict) && (cdict))); /* either dict or cdict, not both */
|
||||
@ -3366,7 +3368,6 @@ size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
|
||||
DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (U32)pledgedSrcSize);
|
||||
if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;
|
||||
params.fParams.contentSizeFlag = 1;
|
||||
params.cParams = ZSTD_getCParamsFromCCtxParams(¶ms, pledgedSrcSize, 0);
|
||||
return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dct_auto, zcs->cdict, params, pledgedSrcSize);
|
||||
}
|
||||
|
||||
@ -3379,6 +3380,7 @@ size_t ZSTD_initCStream_internal(ZSTD_CStream* zcs,
|
||||
ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)
|
||||
{
|
||||
DEBUGLOG(4, "ZSTD_initCStream_internal");
|
||||
params.cParams = ZSTD_getCParamsFromCCtxParams(¶ms, pledgedSrcSize, dictSize);
|
||||
assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
|
||||
assert(!((dict) && (cdict))); /* either dict or cdict, not both */
|
||||
|
||||
@ -3445,25 +3447,21 @@ size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
|
||||
(U32)pledgedSrcSize, params.fParams.contentSizeFlag);
|
||||
CHECK_F( ZSTD_checkCParams(params.cParams) );
|
||||
if ((pledgedSrcSize==0) && (params.fParams.contentSizeFlag==0)) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN; /* for compatibility with older programs relying on this behavior. Users should now specify ZSTD_CONTENTSIZE_UNKNOWN. This line will be removed in the future. */
|
||||
{ ZSTD_CCtx_params const cctxParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
|
||||
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL /*cdict*/, cctxParams, pledgedSrcSize);
|
||||
}
|
||||
zcs->requestedParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
|
||||
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL /*cdict*/, zcs->requestedParams, pledgedSrcSize);
|
||||
}
|
||||
|
||||
size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)
|
||||
{
|
||||
ZSTD_parameters const params = ZSTD_getParams(compressionLevel, 0, dictSize);
|
||||
ZSTD_CCtx_params const cctxParams =
|
||||
ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
|
||||
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, cctxParams, ZSTD_CONTENTSIZE_UNKNOWN);
|
||||
ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel);
|
||||
return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, zcs->requestedParams, ZSTD_CONTENTSIZE_UNKNOWN);
|
||||
}
|
||||
|
||||
size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss)
|
||||
{
|
||||
U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss; /* temporary : 0 interpreted as "unknown" during transition period. Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN. `0` will be interpreted as "empty" in the future */
|
||||
ZSTD_parameters const params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0);
|
||||
ZSTD_CCtx_params const cctxParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
|
||||
return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, cctxParams, pledgedSrcSize);
|
||||
ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel);
|
||||
return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, zcs->requestedParams, pledgedSrcSize);
|
||||
}
|
||||
|
||||
size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
|
||||
|
||||
@ -37,7 +37,9 @@
|
||||
#define ZSTD_RESIZE_SEQPOOL 0
|
||||
|
||||
/* ====== Debug ====== */
|
||||
#if defined(DEBUGLEVEL) && (DEBUGLEVEL>=2) && !defined(_MSC_VER)
|
||||
#if defined(DEBUGLEVEL) && (DEBUGLEVEL>=2) \
|
||||
&& !defined(_MSC_VER) \
|
||||
&& !defined(__MINGW32__)
|
||||
|
||||
# include <stdio.h>
|
||||
# include <unistd.h>
|
||||
|
||||
@ -549,7 +549,8 @@ static BMK_return_t BMK_benchMemAdvancedNoAlloc(
|
||||
double const compressionSpeed = ((double)srcSize / intermediateResultCompress.result.result.nanoSecPerRun) * 1000;
|
||||
int const cSpeedAccuracy = (compressionSpeed < 10.) ? 2 : 1;
|
||||
results.result.cSpeed = compressionSpeed * 1000000;
|
||||
results.result.cSize = intermediateResultCompress.result.result.sumOfReturn;
|
||||
cSize = intermediateResultCompress.result.result.sumOfReturn;
|
||||
results.result.cSize = cSize;
|
||||
ratio = (double)srcSize / results.result.cSize;
|
||||
markNb = (markNb+1) % NB_MARKS;
|
||||
DISPLAYLEVEL(2, "%2s-%-17.17s :%10u ->%10u (%5.*f),%6.*f MB/s\r",
|
||||
|
||||
@ -797,6 +797,14 @@ FIO_compressZstdFrame(const cRess_t* ressPtr,
|
||||
}
|
||||
} while (directive != ZSTD_e_end);
|
||||
|
||||
if (ferror(srcFile)) {
|
||||
EXM_THROW(26, "Read error : I/O error");
|
||||
}
|
||||
if (fileSize != UTIL_FILESIZE_UNKNOWN && *readsize != fileSize) {
|
||||
EXM_THROW(27, "Read error : Incomplete read : %llu / %llu B",
|
||||
(unsigned long long)*readsize, (unsigned long long)fileSize);
|
||||
}
|
||||
|
||||
return compressedfilesize;
|
||||
}
|
||||
|
||||
|
||||
@ -502,7 +502,7 @@ int main(int argCount, const char* argv[])
|
||||
if (!strcmp(argument, "--sparse")) { FIO_setSparseWrite(2); continue; }
|
||||
if (!strcmp(argument, "--no-sparse")) { FIO_setSparseWrite(0); continue; }
|
||||
if (!strcmp(argument, "--test")) { operation=zom_test; continue; }
|
||||
if (!strcmp(argument, "--train")) { operation=zom_train; outFileName=g_defaultDictName; continue; }
|
||||
if (!strcmp(argument, "--train")) { operation=zom_train; if (outFileName==NULL) outFileName=g_defaultDictName; continue; }
|
||||
if (!strcmp(argument, "--maxdict")) { nextArgumentIsMaxDict=1; lastCommand=1; continue; } /* kept available for compatibility with old syntax ; will be removed one day */
|
||||
if (!strcmp(argument, "--dictID")) { nextArgumentIsDictID=1; lastCommand=1; continue; } /* kept available for compatibility with old syntax ; will be removed one day */
|
||||
if (!strcmp(argument, "--no-dictID")) { FIO_setDictIDFlag(0); continue; }
|
||||
@ -526,7 +526,8 @@ int main(int argCount, const char* argv[])
|
||||
#ifndef ZSTD_NODICT
|
||||
if (longCommandWArg(&argument, "--train-cover")) {
|
||||
operation = zom_train;
|
||||
outFileName = g_defaultDictName;
|
||||
if (outFileName == NULL)
|
||||
outFileName = g_defaultDictName;
|
||||
cover = 1;
|
||||
/* Allow optional arguments following an = */
|
||||
if (*argument == 0) { memset(&coverParams, 0, sizeof(coverParams)); }
|
||||
@ -536,7 +537,8 @@ int main(int argCount, const char* argv[])
|
||||
}
|
||||
if (longCommandWArg(&argument, "--train-legacy")) {
|
||||
operation = zom_train;
|
||||
outFileName = g_defaultDictName;
|
||||
if (outFileName == NULL)
|
||||
outFileName = g_defaultDictName;
|
||||
cover = 0;
|
||||
/* Allow optional arguments following an = */
|
||||
if (*argument == 0) { continue; }
|
||||
@ -718,7 +720,7 @@ int main(int argCount, const char* argv[])
|
||||
break;
|
||||
|
||||
/* Select compressibility of synthetic sample */
|
||||
case 'P':
|
||||
case 'P':
|
||||
{ argument++;
|
||||
compressibility = (double)readU32FromChar(&argument) / 100;
|
||||
}
|
||||
@ -841,7 +843,7 @@ int main(int argCount, const char* argv[])
|
||||
if (cLevel > ZSTD_maxCLevel()) cLevel = ZSTD_maxCLevel();
|
||||
if (cLevelLast > ZSTD_maxCLevel()) cLevelLast = ZSTD_maxCLevel();
|
||||
if (cLevelLast < cLevel) cLevelLast = cLevel;
|
||||
if (cLevelLast > cLevel)
|
||||
if (cLevelLast > cLevel)
|
||||
DISPLAYLEVEL(2, "Benchmarking levels from %d to %d\n", cLevel, cLevelLast);
|
||||
if(filenameIdx) {
|
||||
if(separateFiles) {
|
||||
@ -856,7 +858,7 @@ int main(int argCount, const char* argv[])
|
||||
} else {
|
||||
for(; cLevel <= cLevelLast; cLevel++) {
|
||||
BMK_benchFilesAdvanced(filenameTable, filenameIdx, dictFileName, cLevel, &compressionParams, g_displayLevel, &adv);
|
||||
}
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for(; cLevel <= cLevelLast; cLevel++) {
|
||||
|
||||
@ -404,7 +404,13 @@ $ECHO "Hello World" > tmp
|
||||
$ZSTD --train-legacy -q tmp && die "Dictionary training should fail : not enough input source"
|
||||
./datagen -P0 -g10M > tmp
|
||||
$ZSTD --train-legacy -q tmp && die "Dictionary training should fail : source is pure noise"
|
||||
rm tmp*
|
||||
$ECHO "- Test -o before --train"
|
||||
rm -f tmpDict dictionary
|
||||
$ZSTD -o tmpDict --train *.c ../programs/*.c
|
||||
test -f tmpDict
|
||||
$ZSTD --train *.c ../programs/*.c
|
||||
test -f dictionary
|
||||
rm tmp* dictionary
|
||||
|
||||
|
||||
$ECHO "\n===> cover dictionary builder : advanced options "
|
||||
@ -425,12 +431,18 @@ $ZSTD --train-cover=k=46,d=8 *.c ../programs/*.c --dictID=1 -o tmpDict1
|
||||
cmp tmpDict tmpDict1 && die "dictionaries should have different ID !"
|
||||
$ECHO "- Create dictionary with size limit"
|
||||
$ZSTD --train-cover=steps=8 *.c ../programs/*.c -o tmpDict2 --maxdict=4K
|
||||
rm tmp*
|
||||
$ECHO "- Compare size of dictionary from 90% training samples with 80% training samples"
|
||||
$ZSTD --train-cover=split=90 -r *.c ../programs/*.c
|
||||
$ZSTD --train-cover=split=80 -r *.c ../programs/*.c
|
||||
$ECHO "- Create dictionary using all samples for both training and testing"
|
||||
$ZSTD --train-cover=split=100 -r *.c ../programs/*.c
|
||||
$ECHO "- Test -o before --train-cover"
|
||||
rm -f tmpDict dictionary
|
||||
$ZSTD -o tmpDict --train-cover *.c ../programs/*.c
|
||||
test -f tmpDict
|
||||
$ZSTD --train-cover *.c ../programs/*.c
|
||||
test -f dictionary
|
||||
rm tmp* dictionary
|
||||
|
||||
$ECHO "\n===> legacy dictionary builder "
|
||||
|
||||
@ -450,7 +462,13 @@ $ZSTD --train-legacy -s5 *.c ../programs/*.c --dictID=1 -o tmpDict1
|
||||
cmp tmpDict tmpDict1 && die "dictionaries should have different ID !"
|
||||
$ECHO "- Create dictionary with size limit"
|
||||
$ZSTD --train-legacy -s9 *.c ../programs/*.c -o tmpDict2 --maxdict=4K
|
||||
rm tmp*
|
||||
$ECHO "- Test -o before --train-legacy"
|
||||
rm -f tmpDict dictionary
|
||||
$ZSTD -o tmpDict --train-legacy *.c ../programs/*.c
|
||||
test -f tmpDict
|
||||
$ZSTD --train-legacy *.c ../programs/*.c
|
||||
test -f dictionary
|
||||
rm tmp* dictionary
|
||||
|
||||
|
||||
$ECHO "\n===> integrity tests "
|
||||
|
||||
@ -969,6 +969,26 @@ static int basicUnitTests(U32 seed, double compressibility)
|
||||
}
|
||||
DISPLAYLEVEL(3, "OK \n");
|
||||
|
||||
DISPLAYLEVEL(3, "test%3i : ZSTD_initCStream_srcSize sets requestedParams : ", testNb++);
|
||||
{ unsigned level;
|
||||
CHECK_Z(ZSTD_initCStream_srcSize(zc, 11, ZSTD_CONTENTSIZE_UNKNOWN));
|
||||
CHECK_Z(ZSTD_CCtx_getParameter(zc, ZSTD_p_compressionLevel, &level));
|
||||
CHECK(level != 11, "Compression level does not match");
|
||||
ZSTD_resetCStream(zc, ZSTD_CONTENTSIZE_UNKNOWN);
|
||||
CHECK_Z(ZSTD_CCtx_getParameter(zc, ZSTD_p_compressionLevel, &level));
|
||||
CHECK(level != 11, "Compression level does not match");
|
||||
}
|
||||
DISPLAYLEVEL(3, "OK \n");
|
||||
|
||||
DISPLAYLEVEL(3, "test%3i : ZSTD_initCStream_advanced sets requestedParams : ", testNb++);
|
||||
{ ZSTD_parameters const params = ZSTD_getParams(9, 0, 0);
|
||||
CHECK_Z(ZSTD_initCStream_advanced(zc, NULL, 0, params, ZSTD_CONTENTSIZE_UNKNOWN));
|
||||
CHECK(badParameters(zc, params), "Compression parameters do not match");
|
||||
ZSTD_resetCStream(zc, ZSTD_CONTENTSIZE_UNKNOWN);
|
||||
CHECK(badParameters(zc, params), "Compression parameters do not match");
|
||||
}
|
||||
DISPLAYLEVEL(3, "OK \n");
|
||||
|
||||
/* Overlen overwriting window data bug */
|
||||
DISPLAYLEVEL(3, "test%3i : wildcopy doesn't overwrite potential match data : ", testNb++);
|
||||
{ /* This test has a window size of 1024 bytes and consists of 3 blocks:
|
||||
|
||||
Loading…
Reference in New Issue
Block a user