ddidderr/zstd-rs
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #3913 from facebook/lorem2

Browse Source 
Lorem Ipsum generator update
This commit is contained in:
Yann Collet
2024-02-22 09:41:12 -08:00
committed by GitHub
parent 4afe31ba7b 7170f51dd2
commit 3c78b42210
11 changed files with 1149 additions and 684 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.gitignore
+5 -2
View File
@@ -39,12 +39,15 @@ buck-out/
build-* build-*
*.gcda *.gcda
# IDE
.clang_complete
compile_flags.txt
.clang-format
# Other files # Other files
.directory .directory
_codelite/ _codelite/
_zstdbench/ _zstdbench/
.clang_complete
compile_flags.txt
*.idea *.idea
*.swp *.swp
.DS_Store .DS_Store
build/VS2010/datagen/datagen.vcxproj
+2
View File
@@ -157,6 +157,8 @@
<ItemGroup> <ItemGroup>
<ClCompile Include="..\..\..\programs\util.c" /> <ClCompile Include="..\..\..\programs\util.c" />
<ClCompile Include="..\..\..\programs\datagen.c" /> <ClCompile Include="..\..\..\programs\datagen.c" />
<ClCompile Include="..\..\..\programs\lorem.c" />
<ClCompile Include="..\..\..\tests\loremOut.c" />
<ClCompile Include="..\..\..\tests\datagencli.c" /> <ClCompile Include="..\..\..\tests\datagencli.c" />
</ItemGroup> </ItemGroup>
<ItemGroup> <ItemGroup>
build/cmake/tests/CMakeLists.txt
+1 -1
View File
@@ -50,7 +50,7 @@ set(PROGRAMS_DIR ${ZSTD_SOURCE_DIR}/programs)
set(TESTS_DIR ${ZSTD_SOURCE_DIR}/tests) set(TESTS_DIR ${ZSTD_SOURCE_DIR}/tests)
include_directories(${TESTS_DIR} ${PROGRAMS_DIR} ${LIBRARY_DIR} ${LIBRARY_DIR}/common ${LIBRARY_DIR}/compress ${LIBRARY_DIR}/dictBuilder) include_directories(${TESTS_DIR} ${PROGRAMS_DIR} ${LIBRARY_DIR} ${LIBRARY_DIR}/common ${LIBRARY_DIR}/compress ${LIBRARY_DIR}/dictBuilder)
add_executable(datagen ${PROGRAMS_DIR}/datagen.c ${TESTS_DIR}/datagencli.c) add_executable(datagen ${PROGRAMS_DIR}/datagen.c ${PROGRAMS_DIR}/lorem.c ${TESTS_DIR}/loremOut.c ${TESTS_DIR}/datagencli.c)
target_link_libraries(datagen libzstd_static) target_link_libraries(datagen libzstd_static)
# #
build/meson/tests/meson.build
+2 -1
View File
@@ -44,7 +44,8 @@ testcommon_dep = declare_dependency(link_with: testcommon,
dependencies: libzstd_deps, dependencies: libzstd_deps,
include_directories: libzstd_includes) include_directories: libzstd_includes)
datagen_sources = [join_paths(zstd_rootdir, 'tests/datagencli.c')] datagen_sources = [join_paths(zstd_rootdir, 'tests/datagencli.c'),
join_paths(zstd_rootdir, 'tests/loremOut.c')]
datagen = executable('datagen', datagen = executable('datagen',
datagen_sources, datagen_sources,
c_args: [ '-DNDEBUG' ], c_args: [ '-DNDEBUG' ],
lib/common/xxhash.h
-43
View File
@@ -630,7 +630,6 @@ XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32 (const void* input, size_t length, X
* @brief The opaque state struct for the XXH32 streaming API. * @brief The opaque state struct for the XXH32 streaming API.
* *
* @see XXH32_state_s for details. * @see XXH32_state_s for details.
* @see @ref streaming_example "Streaming Example"
*/ */
typedef struct XXH32_state_s XXH32_state_t; typedef struct XXH32_state_s XXH32_state_t;
@@ -641,8 +640,6 @@ typedef struct XXH32_state_s XXH32_state_t;
* @return `NULL` on failure. * @return `NULL` on failure.
* *
* @note Must be freed with XXH32_freeState(). * @note Must be freed with XXH32_freeState().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void); XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void);
/*! /*!
@@ -654,8 +651,6 @@ XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void);
* *
* @note @p statePtr must be allocated with XXH32_createState(). * @note @p statePtr must be allocated with XXH32_createState().
* *
* @see @ref streaming_example "Streaming Example"
*
*/ */
XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr); XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr);
/*! /*!
@@ -681,8 +676,6 @@ XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_
* @return @ref XXH_ERROR on failure. * @return @ref XXH_ERROR on failure.
* *
* @note This function resets and seeds a state. Call it before @ref XXH32_update(). * @note This function resets and seeds a state. Call it before @ref XXH32_update().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed); XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed);
@@ -704,8 +697,6 @@ XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t
* @return @ref XXH_ERROR on failure. * @return @ref XXH_ERROR on failure.
* *
* @note Call this to incrementally consume blocks of data. * @note Call this to incrementally consume blocks of data.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length); XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length);
@@ -722,8 +713,6 @@ XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void*
* @note * @note
* Calling XXH32_digest() will not affect @p statePtr, so you can update, * Calling XXH32_digest() will not affect @p statePtr, so you can update,
* digest, and update again. * digest, and update again.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr); XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr);
#endif /* !XXH_NO_STREAM */ #endif /* !XXH_NO_STREAM */
@@ -900,7 +889,6 @@ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64(XXH_NOESCAPE const void* input, size
* @brief The opaque state struct for the XXH64 streaming API. * @brief The opaque state struct for the XXH64 streaming API.
* *
* @see XXH64_state_s for details. * @see XXH64_state_s for details.
* @see @ref streaming_example "Streaming Example"
*/ */
typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */ typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */
@@ -911,8 +899,6 @@ typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */
* @return `NULL` on failure. * @return `NULL` on failure.
* *
* @note Must be freed with XXH64_freeState(). * @note Must be freed with XXH64_freeState().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void); XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void);
@@ -924,8 +910,6 @@ XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void);
* @return @ref XXH_OK. * @return @ref XXH_OK.
* *
* @note @p statePtr must be allocated with XXH64_createState(). * @note @p statePtr must be allocated with XXH64_createState().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr); XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr);
@@ -952,8 +936,6 @@ XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dst_state, const
* @return @ref XXH_ERROR on failure. * @return @ref XXH_ERROR on failure.
* *
* @note This function resets and seeds a state. Call it before @ref XXH64_update(). * @note This function resets and seeds a state. Call it before @ref XXH64_update().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed); XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed);
@@ -975,8 +957,6 @@ XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr,
* @return @ref XXH_ERROR on failure. * @return @ref XXH_ERROR on failure.
* *
* @note Call this to incrementally consume blocks of data. * @note Call this to incrementally consume blocks of data.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length);
@@ -993,8 +973,6 @@ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr,
* @note * @note
* Calling XXH64_digest() will not affect @p statePtr, so you can update, * Calling XXH64_digest() will not affect @p statePtr, so you can update,
* digest, and update again. * digest, and update again.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr); XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr);
#endif /* !XXH_NO_STREAM */ #endif /* !XXH_NO_STREAM */
@@ -1199,7 +1177,6 @@ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const
* @brief The opaque state struct for the XXH3 streaming API. * @brief The opaque state struct for the XXH3 streaming API.
* *
* @see XXH3_state_s for details. * @see XXH3_state_s for details.
* @see @ref streaming_example "Streaming Example"
*/ */
typedef struct XXH3_state_s XXH3_state_t; typedef struct XXH3_state_s XXH3_state_t;
XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void); XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void);
@@ -1231,8 +1208,6 @@ XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOE
* - Call this function before @ref XXH3_64bits_update(). * - Call this function before @ref XXH3_64bits_update().
* - Digest will be equivalent to `XXH3_64bits()`. * - Digest will be equivalent to `XXH3_64bits()`.
* *
* @see @ref streaming_example "Streaming Example"
*
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr);
@@ -1253,8 +1228,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* stateP
* - Call this function before @ref XXH3_64bits_update(). * - Call this function before @ref XXH3_64bits_update().
* - Digest will be equivalent to `XXH3_64bits_withSeed()`. * - Digest will be equivalent to `XXH3_64bits_withSeed()`.
* *
* @see @ref streaming_example "Streaming Example"
*
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed);
@@ -1279,8 +1252,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_
* (secret's content should look like a bunch of random bytes). * (secret's content should look like a bunch of random bytes).
* When in doubt about the randomness of a candidate `secret`, * When in doubt about the randomness of a candidate `secret`,
* consider employing `XXH3_generateSecret()` instead (see below). * consider employing `XXH3_generateSecret()` instead (see below).
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize);
@@ -1302,8 +1273,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_stat
* @return @ref XXH_ERROR on failure. * @return @ref XXH_ERROR on failure.
* *
* @note Call this to incrementally consume blocks of data. * @note Call this to incrementally consume blocks of data.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length);
@@ -1320,8 +1289,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* stat
* @note * @note
* Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update, * Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update,
* digest, and update again. * digest, and update again.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr);
#endif /* !XXH_NO_STREAM */ #endif /* !XXH_NO_STREAM */
@@ -1442,8 +1409,6 @@ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE cons
* - This function resets `statePtr` and generate a secret with default parameters. * - This function resets `statePtr` and generate a secret with default parameters.
* - Call it before @ref XXH3_128bits_update(). * - Call it before @ref XXH3_128bits_update().
* - Digest will be equivalent to `XXH3_128bits()`. * - Digest will be equivalent to `XXH3_128bits()`.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr);
@@ -1463,8 +1428,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* state
* - This function resets `statePtr` and generate a secret from `seed`. * - This function resets `statePtr` and generate a secret from `seed`.
* - Call it before @ref XXH3_128bits_update(). * - Call it before @ref XXH3_128bits_update().
* - Digest will be equivalent to `XXH3_128bits_withSeed()`. * - Digest will be equivalent to `XXH3_128bits_withSeed()`.
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed);
/*! /*!
@@ -1486,8 +1449,6 @@ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state
* (secret's content should look like a bunch of random bytes). * (secret's content should look like a bunch of random bytes).
* When in doubt about the randomness of a candidate `secret`, * When in doubt about the randomness of a candidate `secret`,
* consider employing `XXH3_generateSecret()` instead (see below). * consider employing `XXH3_generateSecret()` instead (see below).
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize);
@@ -6072,8 +6033,6 @@ static void XXH_alignedFree(void* p)
* @return `NULL` on failure. * @return `NULL` on failure.
* *
* @note Must be freed with XXH3_freeState(). * @note Must be freed with XXH3_freeState().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void) XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void)
{ {
@@ -6092,8 +6051,6 @@ XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void)
* @return @ref XXH_OK. * @return @ref XXH_OK.
* *
* @note Must be allocated with XXH3_createState(). * @note Must be allocated with XXH3_createState().
*
* @see @ref streaming_example "Streaming Example"
*/ */
XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr) XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr)
{ {
programs/benchzstd.c
+627 -287
View File
@@ -8,29 +8,29 @@
* You may select, at your option, one of the above-listed licenses. * You may select, at your option, one of the above-listed licenses.
*/ */
/* ************************************** /* **************************************
* Tuning parameters * Tuning parameters
****************************************/ ****************************************/
#ifndef BMK_TIMETEST_DEFAULT_S /* default minimum time per test */ #ifndef BMK_TIMETEST_DEFAULT_S /* default minimum time per test */
# define BMK_TIMETEST_DEFAULT_S 3 # define BMK_TIMETEST_DEFAULT_S 3
#endif #endif
/* ************************************* /* *************************************
* Includes * Includes
***************************************/ ***************************************/
#include "platform.h" /* Large Files support */ /* this must be included first */
#include "util.h" /* UTIL_getFileSize, UTIL_sleep */ #include "platform.h" /* Large Files support, compiler specifics */
/* then following system includes */
#include <assert.h> /* assert */
#include <errno.h>
#include <stdio.h> /* fprintf, fopen */
#include <stdlib.h> /* malloc, free */ #include <stdlib.h> /* malloc, free */
#include <string.h> /* memset, strerror */ #include <string.h> /* memset, strerror */
#include <stdio.h> /* fprintf, fopen */ #include "util.h" /* UTIL_getFileSize, UTIL_sleep */
#include <errno.h>
#include <assert.h> /* assert */
#include "timefn.h" /* UTIL_time_t */
#include "benchfn.h"
#include "../lib/common/mem.h" #include "../lib/common/mem.h"
#include "benchfn.h"
#include "timefn.h" /* UTIL_time_t */
#ifndef ZSTD_STATIC_LINKING_ONLY #ifndef ZSTD_STATIC_LINKING_ONLY
# define ZSTD_STATIC_LINKING_ONLY # define ZSTD_STATIC_LINKING_ONLY
#endif #endif
@@ -41,74 +41,94 @@
# define XXH_INLINE_ALL # define XXH_INLINE_ALL
#endif #endif
#include "../lib/common/xxhash.h" #include "../lib/common/xxhash.h"
#include "benchzstd.h"
#include "../lib/zstd_errors.h" #include "../lib/zstd_errors.h"
#include "benchzstd.h"
/* ************************************* /* *************************************
* Constants * Constants
***************************************/ ***************************************/
#ifndef ZSTD_GIT_COMMIT #ifndef ZSTD_GIT_COMMIT
# define ZSTD_GIT_COMMIT_STRING "" # define ZSTD_GIT_COMMIT_STRING ""
#else #else
# define ZSTD_GIT_COMMIT_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_GIT_COMMIT) # define ZSTD_GIT_COMMIT_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_GIT_COMMIT)
#endif #endif
#define TIMELOOP_MICROSEC (1*1000000ULL) /* 1 second */ #define TIMELOOP_MICROSEC (1 * 1000000ULL) /* 1 second */
#define TIMELOOP_NANOSEC (1*1000000000ULL) /* 1 second */ #define TIMELOOP_NANOSEC (1 * 1000000000ULL) /* 1 second */
#define ACTIVEPERIOD_MICROSEC (70*TIMELOOP_MICROSEC) /* 70 seconds */ #define ACTIVEPERIOD_MICROSEC (70 * TIMELOOP_MICROSEC) /* 70 seconds */
#define COOLPERIOD_SEC 10 #define COOLPERIOD_SEC 10
#define KB *(1 <<10) #define KB *(1 << 10)
#define MB *(1 <<20) #define MB *(1 << 20)
#define GB *(1U<<30) #define GB *(1U << 30)
#define BMK_RUNTEST_DEFAULT_MS 1000 #define BMK_RUNTEST_DEFAULT_MS 1000
static const size_t maxMemory = (sizeof(size_t)==4) ? static const size_t maxMemory = (sizeof(size_t) == 4)
/* 32-bit */ (2 GB - 64 MB) : ?
/* 64-bit */ (size_t)(1ULL << ((sizeof(size_t)*8)-31)); /* 32-bit */ (2 GB - 64 MB)
:
/* 64-bit */ (size_t)(1ULL << ((sizeof(size_t) * 8) - 31));
/* ************************************* /* *************************************
* console display * console display
***************************************/ ***************************************/
#define DISPLAY(...) { fprintf(stderr, __VA_ARGS__); fflush(NULL); } #define DISPLAY(...) \
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); } { \
/* 0 : no display; 1: errors; 2 : + result + interaction + warnings; 3 : + progression; 4 : + information */ fprintf(stderr, __VA_ARGS__); \
#define OUTPUT(...) { fprintf(stdout, __VA_ARGS__); fflush(NULL); } fflush(NULL); \
#define OUTPUTLEVEL(l, ...) if (displayLevel>=l) { OUTPUT(__VA_ARGS__); } }
#define DISPLAYLEVEL(l, ...) \
if (displayLevel >= l) { \
DISPLAY(__VA_ARGS__); \
}
/* 0 : no display; 1: errors; 2 : + result + interaction + warnings; 3 : +
* progression; 4 : + information */
#define OUTPUT(...) \
{ \
fprintf(stdout, __VA_ARGS__); \
fflush(NULL); \
}
#define OUTPUTLEVEL(l, ...) \
if (displayLevel >= l) { \
OUTPUT(__VA_ARGS__); \
}
/* ************************************* /* *************************************
* Exceptions * Exceptions
***************************************/ ***************************************/
#ifndef DEBUG #ifndef DEBUG
# define DEBUG 0 # define DEBUG 0
#endif #endif
#define DEBUGOUTPUT(...) { if (DEBUG) DISPLAY(__VA_ARGS__); } #define DEBUGOUTPUT(...) \
{ \
if (DEBUG) \
DISPLAY(__VA_ARGS__); \
}
#define RETURN_ERROR_INT(errorNum, ...) { \ #define RETURN_ERROR_INT(errorNum, ...) \
{ \
DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \ DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \
DISPLAYLEVEL(1, "Error %i : ", errorNum); \ DISPLAYLEVEL(1, "Error %i : ", errorNum); \
DISPLAYLEVEL(1, __VA_ARGS__); \ DISPLAYLEVEL(1, __VA_ARGS__); \
DISPLAYLEVEL(1, " \n"); \ DISPLAYLEVEL(1, " \n"); \
return errorNum; \ return errorNum; \
} }
#define CHECK_Z(zf) { \ #define CHECK_Z(zf) \
{ \
size_t const zerr = zf; \ size_t const zerr = zf; \
if (ZSTD_isError(zerr)) { \ if (ZSTD_isError(zerr)) { \
DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \ DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \
DISPLAY("Error : "); \ DISPLAY("Error : "); \
DISPLAY("%s failed : %s", \ DISPLAY("%s failed : %s", #zf, ZSTD_getErrorName(zerr)); \
#zf, ZSTD_getErrorName(zerr)); \
DISPLAY(" \n"); \ DISPLAY(" \n"); \
exit(1); \ exit(1); \
} \ } \
} }
#define RETURN_ERROR(errorNum, retType, ...) { \ #define RETURN_ERROR(errorNum, retType, ...) \
{ \
retType r; \ retType r; \
memset(&r, 0, sizeof(retType)); \ memset(&r, 0, sizeof(retType)); \
DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \ DEBUGOUTPUT("%s: %i: \n", __FILE__, __LINE__); \
@@ -117,14 +137,65 @@ static const size_t maxMemory = (sizeof(size_t)==4) ?
DISPLAYLEVEL(1, " \n"); \ DISPLAYLEVEL(1, " \n"); \
r.tag = errorNum; \ r.tag = errorNum; \
return r; \ return r; \
}
/* replacement for snprintf(), which is not supported by C89
* sprintf() would be the supported one, but it's labelled unsafe,
* so some modern static analyzer will flag it as such, making it unusable.
* formatString_u() replaces snprintf() for the specific case where there are only %u arguments */
static int formatString_u(char* buffer, size_t buffer_size, const char* formatString, unsigned int value)
{
size_t written = 0;
int i;
assert(value <= 100);
for (i = 0; formatString[i] != '\0' && written < buffer_size - 1; ++i) {
if (formatString[i] != '%') {
buffer[written++] = formatString[i];
continue;
}
if (formatString[++i] == 'u') {
/* Handle single digit */
if (value < 10) {
buffer[written++] = '0' + (char)value;
} else if (value < 100) {
/* Handle two digits */
if (written >= buffer_size - 2) {
return -1; /* buffer overflow */
}
buffer[written++] = '0' + (char)(value / 10);
buffer[written++] = '0' + (char)(value % 10);
} else { /* 100 */
if (written >= buffer_size - 3) {
return -1; /* buffer overflow */
}
buffer[written++] = '1';
buffer[written++] = '0';
buffer[written++] = '0';
}
} else if (formatString[i] == '%') { /* Check for escaped percent sign */
buffer[written++] = '%';
} else {
return -1; /* unsupported format */
}
}
if (written < buffer_size) {
buffer[written] = '\0';
} else {
buffer[0] = '\0'; /* Handle truncation */
}
return (int)written;
} }
/* ************************************* /* *************************************
* Benchmark Parameters * Benchmark Parameters
***************************************/ ***************************************/
BMK_advancedParams_t BMK_initAdvancedParams(void) { BMK_advancedParams_t BMK_initAdvancedParams(void)
{
BMK_advancedParams_t const res = { BMK_advancedParams_t const res = {
BMK_both, /* mode */ BMK_both, /* mode */
BMK_TIMETEST_DEFAULT_S, /* nbSeconds */ BMK_TIMETEST_DEFAULT_S, /* nbSeconds */
@@ -143,10 +214,9 @@ BMK_advancedParams_t BMK_initAdvancedParams(void) {
return res; return res;
} }
/* ******************************************************** /* ********************************************************
* Bench functions * Bench functions
**********************************************************/ **********************************************************/
typedef struct { typedef struct {
const void* srcPtr; const void* srcPtr;
size_t srcSize; size_t srcSize;
@@ -159,47 +229,62 @@ typedef struct {
#undef MIN #undef MIN
#undef MAX #undef MAX
#define MIN(a,b) ((a) < (b) ? (a) : (b)) #define MIN(a, b) ((a) < (b) ? (a) : (b))
#define MAX(a,b) ((a) > (b) ? (a) : (b)) #define MAX(a, b) ((a) > (b) ? (a) : (b))
static void static void BMK_initCCtx(
BMK_initCCtx(ZSTD_CCtx* ctx, ZSTD_CCtx* ctx,
const void* dictBuffer, size_t dictBufferSize, const void* dictBuffer,
size_t dictBufferSize,
int cLevel, int cLevel,
const ZSTD_compressionParameters* comprParams, const ZSTD_compressionParameters* comprParams,
const BMK_advancedParams_t* adv) const BMK_advancedParams_t* adv)
{ {
ZSTD_CCtx_reset(ctx, ZSTD_reset_session_and_parameters); ZSTD_CCtx_reset(ctx, ZSTD_reset_session_and_parameters);
if (adv->nbWorkers==1) { if (adv->nbWorkers == 1) {
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_nbWorkers, 0)); CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_nbWorkers, 0));
} else { } else {
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_nbWorkers, adv->nbWorkers)); CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_nbWorkers, adv->nbWorkers));
} }
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_compressionLevel, cLevel)); CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_compressionLevel, cLevel));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_useRowMatchFinder, adv->useRowMatchFinder)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_enableLongDistanceMatching, adv->ldmFlag)); ctx, ZSTD_c_useRowMatchFinder, adv->useRowMatchFinder));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx, ZSTD_c_enableLongDistanceMatching, adv->ldmFlag));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmMinMatch, adv->ldmMinMatch)); CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmMinMatch, adv->ldmMinMatch));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmHashLog, adv->ldmHashLog)); CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmHashLog, adv->ldmHashLog));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmBucketSizeLog, adv->ldmBucketSizeLog)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_ldmHashRateLog, adv->ldmHashRateLog)); ctx, ZSTD_c_ldmBucketSizeLog, adv->ldmBucketSizeLog));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_windowLog, (int)comprParams->windowLog)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_hashLog, (int)comprParams->hashLog)); ctx, ZSTD_c_ldmHashRateLog, adv->ldmHashRateLog));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_chainLog, (int)comprParams->chainLog)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_searchLog, (int)comprParams->searchLog)); ctx, ZSTD_c_windowLog, (int)comprParams->windowLog));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_minMatch, (int)comprParams->minMatch)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_targetLength, (int)comprParams->targetLength)); ctx, ZSTD_c_hashLog, (int)comprParams->hashLog));
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_literalCompressionMode, (int)adv->literalCompressionMode)); CHECK_Z(ZSTD_CCtx_setParameter(
CHECK_Z(ZSTD_CCtx_setParameter(ctx, ZSTD_c_strategy, (int)comprParams->strategy)); ctx, ZSTD_c_chainLog, (int)comprParams->chainLog));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx, ZSTD_c_searchLog, (int)comprParams->searchLog));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx, ZSTD_c_minMatch, (int)comprParams->minMatch));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx, ZSTD_c_targetLength, (int)comprParams->targetLength));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx,
ZSTD_c_literalCompressionMode,
(int)adv->literalCompressionMode));
CHECK_Z(ZSTD_CCtx_setParameter(
ctx, ZSTD_c_strategy, (int)comprParams->strategy));
CHECK_Z(ZSTD_CCtx_loadDictionary(ctx, dictBuffer, dictBufferSize)); CHECK_Z(ZSTD_CCtx_loadDictionary(ctx, dictBuffer, dictBufferSize));
} }
static void BMK_initDCtx(ZSTD_DCtx* dctx, static void
const void* dictBuffer, size_t dictBufferSize) { BMK_initDCtx(ZSTD_DCtx* dctx, const void* dictBuffer, size_t dictBufferSize)
{
CHECK_Z(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_and_parameters)); CHECK_Z(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_and_parameters));
CHECK_Z(ZSTD_DCtx_loadDictionary(dctx, dictBuffer, dictBufferSize)); CHECK_Z(ZSTD_DCtx_loadDictionary(dctx, dictBuffer, dictBufferSize));
} }
typedef struct { typedef struct {
ZSTD_CCtx* cctx; ZSTD_CCtx* cctx;
const void* dictBuffer; const void* dictBuffer;
@@ -209,9 +294,16 @@ typedef struct {
const BMK_advancedParams_t* adv; const BMK_advancedParams_t* adv;
} BMK_initCCtxArgs; } BMK_initCCtxArgs;
static size_t local_initCCtx(void* payload) { static size_t local_initCCtx(void* payload)
{
BMK_initCCtxArgs* ag = (BMK_initCCtxArgs*)payload; BMK_initCCtxArgs* ag = (BMK_initCCtxArgs*)payload;
BMK_initCCtx(ag->cctx, ag->dictBuffer, ag->dictBufferSize, ag->cLevel, ag->comprParams, ag->adv); BMK_initCCtx(
ag->cctx,
ag->dictBuffer,
ag->dictBufferSize,
ag->cLevel,
ag->comprParams,
ag->adv);
return 0; return 0;
} }
@@ -221,17 +313,19 @@ typedef struct {
size_t dictBufferSize; size_t dictBufferSize;
} BMK_initDCtxArgs; } BMK_initDCtxArgs;
static size_t local_initDCtx(void* payload) { static size_t local_initDCtx(void* payload)
{
BMK_initDCtxArgs* ag = (BMK_initDCtxArgs*)payload; BMK_initDCtxArgs* ag = (BMK_initDCtxArgs*)payload;
BMK_initDCtx(ag->dctx, ag->dictBuffer, ag->dictBufferSize); BMK_initDCtx(ag->dctx, ag->dictBuffer, ag->dictBufferSize);
return 0; return 0;
} }
/* `addArgs` is the context */ /* `addArgs` is the context */
static size_t local_defaultCompress( static size_t local_defaultCompress(
const void* srcBuffer, size_t srcSize, const void* srcBuffer,
void* dstBuffer, size_t dstSize, size_t srcSize,
void* dstBuffer,
size_t dstSize,
void* addArgs) void* addArgs)
{ {
ZSTD_CCtx* const cctx = (ZSTD_CCtx*)addArgs; ZSTD_CCtx* const cctx = (ZSTD_CCtx*)addArgs;
@@ -240,18 +334,24 @@ static size_t local_defaultCompress(
/* `addArgs` is the context */ /* `addArgs` is the context */
static size_t local_defaultDecompress( static size_t local_defaultDecompress(
const void* srcBuffer, size_t srcSize, const void* srcBuffer,
void* dstBuffer, size_t dstCapacity, size_t srcSize,
void* dstBuffer,
size_t dstCapacity,
void* addArgs) void* addArgs)
{ {
size_t moreToFlush = 1; size_t moreToFlush = 1;
ZSTD_DCtx* const dctx = (ZSTD_DCtx*)addArgs; ZSTD_DCtx* const dctx = (ZSTD_DCtx*)addArgs;
ZSTD_inBuffer in; ZSTD_inBuffer in;
ZSTD_outBuffer out; ZSTD_outBuffer out;
in.src = srcBuffer; in.size = srcSize; in.pos = 0; in.src = srcBuffer;
out.dst = dstBuffer; out.size = dstCapacity; out.pos = 0; in.size = srcSize;
in.pos = 0;
out.dst = dstBuffer;
out.size = dstCapacity;
out.pos = 0;
while (moreToFlush) { while (moreToFlush) {
if(out.pos == out.size) { if (out.pos == out.size) {
return (size_t)-ZSTD_error_dstSize_tooSmall; return (size_t)-ZSTD_error_dstSize_tooSmall;
} }
moreToFlush = ZSTD_decompressStream(dctx, &out, &in); moreToFlush = ZSTD_decompressStream(dctx, &out, &in);
@@ -260,10 +360,8 @@ static size_t local_defaultDecompress(
} }
} }
return out.pos; return out.pos;
} }
/* ================================================================= */ /* ================================================================= */
/* Benchmark Zstandard, mem-to-mem scenarios */ /* Benchmark Zstandard, mem-to-mem scenarios */
/* ================================================================= */ /* ================================================================= */
@@ -287,7 +385,8 @@ static BMK_benchOutcome_t BMK_benchOutcome_error(void)
return b; return b;
} }
static BMK_benchOutcome_t BMK_benchOutcome_setValidResult(BMK_benchResult_t result) static BMK_benchOutcome_t BMK_benchOutcome_setValidResult(
BMK_benchResult_t result)
{ {
BMK_benchOutcome_t b; BMK_benchOutcome_t b;
b.tag = 0; b.tag = 0;
@@ -295,64 +394,93 @@ static BMK_benchOutcome_t BMK_benchOutcome_setValidResult(BMK_benchResult_t resu
return b; return b;
} }
/* benchMem with no allocation */ /* benchMem with no allocation */
static BMK_benchOutcome_t static BMK_benchOutcome_t BMK_benchMemAdvancedNoAlloc(
BMK_benchMemAdvancedNoAlloc( const void** srcPtrs,
const void** srcPtrs, size_t* srcSizes, size_t* srcSizes,
void** cPtrs, size_t* cCapacities, size_t* cSizes, void** cPtrs,
void** resPtrs, size_t* resSizes, size_t* cCapacities,
void** resultBufferPtr, void* compressedBuffer, size_t* cSizes,
void** resPtrs,
size_t* resSizes,
void** resultBufferPtr,
void* compressedBuffer,
size_t maxCompressedSize, size_t maxCompressedSize,
BMK_timedFnState_t* timeStateCompress, BMK_timedFnState_t* timeStateCompress,
BMK_timedFnState_t* timeStateDecompress, BMK_timedFnState_t* timeStateDecompress,
const void* srcBuffer, size_t srcSize, const void* srcBuffer,
const size_t* fileSizes, unsigned nbFiles, size_t srcSize,
const size_t* fileSizes,
unsigned nbFiles,
const int cLevel, const int cLevel,
const ZSTD_compressionParameters* comprParams, const ZSTD_compressionParameters* comprParams,
const void* dictBuffer, size_t dictBufferSize, const void* dictBuffer,
ZSTD_CCtx* cctx, ZSTD_DCtx* dctx, size_t dictBufferSize,
int displayLevel, const char* displayName, ZSTD_CCtx* cctx,
ZSTD_DCtx* dctx,
int displayLevel,
const char* displayName,
const BMK_advancedParams_t* adv) const BMK_advancedParams_t* adv)
{ {
size_t const blockSize = ((adv->blockSize>=32 && (adv->mode != BMK_decodeOnly)) ? adv->blockSize : srcSize) + (!srcSize); /* avoid div by 0 */ size_t const blockSize =
((adv->blockSize >= 32 && (adv->mode != BMK_decodeOnly))
? adv->blockSize
: srcSize)
+ (!srcSize); /* avoid div by 0 */
BMK_benchResult_t benchResult; BMK_benchResult_t benchResult;
size_t const loadedCompressedSize = srcSize; size_t const loadedCompressedSize = srcSize;
size_t cSize = 0; size_t cSize = 0;
double ratio = 0.; double ratio = 0.;
U32 nbBlocks; U32 nbBlocks;
assert(cctx != NULL); assert(dctx != NULL); assert(cctx != NULL);
assert(dctx != NULL);
/* init */ /* init */
memset(&benchResult, 0, sizeof(benchResult)); memset(&benchResult, 0, sizeof(benchResult));
if (strlen(displayName)>17) displayName += strlen(displayName) - 17; /* display last 17 characters */ if (strlen(displayName) > 17)
displayName +=
strlen(displayName) - 17; /* display last 17 characters */
if (adv->mode == BMK_decodeOnly) { if (adv->mode == BMK_decodeOnly) {
/* benchmark only decompression : source must be already compressed */ /* benchmark only decompression : source must be already compressed */
const char* srcPtr = (const char*)srcBuffer; const char* srcPtr = (const char*)srcBuffer;
U64 totalDSize64 = 0; U64 totalDSize64 = 0;
U32 fileNb; U32 fileNb;
for (fileNb=0; fileNb<nbFiles; fileNb++) { for (fileNb = 0; fileNb < nbFiles; fileNb++) {
U64 const fSize64 = ZSTD_findDecompressedSize(srcPtr, fileSizes[fileNb]); U64 const fSize64 =
ZSTD_findDecompressedSize(srcPtr, fileSizes[fileNb]);
if (fSize64 == ZSTD_CONTENTSIZE_UNKNOWN) { if (fSize64 == ZSTD_CONTENTSIZE_UNKNOWN) {
RETURN_ERROR(32, BMK_benchOutcome_t, "Decompressed size cannot be determined: cannot benchmark"); RETURN_ERROR(
32,
BMK_benchOutcome_t,
"Decompressed size cannot be determined: cannot benchmark");
} }
if (fSize64 == ZSTD_CONTENTSIZE_ERROR) { if (fSize64 == ZSTD_CONTENTSIZE_ERROR) {
RETURN_ERROR(32, BMK_benchOutcome_t, "Error while trying to assess decompressed size: data may be invalid"); RETURN_ERROR(
32,
BMK_benchOutcome_t,
"Error while trying to assess decompressed size: data may be invalid");
} }
totalDSize64 += fSize64; totalDSize64 += fSize64;
srcPtr += fileSizes[fileNb]; srcPtr += fileSizes[fileNb];
} }
{ size_t const decodedSize = (size_t)totalDSize64; {
size_t const decodedSize = (size_t)totalDSize64;
assert((U64)decodedSize == totalDSize64); /* check overflow */ assert((U64)decodedSize == totalDSize64); /* check overflow */
free(*resultBufferPtr); free(*resultBufferPtr);
if (totalDSize64 > decodedSize) { /* size_t overflow */ if (totalDSize64 > decodedSize) { /* size_t overflow */
RETURN_ERROR(32, BMK_benchOutcome_t, "decompressed size is too large for local system"); RETURN_ERROR(
32,
BMK_benchOutcome_t,
"decompressed size is too large for local system");
} }
*resultBufferPtr = malloc(decodedSize); *resultBufferPtr = malloc(decodedSize);
if (!(*resultBufferPtr)) { if (!(*resultBufferPtr)) {
RETURN_ERROR(33, BMK_benchOutcome_t, "allocation error: not enough memory"); RETURN_ERROR(
33,
BMK_benchOutcome_t,
"allocation error: not enough memory");
} }
cSize = srcSize; cSize = srcSize;
srcSize = decodedSize; srcSize = decodedSize;
@@ -361,22 +489,30 @@ BMK_benchMemAdvancedNoAlloc(
} }
/* Init data blocks */ /* Init data blocks */
{ const char* srcPtr = (const char*)srcBuffer; {
const char* srcPtr = (const char*)srcBuffer;
char* cPtr = (char*)compressedBuffer; char* cPtr = (char*)compressedBuffer;
char* resPtr = (char*)(*resultBufferPtr); char* resPtr = (char*)(*resultBufferPtr);
U32 fileNb; U32 fileNb;
for (nbBlocks=0, fileNb=0; fileNb<nbFiles; fileNb++) { for (nbBlocks = 0, fileNb = 0; fileNb < nbFiles; fileNb++) {
size_t remaining = fileSizes[fileNb]; size_t remaining = fileSizes[fileNb];
U32 const nbBlocksforThisFile = (adv->mode == BMK_decodeOnly) ? 1 : (U32)((remaining + (blockSize-1)) / blockSize); U32 const nbBlocksforThisFile = (adv->mode == BMK_decodeOnly)
? 1
: (U32)((remaining + (blockSize - 1)) / blockSize);
U32 const blockEnd = nbBlocks + nbBlocksforThisFile; U32 const blockEnd = nbBlocks + nbBlocksforThisFile;
for ( ; nbBlocks<blockEnd; nbBlocks++) { for (; nbBlocks < blockEnd; nbBlocks++) {
size_t const thisBlockSize = MIN(remaining, blockSize); size_t const thisBlockSize = MIN(remaining, blockSize);
srcPtrs[nbBlocks] = srcPtr; srcPtrs[nbBlocks] = srcPtr;
srcSizes[nbBlocks] = thisBlockSize; srcSizes[nbBlocks] = thisBlockSize;
cPtrs[nbBlocks] = cPtr; cPtrs[nbBlocks] = cPtr;
cCapacities[nbBlocks] = (adv->mode == BMK_decodeOnly) ? thisBlockSize : ZSTD_compressBound(thisBlockSize); cCapacities[nbBlocks] = (adv->mode == BMK_decodeOnly)
? thisBlockSize
: ZSTD_compressBound(thisBlockSize);
resPtrs[nbBlocks] = resPtr; resPtrs[nbBlocks] = resPtr;
resSizes[nbBlocks] = (adv->mode == BMK_decodeOnly) ? (size_t) ZSTD_findDecompressedSize(srcPtr, thisBlockSize) : thisBlockSize; resSizes[nbBlocks] = (adv->mode == BMK_decodeOnly)
? (size_t)ZSTD_findDecompressedSize(
srcPtr, thisBlockSize)
: thisBlockSize;
srcPtr += thisBlockSize; srcPtr += thisBlockSize;
cPtr += cCapacities[nbBlocks]; cPtr += cCapacities[nbBlocks];
resPtr += thisBlockSize; resPtr += thisBlockSize;
@@ -384,7 +520,10 @@ BMK_benchMemAdvancedNoAlloc(
if (adv->mode == BMK_decodeOnly) { if (adv->mode == BMK_decodeOnly) {
cSizes[nbBlocks] = thisBlockSize; cSizes[nbBlocks] = thisBlockSize;
benchResult.cSize = thisBlockSize; benchResult.cSize = thisBlockSize;
} } } } }
}
}
}
/* warming up `compressedBuffer` */ /* warming up `compressedBuffer` */
if (adv->mode == BMK_decodeOnly) { if (adv->mode == BMK_decodeOnly) {
@@ -394,12 +533,17 @@ BMK_benchMemAdvancedNoAlloc(
} }
if (!UTIL_support_MT_measurements() && adv->nbWorkers > 1) { if (!UTIL_support_MT_measurements() && adv->nbWorkers > 1) {
OUTPUTLEVEL(2, "Warning : time measurements may be incorrect in multithreading mode... \n") OUTPUTLEVEL(
2,
"Warning : time measurements may be incorrect in multithreading mode... \n")
} }
/* Bench */ /* Bench */
{ U64 const crcOrig = (adv->mode == BMK_decodeOnly) ? 0 : XXH64(srcBuffer, srcSize, 0); {
# define NB_MARKS 4 U64 const crcOrig = (adv->mode == BMK_decodeOnly)
? 0
: XXH64(srcBuffer, srcSize, 0);
#define NB_MARKS 4
const char* marks[NB_MARKS] = { " |", " /", " =", " \\" }; const char* marks[NB_MARKS] = { " |", " /", " =", " \\" };
U32 markNb = 0; U32 markNb = 0;
int compressionCompleted = (adv->mode == BMK_decodeOnly); int compressionCompleted = (adv->mode == BMK_decodeOnly);
@@ -433,7 +577,7 @@ BMK_benchMemAdvancedNoAlloc(
dbp.initPayload = &dctxprep; dbp.initPayload = &dctxprep;
dbp.errorFn = ZSTD_isError; dbp.errorFn = ZSTD_isError;
dbp.blockCount = nbBlocks; dbp.blockCount = nbBlocks;
dbp.srcBuffers = (const void* const *) cPtrs; dbp.srcBuffers = (const void* const*)cPtrs;
dbp.srcSizes = cSizes; dbp.srcSizes = cSizes;
dbp.dstBuffers = resPtrs; dbp.dstBuffers = resPtrs;
dbp.dstCapacities = resSizes; dbp.dstCapacities = resSizes;
@@ -445,185 +589,273 @@ BMK_benchMemAdvancedNoAlloc(
OUTPUTLEVEL(2, "\r%70s\r", ""); /* blank line */ OUTPUTLEVEL(2, "\r%70s\r", ""); /* blank line */
assert(srcSize < UINT_MAX); assert(srcSize < UINT_MAX);
OUTPUTLEVEL(2, "%2s-%-17.17s :%10u -> \r", marks[markNb], displayName, (unsigned)srcSize); OUTPUTLEVEL(
2,
"%2s-%-17.17s :%10u -> \r",
marks[markNb],
displayName,
(unsigned)srcSize);
while (!(compressionCompleted && decompressionCompleted)) { while (!(compressionCompleted && decompressionCompleted)) {
if (!compressionCompleted) { if (!compressionCompleted) {
BMK_runOutcome_t const cOutcome = BMK_benchTimedFn( timeStateCompress, cbp); BMK_runOutcome_t const cOutcome =
BMK_benchTimedFn(timeStateCompress, cbp);
if (!BMK_isSuccessful_runOutcome(cOutcome)) { if (!BMK_isSuccessful_runOutcome(cOutcome)) {
RETURN_ERROR(30, BMK_benchOutcome_t, "compression error"); RETURN_ERROR(30, BMK_benchOutcome_t, "compression error");
} }
{ BMK_runTime_t const cResult = BMK_extract_runTime(cOutcome); {
BMK_runTime_t const cResult = BMK_extract_runTime(cOutcome);
cSize = cResult.sumOfReturn; cSize = cResult.sumOfReturn;
ratio = (double)srcSize / (double)cSize; ratio = (double)srcSize / (double)cSize;
{ BMK_benchResult_t newResult; {
newResult.cSpeed = (U64)((double)srcSize * TIMELOOP_NANOSEC / cResult.nanoSecPerRun); BMK_benchResult_t newResult;
newResult.cSpeed =
(U64)((double)srcSize * TIMELOOP_NANOSEC
/ cResult.nanoSecPerRun);
benchResult.cSize = cSize; benchResult.cSize = cSize;
if (newResult.cSpeed > benchResult.cSpeed) if (newResult.cSpeed > benchResult.cSpeed)
benchResult.cSpeed = newResult.cSpeed; benchResult.cSpeed = newResult.cSpeed;
} } }
}
{ int const ratioAccuracy = (ratio < 10.) ? 3 : 2; {
int const ratioAccuracy = (ratio < 10.) ? 3 : 2;
assert(cSize < UINT_MAX); assert(cSize < UINT_MAX);
OUTPUTLEVEL(2, "%2s-%-17.17s :%10u ->%10u (x%5.*f), %6.*f MB/s \r", OUTPUTLEVEL(
marks[markNb], displayName, 2,
(unsigned)srcSize, (unsigned)cSize, "%2s-%-17.17s :%10u ->%10u (x%5.*f), %6.*f MB/s \r",
ratioAccuracy, ratio, marks[markNb],
benchResult.cSpeed < (10 * MB_UNIT) ? 2 : 1, (double)benchResult.cSpeed / MB_UNIT); displayName,
(unsigned)srcSize,
(unsigned)cSize,
ratioAccuracy,
ratio,
benchResult.cSpeed < (10 * MB_UNIT) ? 2 : 1,
(double)benchResult.cSpeed / MB_UNIT);
} }
compressionCompleted = BMK_isCompleted_TimedFn(timeStateCompress); compressionCompleted =
BMK_isCompleted_TimedFn(timeStateCompress);
} }
if(!decompressionCompleted) { if (!decompressionCompleted) {
BMK_runOutcome_t const dOutcome = BMK_benchTimedFn(timeStateDecompress, dbp); BMK_runOutcome_t const dOutcome =
BMK_benchTimedFn(timeStateDecompress, dbp);
if(!BMK_isSuccessful_runOutcome(dOutcome)) { if (!BMK_isSuccessful_runOutcome(dOutcome)) {
RETURN_ERROR(30, BMK_benchOutcome_t, "decompression error"); RETURN_ERROR(30, BMK_benchOutcome_t, "decompression error");
} }
{ BMK_runTime_t const dResult = BMK_extract_runTime(dOutcome); {
U64 const newDSpeed = (U64)((double)srcSize * TIMELOOP_NANOSEC / dResult.nanoSecPerRun); BMK_runTime_t const dResult = BMK_extract_runTime(dOutcome);
U64 const newDSpeed =
(U64)((double)srcSize * TIMELOOP_NANOSEC
/ dResult.nanoSecPerRun);
if (newDSpeed > benchResult.dSpeed) if (newDSpeed > benchResult.dSpeed)
benchResult.dSpeed = newDSpeed; benchResult.dSpeed = newDSpeed;
} }
{ int const ratioAccuracy = (ratio < 10.) ? 3 : 2; {
OUTPUTLEVEL(2, "%2s-%-17.17s :%10u ->%10u (x%5.*f), %6.*f MB/s, %6.1f MB/s\r", int const ratioAccuracy = (ratio < 10.) ? 3 : 2;
marks[markNb], displayName, OUTPUTLEVEL(
(unsigned)srcSize, (unsigned)cSize, 2,
ratioAccuracy, ratio, "%2s-%-17.17s :%10u ->%10u (x%5.*f), %6.*f MB/s, %6.1f MB/s\r",
benchResult.cSpeed < (10 * MB_UNIT) ? 2 : 1, (double)benchResult.cSpeed / MB_UNIT, marks[markNb],
displayName,
(unsigned)srcSize,
(unsigned)cSize,
ratioAccuracy,
ratio,
benchResult.cSpeed < (10 * MB_UNIT) ? 2 : 1,
(double)benchResult.cSpeed / MB_UNIT,
(double)benchResult.dSpeed / MB_UNIT); (double)benchResult.dSpeed / MB_UNIT);
} }
decompressionCompleted = BMK_isCompleted_TimedFn(timeStateDecompress); decompressionCompleted =
BMK_isCompleted_TimedFn(timeStateDecompress);
} }
markNb = (markNb+1) % NB_MARKS; markNb = (markNb + 1) % NB_MARKS;
} /* while (!(compressionCompleted && decompressionCompleted)) */ } /* while (!(compressionCompleted && decompressionCompleted)) */
/* CRC Checking */ /* CRC Checking */
{ const BYTE* resultBuffer = (const BYTE*)(*resultBufferPtr); {
const BYTE* resultBuffer = (const BYTE*)(*resultBufferPtr);
U64 const crcCheck = XXH64(resultBuffer, srcSize, 0); U64 const crcCheck = XXH64(resultBuffer, srcSize, 0);
if ((adv->mode == BMK_both) && (crcOrig!=crcCheck)) { if ((adv->mode == BMK_both) && (crcOrig != crcCheck)) {
size_t u; size_t u;
DISPLAY("!!! WARNING !!! %14s : Invalid Checksum : %x != %x \n", DISPLAY("!!! WARNING !!! %14s : Invalid Checksum : %x != %x \n",
displayName, (unsigned)crcOrig, (unsigned)crcCheck); displayName,
for (u=0; u<srcSize; u++) { (unsigned)crcOrig,
(unsigned)crcCheck);
for (u = 0; u < srcSize; u++) {
if (((const BYTE*)srcBuffer)[u] != resultBuffer[u]) { if (((const BYTE*)srcBuffer)[u] != resultBuffer[u]) {
unsigned segNb, bNb, pos; unsigned segNb, bNb, pos;
size_t bacc = 0; size_t bacc = 0;
DISPLAY("Decoding error at pos %u ", (unsigned)u); DISPLAY("Decoding error at pos %u ", (unsigned)u);
for (segNb = 0; segNb < nbBlocks; segNb++) { for (segNb = 0; segNb < nbBlocks; segNb++) {
if (bacc + srcSizes[segNb] > u) break; if (bacc + srcSizes[segNb] > u)
break;
bacc += srcSizes[segNb]; bacc += srcSizes[segNb];
} }
pos = (U32)(u - bacc); pos = (U32)(u - bacc);
bNb = pos / (128 KB); bNb = pos / (128 KB);
DISPLAY("(sample %u, block %u, pos %u) \n", segNb, bNb, pos); DISPLAY("(sample %u, block %u, pos %u) \n",
{ size_t const lowest = (u>5) ? 5 : u; segNb,
bNb,
pos);
{
size_t const lowest = (u > 5) ? 5 : u;
size_t n; size_t n;
DISPLAY("origin: "); DISPLAY("origin: ");
for (n=lowest; n>0; n--) for (n = lowest; n > 0; n--)
DISPLAY("%02X ", ((const BYTE*)srcBuffer)[u-n]); DISPLAY("%02X ",
((const BYTE*)srcBuffer)[u - n]);
DISPLAY(" :%02X: ", ((const BYTE*)srcBuffer)[u]); DISPLAY(" :%02X: ", ((const BYTE*)srcBuffer)[u]);
for (n=1; n<3; n++) for (n = 1; n < 3; n++)
DISPLAY("%02X ", ((const BYTE*)srcBuffer)[u+n]); DISPLAY("%02X ",
((const BYTE*)srcBuffer)[u + n]);
DISPLAY(" \n"); DISPLAY(" \n");
DISPLAY("decode: "); DISPLAY("decode: ");
for (n=lowest; n>0; n--) for (n = lowest; n > 0; n--)
DISPLAY("%02X ", resultBuffer[u-n]); DISPLAY("%02X ", resultBuffer[u - n]);
DISPLAY(" :%02X: ", resultBuffer[u]); DISPLAY(" :%02X: ", resultBuffer[u]);
for (n=1; n<3; n++) for (n = 1; n < 3; n++)
DISPLAY("%02X ", resultBuffer[u+n]); DISPLAY("%02X ", resultBuffer[u + n]);
DISPLAY(" \n"); DISPLAY(" \n");
} }
break; break;
} }
if (u==srcSize-1) { /* should never happen */ if (u == srcSize - 1) { /* should never happen */
DISPLAY("no difference detected\n"); DISPLAY("no difference detected\n");
} }
} /* for (u=0; u<srcSize; u++) */ } /* for (u=0; u<srcSize; u++) */
} /* if ((adv->mode == BMK_both) && (crcOrig!=crcCheck)) */ } /* if ((adv->mode == BMK_both) && (crcOrig!=crcCheck)) */
} /* CRC Checking */ } /* CRC Checking */
if (displayLevel == 1) { /* hidden display mode -q, used by python speed benchmark */ if (displayLevel
== 1) { /* hidden display mode -q, used by python speed benchmark */
double const cSpeed = (double)benchResult.cSpeed / MB_UNIT; double const cSpeed = (double)benchResult.cSpeed / MB_UNIT;
double const dSpeed = (double)benchResult.dSpeed / MB_UNIT; double const dSpeed = (double)benchResult.dSpeed / MB_UNIT;
if (adv->additionalParam) { if (adv->additionalParam) {
OUTPUT("-%-3i%11i (%5.3f) %6.2f MB/s %6.1f MB/s %s (param=%d)\n", cLevel, (int)cSize, ratio, cSpeed, dSpeed, displayName, adv->additionalParam); OUTPUT("-%-3i%11i (%5.3f) %6.2f MB/s %6.1f MB/s %s (param=%d)\n",
cLevel,
(int)cSize,
ratio,
cSpeed,
dSpeed,
displayName,
adv->additionalParam);
} else { } else {
OUTPUT("-%-3i%11i (%5.3f) %6.2f MB/s %6.1f MB/s %s\n", cLevel, (int)cSize, ratio, cSpeed, dSpeed, displayName); OUTPUT("-%-3i%11i (%5.3f) %6.2f MB/s %6.1f MB/s %s\n",
cLevel,
(int)cSize,
ratio,
cSpeed,
dSpeed,
displayName);
} }
} }
OUTPUTLEVEL(2, "%2i#\n", cLevel); OUTPUTLEVEL(2, "%2i#\n", cLevel);
} /* Bench */ } /* Bench */
benchResult.cMem = (1ULL << (comprParams->windowLog)) + ZSTD_sizeof_CCtx(cctx); benchResult.cMem =
(1ULL << (comprParams->windowLog)) + ZSTD_sizeof_CCtx(cctx);
return BMK_benchOutcome_setValidResult(benchResult); return BMK_benchOutcome_setValidResult(benchResult);
} }
BMK_benchOutcome_t BMK_benchMemAdvanced(const void* srcBuffer, size_t srcSize, BMK_benchOutcome_t BMK_benchMemAdvanced(
void* dstBuffer, size_t dstCapacity, const void* srcBuffer,
const size_t* fileSizes, unsigned nbFiles, size_t srcSize,
int cLevel, const ZSTD_compressionParameters* comprParams, void* dstBuffer,
const void* dictBuffer, size_t dictBufferSize, size_t dstCapacity,
int displayLevel, const char* displayName, const BMK_advancedParams_t* adv) const size_t* fileSizes,
unsigned nbFiles,
int cLevel,
const ZSTD_compressionParameters* comprParams,
const void* dictBuffer,
size_t dictBufferSize,
int displayLevel,
const char* displayName,
const BMK_advancedParams_t* adv)
{ {
int const dstParamsError = !dstBuffer ^ !dstCapacity; /* must be both NULL or none */ int const dstParamsError =
!dstBuffer ^ !dstCapacity; /* must be both NULL or none */
size_t const blockSize = ((adv->blockSize>=32 && (adv->mode != BMK_decodeOnly)) ? adv->blockSize : srcSize) + (!srcSize) /* avoid div by 0 */ ; size_t const blockSize =
U32 const maxNbBlocks = (U32) ((srcSize + (blockSize-1)) / blockSize) + nbFiles; ((adv->blockSize >= 32 && (adv->mode != BMK_decodeOnly))
? adv->blockSize
: srcSize)
+ (!srcSize) /* avoid div by 0 */;
U32 const maxNbBlocks =
(U32)((srcSize + (blockSize - 1)) / blockSize) + nbFiles;
/* these are the blockTable parameters, just split up */ /* these are the blockTable parameters, just split up */
const void ** const srcPtrs = (const void**)malloc(maxNbBlocks * sizeof(void*)); const void** const srcPtrs =
(const void**)malloc(maxNbBlocks * sizeof(void*));
size_t* const srcSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t)); size_t* const srcSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t));
void** const cPtrs = (void**)malloc(maxNbBlocks * sizeof(void*));
void ** const cPtrs = (void**)malloc(maxNbBlocks * sizeof(void*));
size_t* const cSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t)); size_t* const cSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t));
size_t* const cCapacities = (size_t*)malloc(maxNbBlocks * sizeof(size_t)); size_t* const cCapacities = (size_t*)malloc(maxNbBlocks * sizeof(size_t));
void ** const resPtrs = (void**)malloc(maxNbBlocks * sizeof(void*)); void** const resPtrs = (void**)malloc(maxNbBlocks * sizeof(void*));
size_t* const resSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t)); size_t* const resSizes = (size_t*)malloc(maxNbBlocks * sizeof(size_t));
BMK_timedFnState_t* timeStateCompress = BMK_createTimedFnState(adv->nbSeconds * 1000, BMK_RUNTEST_DEFAULT_MS); BMK_timedFnState_t* timeStateCompress = BMK_createTimedFnState(
BMK_timedFnState_t* timeStateDecompress = BMK_createTimedFnState(adv->nbSeconds * 1000, BMK_RUNTEST_DEFAULT_MS); adv->nbSeconds * 1000, BMK_RUNTEST_DEFAULT_MS);
BMK_timedFnState_t* timeStateDecompress = BMK_createTimedFnState(
adv->nbSeconds * 1000, BMK_RUNTEST_DEFAULT_MS);
ZSTD_CCtx* const cctx = ZSTD_createCCtx(); ZSTD_CCtx* const cctx = ZSTD_createCCtx();
ZSTD_DCtx* const dctx = ZSTD_createDCtx(); ZSTD_DCtx* const dctx = ZSTD_createDCtx();
const size_t maxCompressedSize = dstCapacity ? dstCapacity : ZSTD_compressBound(srcSize) + (maxNbBlocks * 1024); const size_t maxCompressedSize = dstCapacity
? dstCapacity
: ZSTD_compressBound(srcSize) + (maxNbBlocks * 1024);
void* const internalDstBuffer = dstBuffer ? NULL : malloc(maxCompressedSize); void* const internalDstBuffer =
dstBuffer ? NULL : malloc(maxCompressedSize);
void* const compressedBuffer = dstBuffer ? dstBuffer : internalDstBuffer; void* const compressedBuffer = dstBuffer ? dstBuffer : internalDstBuffer;
BMK_benchOutcome_t outcome = BMK_benchOutcome_error(); /* error by default */ BMK_benchOutcome_t outcome =
BMK_benchOutcome_error(); /* error by default */
void* resultBuffer = srcSize ? malloc(srcSize) : NULL; void* resultBuffer = srcSize ? malloc(srcSize) : NULL;
int const allocationincomplete = !srcPtrs || !srcSizes || !cPtrs || int const allocationincomplete = !srcPtrs || !srcSizes || !cPtrs || !cSizes
!cSizes || !cCapacities || !resPtrs || !resSizes || || !cCapacities || !resPtrs || !resSizes || !timeStateCompress
!timeStateCompress || !timeStateDecompress || || !timeStateDecompress || !cctx || !dctx || !compressedBuffer
!cctx || !dctx || || !resultBuffer;
!compressedBuffer || !resultBuffer;
if (!allocationincomplete && !dstParamsError) { if (!allocationincomplete && !dstParamsError) {
outcome = BMK_benchMemAdvancedNoAlloc(srcPtrs, srcSizes, outcome = BMK_benchMemAdvancedNoAlloc(
cPtrs, cCapacities, cSizes, srcPtrs,
resPtrs, resSizes, srcSizes,
cPtrs,
cCapacities,
cSizes,
resPtrs,
resSizes,
&resultBuffer, &resultBuffer,
compressedBuffer, maxCompressedSize, compressedBuffer,
timeStateCompress, timeStateDecompress, maxCompressedSize,
srcBuffer, srcSize, timeStateCompress,
fileSizes, nbFiles, timeStateDecompress,
cLevel, comprParams, srcBuffer,
dictBuffer, dictBufferSize, srcSize,
cctx, dctx, fileSizes,
displayLevel, displayName, adv); nbFiles,
cLevel,
comprParams,
dictBuffer,
dictBufferSize,
cctx,
dctx,
displayLevel,
displayName,
adv);
} }
/* clean up */ /* clean up */
@@ -644,41 +876,64 @@ BMK_benchOutcome_t BMK_benchMemAdvanced(const void* srcBuffer, size_t srcSize,
free(resPtrs); free(resPtrs);
free(resSizes); free(resSizes);
if(allocationincomplete) { if (allocationincomplete) {
RETURN_ERROR(31, BMK_benchOutcome_t, "allocation error : not enough memory"); RETURN_ERROR(
31, BMK_benchOutcome_t, "allocation error : not enough memory");
} }
if(dstParamsError) { if (dstParamsError) {
RETURN_ERROR(32, BMK_benchOutcome_t, "Dst parameters not coherent"); RETURN_ERROR(32, BMK_benchOutcome_t, "Dst parameters not coherent");
} }
return outcome; return outcome;
} }
BMK_benchOutcome_t BMK_benchMem(const void* srcBuffer, size_t srcSize, BMK_benchOutcome_t BMK_benchMem(
const size_t* fileSizes, unsigned nbFiles, const void* srcBuffer,
int cLevel, const ZSTD_compressionParameters* comprParams, size_t srcSize,
const void* dictBuffer, size_t dictBufferSize, const size_t* fileSizes,
int displayLevel, const char* displayName) { unsigned nbFiles,
int cLevel,
const ZSTD_compressionParameters* comprParams,
const void* dictBuffer,
size_t dictBufferSize,
int displayLevel,
const char* displayName)
{
BMK_advancedParams_t const adv = BMK_initAdvancedParams(); BMK_advancedParams_t const adv = BMK_initAdvancedParams();
return BMK_benchMemAdvanced(srcBuffer, srcSize, return BMK_benchMemAdvanced(
NULL, 0, srcBuffer,
fileSizes, nbFiles, srcSize,
cLevel, comprParams, NULL,
dictBuffer, dictBufferSize, 0,
displayLevel, displayName, &adv); fileSizes,
nbFiles,
cLevel,
comprParams,
dictBuffer,
dictBufferSize,
displayLevel,
displayName,
&adv);
} }
static BMK_benchOutcome_t BMK_benchCLevel(const void* srcBuffer, size_t benchedSize, static BMK_benchOutcome_t BMK_benchCLevel(
const size_t* fileSizes, unsigned nbFiles, const void* srcBuffer,
int cLevel, const ZSTD_compressionParameters* comprParams, size_t benchedSize,
const void* dictBuffer, size_t dictBufferSize, const size_t* fileSizes,
int displayLevel, const char* displayName, unsigned nbFiles,
BMK_advancedParams_t const * const adv) int cLevel,
const ZSTD_compressionParameters* comprParams,
const void* dictBuffer,
size_t dictBufferSize,
int displayLevel,
const char* displayName,
BMK_advancedParams_t const* const adv)
{ {
const char* pch = strrchr(displayName, '\\'); /* Windows */ const char* pch = strrchr(displayName, '\\'); /* Windows */
if (!pch) pch = strrchr(displayName, '/'); /* Linux */ if (!pch)
if (pch) displayName = pch+1; pch = strrchr(displayName, '/'); /* Linux */
if (pch)
displayName = pch + 1;
if (adv->realTime) { if (adv->realTime) {
DISPLAYLEVEL(2, "Note : switching to real-time priority \n"); DISPLAYLEVEL(2, "Note : switching to real-time priority \n");
@@ -687,24 +942,38 @@ static BMK_benchOutcome_t BMK_benchCLevel(const void* srcBuffer, size_t benchedS
if (displayLevel == 1 && !adv->additionalParam) /* --quiet mode */ if (displayLevel == 1 && !adv->additionalParam) /* --quiet mode */
OUTPUT("bench %s %s: input %u bytes, %u seconds, %u KB blocks\n", OUTPUT("bench %s %s: input %u bytes, %u seconds, %u KB blocks\n",
ZSTD_VERSION_STRING, ZSTD_GIT_COMMIT_STRING, ZSTD_VERSION_STRING,
(unsigned)benchedSize, adv->nbSeconds, (unsigned)(adv->blockSize>>10)); ZSTD_GIT_COMMIT_STRING,
(unsigned)benchedSize,
adv->nbSeconds,
(unsigned)(adv->blockSize >> 10));
return BMK_benchMemAdvanced(srcBuffer, benchedSize, return BMK_benchMemAdvanced(
NULL, 0, srcBuffer,
fileSizes, nbFiles, benchedSize,
cLevel, comprParams, NULL,
dictBuffer, dictBufferSize, 0,
displayLevel, displayName, adv); fileSizes,
nbFiles,
cLevel,
comprParams,
dictBuffer,
dictBufferSize,
displayLevel,
displayName,
adv);
} }
int BMK_syntheticTest(int cLevel, double compressibility, int BMK_syntheticTest(
int cLevel,
double compressibility,
const ZSTD_compressionParameters* compressionParams, const ZSTD_compressionParameters* compressionParams,
int displayLevel, const BMK_advancedParams_t* adv) int displayLevel,
const BMK_advancedParams_t* adv)
{ {
char nameBuff[20] = {0}; char nameBuff[20] = { 0 };
const char* name = nameBuff; const char* name = nameBuff;
size_t const benchedSize = 10000000; size_t const benchedSize = adv->blockSize ? adv->blockSize : 10000000;
void* srcBuffer; void* srcBuffer;
BMK_benchOutcome_t res; BMK_benchOutcome_t res;
@@ -726,15 +995,26 @@ int BMK_syntheticTest(int cLevel, double compressibility,
name = "Lorem ipsum"; name = "Lorem ipsum";
} else { } else {
RDG_genBuffer(srcBuffer, benchedSize, compressibility, 0.0, 0); RDG_genBuffer(srcBuffer, benchedSize, compressibility, 0.0, 0);
snprintf (nameBuff, sizeof(nameBuff), "Synthetic %2u%%", (unsigned)(compressibility*100)); formatString_u(
nameBuff,
sizeof(nameBuff),
"Synthetic %u%%",
(unsigned)(compressibility * 100));
} }
/* Bench */ /* Bench */
res = BMK_benchCLevel(srcBuffer, benchedSize, res = BMK_benchCLevel(
&benchedSize /* ? */, 1 /* ? */, srcBuffer,
cLevel, compressionParams, benchedSize,
NULL, 0, /* dictionary */ &benchedSize /* ? */,
displayLevel, name, adv); 1 /* ? */,
cLevel,
compressionParams,
NULL,
0, /* dictionary */
displayLevel,
name,
adv);
/* clean up */ /* clean up */
free(srcBuffer); free(srcBuffer);
@@ -742,8 +1022,6 @@ int BMK_syntheticTest(int cLevel, double compressibility,
return !BMK_isSuccessful_benchOutcome(res); return !BMK_isSuccessful_benchOutcome(res);
} }
static size_t BMK_findMaxMem(U64 requiredMem) static size_t BMK_findMaxMem(U64 requiredMem)
{ {
size_t const step = 64 MB; size_t const step = 64 MB;
@@ -751,7 +1029,8 @@ static size_t BMK_findMaxMem(U64 requiredMem)
requiredMem = (((requiredMem >> 26) + 1) << 26); requiredMem = (((requiredMem >> 26) + 1) << 26);
requiredMem += step; requiredMem += step;
if (requiredMem > maxMemory) requiredMem = maxMemory; if (requiredMem > maxMemory)
requiredMem = maxMemory;
do { do {
testmem = (BYTE*)malloc((size_t)requiredMem); testmem = (BYTE*)malloc((size_t)requiredMem);
@@ -765,47 +1044,69 @@ static size_t BMK_findMaxMem(U64 requiredMem)
/*! BMK_loadFiles() : /*! BMK_loadFiles() :
* Loads `buffer` with content of files listed within `fileNamesTable`. * Loads `buffer` with content of files listed within `fileNamesTable`.
* At most, fills `buffer` entirely. */ * At most, fills `buffer` entirely. */
static int BMK_loadFiles(void* buffer, size_t bufferSize, static int BMK_loadFiles(
void* buffer,
size_t bufferSize,
size_t* fileSizes, size_t* fileSizes,
const char* const * fileNamesTable, unsigned nbFiles, const char* const* fileNamesTable,
unsigned nbFiles,
int displayLevel) int displayLevel)
{ {
size_t pos = 0, totalSize = 0; size_t pos = 0, totalSize = 0;
unsigned n; unsigned n;
for (n=0; n<nbFiles; n++) { for (n = 0; n < nbFiles; n++) {
U64 fileSize = UTIL_getFileSize(fileNamesTable[n]); /* last file may be shortened */ U64 fileSize = UTIL_getFileSize(
fileNamesTable[n]); /* last file may be shortened */
if (UTIL_isDirectory(fileNamesTable[n])) { if (UTIL_isDirectory(fileNamesTable[n])) {
DISPLAYLEVEL(2, "Ignoring %s directory... \n", fileNamesTable[n]); DISPLAYLEVEL(
2, "Ignoring %s directory... \n", fileNamesTable[n]);
fileSizes[n] = 0; fileSizes[n] = 0;
continue; continue;
} }
if (fileSize == UTIL_FILESIZE_UNKNOWN) { if (fileSize == UTIL_FILESIZE_UNKNOWN) {
DISPLAYLEVEL(2, "Cannot evaluate size of %s, ignoring ... \n", fileNamesTable[n]); DISPLAYLEVEL(
2,
"Cannot evaluate size of %s, ignoring ... \n",
fileNamesTable[n]);
fileSizes[n] = 0; fileSizes[n] = 0;
continue; continue;
} }
{ FILE* const f = fopen(fileNamesTable[n], "rb"); {
if (f==NULL) RETURN_ERROR_INT(10, "impossible to open file %s", fileNamesTable[n]); FILE* const f = fopen(fileNamesTable[n], "rb");
if (f == NULL)
RETURN_ERROR_INT(
10, "impossible to open file %s", fileNamesTable[n]);
OUTPUTLEVEL(2, "Loading %s... \r", fileNamesTable[n]); OUTPUTLEVEL(2, "Loading %s... \r", fileNamesTable[n]);
if (fileSize > bufferSize-pos) fileSize = bufferSize-pos, nbFiles=n; /* buffer too small - stop after this file */ if (fileSize > bufferSize - pos)
{ size_t const readSize = fread(((char*)buffer)+pos, 1, (size_t)fileSize, f); fileSize = bufferSize - pos,
if (readSize != (size_t)fileSize) RETURN_ERROR_INT(11, "could not read %s", fileNamesTable[n]); nbFiles = n; /* buffer too small - stop after this file */
{
size_t const readSize =
fread(((char*)buffer) + pos, 1, (size_t)fileSize, f);
if (readSize != (size_t)fileSize)
RETURN_ERROR_INT(
11, "could not read %s", fileNamesTable[n]);
pos += readSize; pos += readSize;
} }
fileSizes[n] = (size_t)fileSize; fileSizes[n] = (size_t)fileSize;
totalSize += (size_t)fileSize; totalSize += (size_t)fileSize;
fclose(f); fclose(f);
} } }
}
if (totalSize == 0) RETURN_ERROR_INT(12, "no data to bench"); if (totalSize == 0)
RETURN_ERROR_INT(12, "no data to bench");
return 0; return 0;
} }
int BMK_benchFilesAdvanced( int BMK_benchFilesAdvanced(
const char* const * fileNamesTable, unsigned nbFiles, const char* const* fileNamesTable,
const char* dictFileName, int cLevel, unsigned nbFiles,
const char* dictFileName,
int cLevel,
const ZSTD_compressionParameters* compressionParams, const ZSTD_compressionParameters* compressionParams,
int displayLevel, const BMK_advancedParams_t* adv) int displayLevel,
const BMK_advancedParams_t* adv)
{ {
void* srcBuffer = NULL; void* srcBuffer = NULL;
size_t benchedSize; size_t benchedSize;
@@ -840,7 +1141,11 @@ int BMK_benchFilesAdvanced(
if (dictFileName != NULL) { if (dictFileName != NULL) {
U64 const dictFileSize = UTIL_getFileSize(dictFileName); U64 const dictFileSize = UTIL_getFileSize(dictFileName);
if (dictFileSize == UTIL_FILESIZE_UNKNOWN) { if (dictFileSize == UTIL_FILESIZE_UNKNOWN) {
DISPLAYLEVEL(1, "error loading %s : %s \n", dictFileName, strerror(errno)); DISPLAYLEVEL(
1,
"error loading %s : %s \n",
dictFileName,
strerror(errno));
free(fileSizes); free(fileSizes);
DISPLAYLEVEL(1, "benchmark aborted"); DISPLAYLEVEL(1, "benchmark aborted");
return 17; return 17;
@@ -852,27 +1157,37 @@ int BMK_benchFilesAdvanced(
} }
dictBufferSize = (size_t)dictFileSize; dictBufferSize = (size_t)dictFileSize;
dictBuffer = malloc(dictBufferSize); dictBuffer = malloc(dictBufferSize);
if (dictBuffer==NULL) { if (dictBuffer == NULL) {
free(fileSizes); free(fileSizes);
DISPLAYLEVEL(1, "not enough memory for dictionary (%u bytes)", DISPLAYLEVEL(
1,
"not enough memory for dictionary (%u bytes)",
(unsigned)dictBufferSize); (unsigned)dictBufferSize);
return 19; return 19;
} }
{ int const errorCode = BMK_loadFiles(dictBuffer, dictBufferSize, {
fileSizes, &dictFileName /*?*/, int const errorCode = BMK_loadFiles(
1 /*?*/, displayLevel); dictBuffer,
dictBufferSize,
fileSizes,
&dictFileName /*?*/,
1 /*?*/,
displayLevel);
if (errorCode) { if (errorCode) {
res = BMK_benchOutcome_error(); res = BMK_benchOutcome_error();
goto _cleanUp; goto _cleanUp;
} } }
}
} }
/* Memory allocation & restrictions */ /* Memory allocation & restrictions */
benchedSize = BMK_findMaxMem(totalSizeToLoad * 3) / 3; benchedSize = BMK_findMaxMem(totalSizeToLoad * 3) / 3;
if ((U64)benchedSize > totalSizeToLoad) benchedSize = (size_t)totalSizeToLoad; if ((U64)benchedSize > totalSizeToLoad)
benchedSize = (size_t)totalSizeToLoad;
if (benchedSize < totalSizeToLoad) if (benchedSize < totalSizeToLoad)
DISPLAY("Not enough memory; testing %u MB only...\n", (unsigned)(benchedSize >> 20)); DISPLAY("Not enough memory; testing %u MB only...\n",
(unsigned)(benchedSize >> 20));
srcBuffer = benchedSize ? malloc(benchedSize) : NULL; srcBuffer = benchedSize ? malloc(benchedSize) : NULL;
if (!srcBuffer) { if (!srcBuffer) {
@@ -883,25 +1198,41 @@ int BMK_benchFilesAdvanced(
} }
/* Load input buffer */ /* Load input buffer */
{ int const errorCode = BMK_loadFiles(srcBuffer, benchedSize, {
fileSizes, fileNamesTable, nbFiles, int const errorCode = BMK_loadFiles(
srcBuffer,
benchedSize,
fileSizes,
fileNamesTable,
nbFiles,
displayLevel); displayLevel);
if (errorCode) { if (errorCode) {
res = BMK_benchOutcome_error(); res = BMK_benchOutcome_error();
goto _cleanUp; goto _cleanUp;
} } }
}
/* Bench */ /* Bench */
{ char mfName[20] = {0}; {
snprintf (mfName, sizeof(mfName), " %u files", nbFiles); char mfName[20] = { 0 };
{ const char* const displayName = (nbFiles > 1) ? mfName : fileNamesTable[0]; formatString_u(mfName, sizeof(mfName), " %u files", nbFiles);
res = BMK_benchCLevel(srcBuffer, benchedSize, {
fileSizes, nbFiles, const char* const displayName =
cLevel, compressionParams, (nbFiles > 1) ? mfName : fileNamesTable[0];
dictBuffer, dictBufferSize, res = BMK_benchCLevel(
displayLevel, displayName, srcBuffer,
benchedSize,
fileSizes,
nbFiles,
cLevel,
compressionParams,
dictBuffer,
dictBufferSize,
displayLevel,
displayName,
adv); adv);
} } }
}
_cleanUp: _cleanUp:
free(srcBuffer); free(srcBuffer);
@@ -910,12 +1241,21 @@ _cleanUp:
return !BMK_isSuccessful_benchOutcome(res); return !BMK_isSuccessful_benchOutcome(res);
} }
int BMK_benchFiles(
int BMK_benchFiles(const char* const * fileNamesTable, unsigned nbFiles, const char* const* fileNamesTable,
unsigned nbFiles,
const char* dictFileName, const char* dictFileName,
int cLevel, const ZSTD_compressionParameters* compressionParams, int cLevel,
const ZSTD_compressionParameters* compressionParams,
int displayLevel) int displayLevel)
{ {
BMK_advancedParams_t const adv = BMK_initAdvancedParams(); BMK_advancedParams_t const adv = BMK_initAdvancedParams();
return BMK_benchFilesAdvanced(fileNamesTable, nbFiles, dictFileName, cLevel, compressionParams, displayLevel, &adv); return BMK_benchFilesAdvanced(
fileNamesTable,
nbFiles,
dictFileName,
cLevel,
compressionParams,
displayLevel,
&adv);
} }
programs/lorem.c
+127 -49
View File
@@ -8,7 +8,6 @@
* You may select, at your option, one of the above-listed licenses. * You may select, at your option, one of the above-listed licenses.
*/ */
/* Implementation notes: /* Implementation notes:
* *
* This is a very simple lorem ipsum generator * This is a very simple lorem ipsum generator
@@ -24,22 +23,22 @@
* and lacks a regularity more representative of text. * and lacks a regularity more representative of text.
* *
* The compression ratio achievable on the generated lorem ipsum * The compression ratio achievable on the generated lorem ipsum
* is still a bit too good, presumably because the dictionary is too small. * is still a bit too good, presumably because the dictionary is a bit too
* It would be possible to create some more complex scheme, * small. It would be possible to create some more complex scheme, notably by
* notably by enlarging the dictionary with a word generator, * enlarging the dictionary with a word generator, and adding grammatical rules
* and adding grammatical rules (composition) and syntax rules. * (composition) and syntax rules. But that's probably overkill for the intended
* But that's probably overkill for the intended goal. * goal.
*/ */
#include "lorem.h" #include "lorem.h"
#include <string.h> /* memcpy */
#include <limits.h> /* INT_MAX */
#include <assert.h> #include <assert.h>
#include <limits.h> /* INT_MAX */
#include <string.h> /* memcpy */
#define WORD_MAX_SIZE 20 #define WORD_MAX_SIZE 20
/* Define the word pool */ /* Define the word pool */
static const char *words[] = { static const char* kWords[] = {
"lorem", "ipsum", "dolor", "sit", "amet", "lorem", "ipsum", "dolor", "sit", "amet",
"consectetur", "adipiscing", "elit", "sed", "do", "consectetur", "adipiscing", "elit", "sed", "do",
"eiusmod", "tempor", "incididunt", "ut", "labore", "eiusmod", "tempor", "incididunt", "ut", "labore",
@@ -54,35 +53,105 @@ static const char *words[] = {
"dui", "faucibus", "ornare", "at", "elementum", "dui", "faucibus", "ornare", "at", "elementum",
"eu", "facilisis", "odio", "morbi", "quis", "eu", "facilisis", "odio", "morbi", "quis",
"eros", "donec", "ac", "orci", "purus", "eros", "donec", "ac", "orci", "purus",
"turpis", "cursus", "leo", "vel", "porta"}; "turpis", "cursus", "leo", "vel", "porta",
"consequat", "interdum", "varius", "vulputate", "aliquet",
"pharetra", "nunc", "auctor", "urna", "id",
"metus", "viverra", "nibh", "cras", "mi",
"unde", "omnis", "iste", "natus", "error",
"perspiciatis", "voluptatem", "accusantium", "doloremque", "laudantium",
"totam", "rem", "aperiam", "eaque", "ipsa",
"quae", "ab", "illo", "inventore", "veritatis",
"quasi", "architecto", "beatae", "vitae", "dicta",
"sunt", "explicabo", "nemo", "ipsam", "quia",
"voluptas", "aspernatur", "aut", "odit", "fugit",
"consequuntur", "magni", "dolores", "eos", "qui",
"ratione", "sequi", "nesciunt", "neque", "porro",
"quisquam", "est", "dolorem", "adipisci", "numquam",
"eius", "modi", "tempora", "incidunt", "magnam",
"quaerat", "ad", "minima", "veniam", "nostrum",
"ullam", "corporis", "suscipit", "laboriosam", "nisi",
"aliquid", "ex", "ea", "commodi", "consequatur",
"autem", "eum", "iure", "voluptate", "esse",
"quam", "nihil", "molestiae", "illum", "fugiat",
"quo", "pariatur", "vero", "accusamus", "iusto",
"dignissimos", "ducimus", "blanditiis", "praesentium", "voluptatum",
"deleniti", "atque", "corrupti", "quos", "quas",
"molestias", "excepturi", "sint", "occaecati", "cupiditate",
"provident", "similique", "culpa", "officia", "deserunt",
"mollitia", "animi", "laborum", "dolorum", "fuga",
"harum", "quidem", "rerum", "facilis", "expedita",
"distinctio", "libero", "tempore", "cum", "soluta",
"nobis", "eligendi", "optio", "cumque", "impedit",
"minus", "quod", "maxime", "placeat", "facere",
"possimus", "assumenda", "repellendus", "temporibus", "quibusdam",
"officiis", "debitis", "saepe", "eveniet", "voluptates",
"repudiandae", "recusandae", "itaque", "earum", "hic",
"tenetur", "sapiente", "delectus", "reiciendis", "cillum",
"maiores", "alias", "perferendis", "doloribus", "asperiores",
"repellat", "minim", "nostrud", "exercitation", "ullamco",
"laboris", "aliquip", "duis", "aute", "irure",
};
static const unsigned kNbWords = sizeof(kWords) / sizeof(kWords[0]);
/* simple distribution that favors small words : /* simple 1-dimension distribution, based on word's length, favors small words
* 1 letter : weight 3 */
* 2-3 letters : weight 2 static const int kWeights[] = { 0, 8, 6, 4, 3, 2 };
* 4+ letters : weight 1 static const size_t kNbWeights = sizeof(kWeights) / sizeof(kWeights[0]);
* This is expected to be a bit more difficult to compress */
static const int distrib[] = { #define DISTRIB_SIZE_MAX 650
0, 1, 2, 3, 3, 4, 5, 6, 7, 8, static int g_distrib[DISTRIB_SIZE_MAX] = { 0 };
8,9, 9, 10, 11, 12, 13, 13, 14, 15, static unsigned g_distribCount = 0;
15, 16, 17, 18, 19, 19, 20, 21, 22, 23,
24, 25, 26, 26, 26, 27, 28, 29, 30, 31, static void countFreqs(
32, 33, 34, 34, 35, 36, 37, 38, 39, 40, const char* words[],
41, 41, 42, 43, 43, 44, 45, 45, 46, 47, size_t nbWords,
48, 49, 50, 51, 52, 53, 54, 55, 55, 56, const int* weights,
57, 58, 58, 59, 60, 60, 61, 62, 63, 64, size_t nbWeights)
65, 66, 67, 67, 68, 69, 70, 71, 72, 72, {
73, 73, 74 }; unsigned total = 0;
static const unsigned distribCount = sizeof(distrib) / sizeof(distrib[0]); size_t w;
for (w = 0; w < nbWords; w++) {
size_t len = strlen(words[w]);
int lmax;
if (len >= nbWeights)
len = nbWeights - 1;
lmax = weights[len];
total += (unsigned)lmax;
}
g_distribCount = total;
assert(g_distribCount <= DISTRIB_SIZE_MAX);
}
static void init_word_distrib(
const char* words[],
size_t nbWords,
const int* weights,
size_t nbWeights)
{
size_t w, d = 0;
countFreqs(words, nbWords, weights, nbWeights);
for (w = 0; w < nbWords; w++) {
size_t len = strlen(words[w]);
int l, lmax;
if (len >= nbWeights)
len = nbWeights - 1;
lmax = weights[len];
for (l = 0; l < lmax; l++) {
g_distrib[d++] = (int)w;
}
}
}
/* Note: this unit only works when invoked sequentially. /* Note: this unit only works when invoked sequentially.
* No concurrent access is allowed */ * No concurrent access is allowed */
static char *g_ptr = NULL; static char* g_ptr = NULL;
static size_t g_nbChars = 0; static size_t g_nbChars = 0;
static size_t g_maxChars = 10000000; static size_t g_maxChars = 10000000;
static unsigned g_randRoot = 0; static unsigned g_randRoot = 0;
#define RDG_rotl32(x, r) ((x << r) | (x >> (32 - r))) #define RDG_rotl32(x, r) ((x << r) | (x >> (32 - r)))
static unsigned LOREM_rand(unsigned range) { static unsigned LOREM_rand(unsigned range)
{
static const unsigned prime1 = 2654435761U; static const unsigned prime1 = 2654435761U;
static const unsigned prime2 = 2246822519U; static const unsigned prime2 = 2246822519U;
unsigned rand32 = g_randRoot; unsigned rand32 = g_randRoot;
@@ -93,7 +162,8 @@ static unsigned LOREM_rand(unsigned range) {
return (unsigned)(((unsigned long long)rand32 * range) >> 32); return (unsigned)(((unsigned long long)rand32 * range) >> 32);
} }
static void writeLastCharacters(void) { static void writeLastCharacters(void)
{
size_t lastChars = g_maxChars - g_nbChars; size_t lastChars = g_maxChars - g_nbChars;
assert(g_maxChars >= g_nbChars); assert(g_maxChars >= g_nbChars);
if (lastChars == 0) if (lastChars == 0)
@@ -103,12 +173,12 @@ static void writeLastCharacters(void) {
memset(g_ptr + g_nbChars, ' ', lastChars - 2); memset(g_ptr + g_nbChars, ' ', lastChars - 2);
} }
if (lastChars > 1) { if (lastChars > 1) {
g_ptr[g_maxChars-1] = '\n'; g_ptr[g_maxChars - 1] = '\n';
} }
g_nbChars = g_maxChars; g_nbChars = g_maxChars;
} }
static void generateWord(const char *word, const char *separator, int upCase) static void generateWord(const char* word, const char* separator, int upCase)
{ {
size_t const len = strlen(word) + strlen(separator); size_t const len = strlen(word) + strlen(separator);
if (g_nbChars + len > g_maxChars) { if (g_nbChars + len > g_maxChars) {
@@ -125,33 +195,38 @@ static void generateWord(const char *word, const char *separator, int upCase)
g_nbChars += strlen(separator); g_nbChars += strlen(separator);
} }
static int about(unsigned target) { static int about(unsigned target)
{
return (int)(LOREM_rand(target) + LOREM_rand(target) + 1); return (int)(LOREM_rand(target) + LOREM_rand(target) + 1);
} }
/* Function to generate a random sentence */ /* Function to generate a random sentence */
static void generateSentence(int nbWords) { static void generateSentence(int nbWords)
{
int commaPos = about(9); int commaPos = about(9);
int comma2 = commaPos + about(7); int comma2 = commaPos + about(7);
int qmark = (LOREM_rand(11) == 7);
const char* endSep = qmark ? "? " : ". ";
int i; int i;
for (i = 0; i < nbWords; i++) { for (i = 0; i < nbWords; i++) {
int const wordID = distrib[LOREM_rand(distribCount)]; int const wordID = g_distrib[LOREM_rand(g_distribCount)];
const char *const word = words[wordID]; const char* const word = kWords[wordID];
const char* sep = " "; const char* sep = " ";
if (i == commaPos) if (i == commaPos)
sep = ", "; sep = ", ";
if (i == comma2) if (i == comma2)
sep = ", "; sep = ", ";
if (i == nbWords - 1) if (i == nbWords - 1)
sep = ". "; sep = endSep;
generateWord(word, sep, i==0); generateWord(word, sep, i == 0);
} }
} }
static void generateParagraph(int nbSentences) { static void generateParagraph(int nbSentences)
{
int i; int i;
for (i = 0; i < nbSentences; i++) { for (i = 0; i < nbSentences; i++) {
int wordsPerSentence = about(8); int wordsPerSentence = about(11);
generateSentence(wordsPerSentence); generateSentence(wordsPerSentence);
} }
if (g_nbChars < g_maxChars) { if (g_nbChars < g_maxChars) {
@@ -164,29 +239,33 @@ static void generateParagraph(int nbSentences) {
/* It's "common" for lorem ipsum generators to start with the same first /* It's "common" for lorem ipsum generators to start with the same first
* pre-defined sentence */ * pre-defined sentence */
static void generateFirstSentence(void) { static void generateFirstSentence(void)
{
int i; int i;
for (i = 0; i < 18; i++) { for (i = 0; i < 18; i++) {
const char *word = words[i]; const char* word = kWords[i];
const char *separator = " "; const char* separator = " ";
if (i == 4) if (i == 4)
separator = ", "; separator = ", ";
if (i == 7) if (i == 7)
separator = ", "; separator = ", ";
generateWord(word, separator, i==0); generateWord(word, separator, i == 0);
} }
generateWord(words[18], ". ", 0); generateWord(kWords[18], ". ", 0);
} }
size_t LOREM_genBlock(void* buffer, size_t size, size_t
unsigned seed, LOREM_genBlock(void* buffer, size_t size, unsigned seed, int first, int fill)
int first, int fill)
{ {
g_ptr = (char*)buffer; g_ptr = (char*)buffer;
assert(size < INT_MAX); assert(size < INT_MAX);
g_maxChars = size; g_maxChars = size;
g_nbChars = 0; g_nbChars = 0;
g_randRoot = seed; g_randRoot = seed;
if (g_distribCount == 0) {
init_word_distrib(kWords, kNbWords, kWeights, kNbWeights);
}
if (first) { if (first) {
generateFirstSentence(); generateFirstSentence();
} }
@@ -204,4 +283,3 @@ void LOREM_genBuffer(void* buffer, size_t size, unsigned seed)
{ {
LOREM_genBlock(buffer, size, seed, 1, 1); LOREM_genBlock(buffer, size, seed, 1, 1);
} }
tests/Makefile
+1 -1
View File
@@ -206,7 +206,7 @@ paramgrill : LDLIBS += -lm
paramgrill : $(ZSTD_FILES) $(PRGDIR)/util.c $(PRGDIR)/timefn.c $(PRGDIR)/benchfn.c $(PRGDIR)/benchzstd.c $(PRGDIR)/datagen.c $(PRGDIR)/lorem.c paramgrill.c paramgrill : $(ZSTD_FILES) $(PRGDIR)/util.c $(PRGDIR)/timefn.c $(PRGDIR)/benchfn.c $(PRGDIR)/benchzstd.c $(PRGDIR)/datagen.c $(PRGDIR)/lorem.c paramgrill.c
CLEAN += datagen CLEAN += datagen
datagen : $(PRGDIR)/datagen.c datagencli.c datagen : $(PRGDIR)/datagen.c $(PRGDIR)/lorem.c loremOut.c datagencli.c
$(LINK.c) $^ -o $@$(EXT) $(LINK.c) $^ -o $@$(EXT)
CLEAN += roundTripCrash CLEAN += roundTripCrash
tests/datagencli.c
+75 -56
View File
@@ -8,54 +8,52 @@
* You may select, at your option, one of the above-listed licenses. * You may select, at your option, one of the above-listed licenses.
*/ */
/*-************************************ /*-************************************
* Dependencies * Dependencies
**************************************/ **************************************/
#include "util.h" /* Compiler options */
#include <stdio.h> /* fprintf, stderr */ #include <stdio.h> /* fprintf, stderr */
#include "datagen.h" /* RDG_generate */ #include "datagen.h" /* RDG_generate */
#include "loremOut.h" /* LOREM_genOut */
#include "util.h" /* Compiler options */
/*-************************************ /*-************************************
* Constants * Constants
**************************************/ **************************************/
#define KB *(1 <<10) #define KB *(1 << 10)
#define MB *(1 <<20) #define MB *(1 << 20)
#define GB *(1U<<30) #define GB *(1U << 30)
#define SIZE_DEFAULT ((64 KB) + 1) #define SIZE_DEFAULT ((64 KB) + 1)
#define SEED_DEFAULT 0 #define SEED_DEFAULT 0
#define COMPRESSIBILITY_DEFAULT 50 #define COMPRESSIBILITY_DEFAULT 9999
/*-************************************ /*-************************************
* Macros * Macros
**************************************/ **************************************/
#define DISPLAY(...) fprintf(stderr, __VA_ARGS__) #define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
#define DISPLAYLEVEL(l, ...) if (displayLevel>=l) { DISPLAY(__VA_ARGS__); } #define DISPLAYLEVEL(l, ...) \
if (displayLevel >= l) { \
DISPLAY(__VA_ARGS__); \
}
static unsigned displayLevel = 2; static unsigned displayLevel = 2;
/*-******************************************************* /*-*******************************************************
* Command line * Command line
*********************************************************/ *********************************************************/
static int usage(const char* programName) static int usage(const char* programName)
{ {
DISPLAY( "Compressible data generator\n"); DISPLAY("Compressible data generator\n");
DISPLAY( "Usage :\n"); DISPLAY("Usage :\n");
DISPLAY( " %s [args]\n", programName); DISPLAY(" %s [args]\n", programName);
DISPLAY( "\n"); DISPLAY("\n");
DISPLAY( "Arguments :\n"); DISPLAY("Arguments :\n");
DISPLAY( " -g# : generate # data (default:%i)\n", SIZE_DEFAULT); DISPLAY(" -g# : generate # data (default:%i)\n", SIZE_DEFAULT);
DISPLAY( " -s# : Select seed (default:%i)\n", SEED_DEFAULT); DISPLAY(" -s# : Select seed (default:%i)\n", SEED_DEFAULT);
DISPLAY( " -P# : Select compressibility in %% (default:%i%%)\n", DISPLAY(" -P# : Select compressibility in %% (range [0-100])\n");
COMPRESSIBILITY_DEFAULT); DISPLAY(" -h : display help and exit\n");
DISPLAY( " -h : display help and exit\n");
return 0; return 0;
} }
int main(int argc, const char** argv) int main(int argc, const char** argv)
{ {
unsigned probaU32 = COMPRESSIBILITY_DEFAULT; unsigned probaU32 = COMPRESSIBILITY_DEFAULT;
@@ -65,48 +63,63 @@ int main(int argc, const char** argv)
const char* const programName = argv[0]; const char* const programName = argv[0];
int argNb; int argNb;
for(argNb=1; argNb<argc; argNb++) { for (argNb = 1; argNb < argc; argNb++) {
const char* argument = argv[argNb]; const char* argument = argv[argNb];
if(!argument) continue; /* Protection if argument empty */ if (!argument)
continue; /* Protection if argument empty */
/* Handle commands. Aggregated commands are allowed */ /* Handle commands. Aggregated commands are allowed */
if (*argument=='-') { if (*argument == '-') {
argument++; argument++;
while (*argument!=0) { while (*argument != 0) {
switch(*argument) switch (*argument) {
{
case 'h': case 'h':
return usage(programName); return usage(programName);
case 'g': case 'g':
argument++; argument++;
size=0; size = 0;
while ((*argument>='0') && (*argument<='9')) while ((*argument >= '0') && (*argument <= '9'))
size *= 10, size += *argument++ - '0'; size *= 10, size += (U64)(*argument++ - '0');
if (*argument=='K') { size <<= 10; argument++; } if (*argument == 'K') {
if (*argument=='M') { size <<= 20; argument++; } size <<= 10;
if (*argument=='G') { size <<= 30; argument++; } argument++;
if (*argument=='B') { argument++; } }
if (*argument == 'M') {
size <<= 20;
argument++;
}
if (*argument == 'G') {
size <<= 30;
argument++;
}
if (*argument == 'B') {
argument++;
}
break; break;
case 's': case 's':
argument++; argument++;
seed=0; seed = 0;
while ((*argument>='0') && (*argument<='9')) while ((*argument >= '0') && (*argument <= '9'))
seed *= 10, seed += *argument++ - '0'; seed *= 10, seed += (U32)(*argument++ - '0');
break; break;
case 'P': case 'P':
argument++; argument++;
probaU32 = 0; probaU32 = 0;
while ((*argument>='0') && (*argument<='9')) while ((*argument >= '0') && (*argument <= '9'))
probaU32 *= 10, probaU32 += *argument++ - '0'; probaU32 *= 10,
if (probaU32>100) probaU32 = 100; probaU32 += (U32)(*argument++ - '0');
if (probaU32 > 100)
probaU32 = 100;
break; break;
case 'L': /* hidden argument : Literal distribution probability */ case 'L': /* hidden argument : Literal distribution
probability */
argument++; argument++;
litProba=0.; litProba = 0.;
while ((*argument>='0') && (*argument<='9')) while ((*argument >= '0') && (*argument <= '9'))
litProba *= 10, litProba += *argument++ - '0'; litProba *= 10, litProba += *argument++ - '0';
if (litProba>100.) litProba=100.; if (litProba > 100.)
litProba = 100.;
litProba /= 100.; litProba /= 100.;
break; break;
case 'v': case 'v':
@@ -116,14 +129,20 @@ int main(int argc, const char** argv)
default: default:
return usage(programName); return usage(programName);
} }
} } } /* for(argNb=1; argNb<argc; argNb++) */ }
}
} /* for(argNb=1; argNb<argc; argNb++) */
DISPLAYLEVEL(4, "Compressible data Generator \n"); DISPLAYLEVEL(4, "Compressible data Generator \n");
if (probaU32!=COMPRESSIBILITY_DEFAULT)
DISPLAYLEVEL(3, "Compressibility : %i%%\n", probaU32);
DISPLAYLEVEL(3, "Seed = %u \n", (unsigned)seed); DISPLAYLEVEL(3, "Seed = %u \n", (unsigned)seed);
RDG_genStdout(size, (double)probaU32/100, litProba, seed); if (probaU32 != COMPRESSIBILITY_DEFAULT) {
DISPLAYLEVEL(3, "Compressibility : %i%%\n", probaU32);
RDG_genStdout(size, (double)probaU32 / 100, litProba, seed);
} else {
LOREM_genOut(size, seed);
}
DISPLAYLEVEL(3, "\n"); DISPLAYLEVEL(3, "\n");
return 0; return 0;
tests/loremOut.c
+50
View File
@@ -0,0 +1,50 @@
/*
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
/* Implementation notes:
* Generates a stream of Lorem ipsum paragraphs to stdout,
* up to the requested size, which can be very large (> 4 GB).
* Note that, beyond 1 paragraph, this generator produces
* a different content than LOREM_genBuffer (even when using same seed).
*/
#include "loremOut.h"
#include <assert.h>
#include <stdio.h>
#include "lorem.h" /* LOREM_genBlock */
#include "platform.h" /* Compiler options, SET_BINARY_MODE */
#define MIN(a, b) ((a) < (b) ? (a) : (b))
#define LOREM_BLOCKSIZE (1 << 10)
void LOREM_genOut(unsigned long long size, unsigned seed)
{
char buff[LOREM_BLOCKSIZE] = { 0 };
unsigned long long total = 0;
size_t genBlockSize = (size_t)MIN(size, LOREM_BLOCKSIZE);
/* init */
SET_BINARY_MODE(stdout);
/* Generate Ipsum text, one paragraph at a time */
while (total < size) {
size_t generated =
LOREM_genBlock(buff, genBlockSize, seed++, total == 0, 0);
assert(generated <= genBlockSize);
total += generated;
assert(total <= size);
fwrite(buff,
1,
generated,
stdout); /* note: should check potential write error */
if (size - total < genBlockSize)
genBlockSize = (size_t)(size - total);
}
assert(total == size);
}
tests/loremOut.h
+15
View File
@@ -0,0 +1,15 @@
/*
* Copyright (c) Meta Platforms, Inc. and affiliates.
* All rights reserved.
*
* This source code is licensed under both the BSD-style license (found in the
* LICENSE file in the root directory of this source tree) and the GPLv2 (found
* in the COPYING file in the root directory of this source tree).
* You may select, at your option, one of the above-listed licenses.
*/
/* LOREM_genOut():
* Generate @size bytes of compressible data using lorem ipsum generator into
* stdout.
*/
void LOREM_genOut(unsigned long long size, unsigned seed);