Commit Graph

354 Commits (b668badf00860bb2ef9f275e237640883feb9f64)

Author SHA1 Message Date
Yann Collet a51cab6e68 Merge pull request #678 from facebook/apiChange
Breaking API Change around CDict
2017-04-28 10:02:45 -07:00
Yann Collet 29297c6751 Changed default level 18 (large input)
Previous -18 : 4.7 MB/s, R:3.833
New -18 : 5.1 MB/s. R:3.825

It's a better fit within -17 (6.8 MB/s) and -19 (4.0 MB/s)
The new level 18 also uses significantly less memory.
And, it makes a good transition between level 17 (mml5)
and level 19 (mml3).
Up to now, there was no level with mml4.

(note : minmatch setting can have a large impact on some (specific) datasets)
2017-04-27 17:44:01 -07:00
Yann Collet 1c3ab0c77f fixed init error on Visual 2008 2017-04-27 12:57:11 -07:00
Yann Collet 77bf59ef50 added ZSTD_initCStream_usingCDict_advanced() 2017-04-27 11:43:04 -07:00
Yann Collet f4bd857d81 created ZSTD_compress_usingCDict_advanced() 2017-04-27 11:31:55 -07:00
Yann Collet 69a54d138a fixed compilation warning : declaration-after-statement 2017-04-27 01:11:26 -07:00
Yann Collet 31533bacce Changed ZSTD_createCDict_advanced()
It now only uses compressionParameters as argument.
It produces many changes throughout user code,
though hopefully they tend to be simple :
just provide the cParams part from existing ZSTD_parameters.

Some programs might depend on ZSTD_createCDict_advanced() to pass frame parameters.
This change will force them to revisit this strategy and fix it,
since frame parameters are effectively silently ignored in current version.
2017-04-27 00:29:04 -07:00
Yann Collet 768df129d2 changed ZSTD_compressBegin_usingCDict()
No longer takes `pledgedSrcSize` as argument
this is in line with similar functions ZSTD_compress_usingCDict()
and ZSTD_initCStream_usingCDict().
2017-04-26 15:42:10 -07:00
Yann Collet e42afbc6fa Comply with suggested comments by @terrelln
created FSE_CTABLE_SIZE() and FSE_DTABLE_SIZE()
2017-04-26 11:39:35 -07:00
Yann Collet 7271203bdb transferred entropy scratch space from CCtx into workSpace
Saved 6 KB
2017-04-20 23:21:19 -07:00
Yann Collet a408645f50 made some room for entropy scratch space 2017-04-20 23:09:39 -07:00
Yann Collet 71aaa32c3c transferred FSE tables from CCtx into workspace
Saved 5 KB from CCtx
2017-04-20 23:03:38 -07:00
Yann Collet 71ddeb67b1 made room in workspace for FSE tables
still need to be transferred from CCtx into workspace
2017-04-20 22:54:54 -07:00
Yann Collet a34a39c183 changed size evaluation of entropy tables
so that memcpy() does no longer depends on fse pointer being a static table
2017-04-20 18:26:25 -07:00
Yann Collet 7bb60b17d8 init entropy table pointers only once
per workSpace resize
2017-04-20 17:38:56 -07:00
Yann Collet e6fa70a0a1 reorganized ZSTD_resetCCtx_internal()
clearer separation between variables and buffers
clearer buffers category
kept static buffers at the beginning, favoring cache locality
(it will be easier to add FSE tables there later)

This break a few assumptions that hashTable was always at the beginning.
This is fixed.
And remaining assumptions (namely that tables stand next to each other in memory)
are now tested with assert.
2017-04-20 17:28:31 -07:00
Yann Collet c17e020c9a disable assert when compiling paramgrill
paramgrill is a benchmark calibration function.
Speed accuracy is critical, it cannot be altered by assert.
2017-04-20 12:50:02 -07:00
Yann Collet 16f9c572fc Merge branch 'dev' into compressionFlow 2017-04-20 11:16:40 -07:00
Yann Collet e348dad305 minor long line reformatting 2017-04-20 11:14:13 -07:00
Yann Collet 2c5514c759 fixed ZSTDMT_initCStream_advanced()
Must use the new ZSTD_compressBegin_usingCDict_advanced()
to enforce correct frame parameters
2017-04-18 22:52:41 -07:00
Yann Collet a4cab80183 added ZSTD_copyCCtx_internal()
which respects provided fParams.
2017-04-18 14:54:54 -07:00
Yann Collet 30fb499208 Changed ZSTD_resetCCtx_advanced() into ZSTD_resetCCtx_internal()
for naming consistency :
_advanced() can be invoked
while _internal() are strictly static
2017-04-18 14:08:50 -07:00
Yann Collet 715b9aa113 created ZSTD_compressBegin_usingCDict_advanced() 2017-04-18 13:55:53 -07:00
Yann Collet 4f818182b8 clarified frame parameters for ZSTD_compress*_usingCDict()
created ZSTD_compressBegin_usingCDict_internal(),
which gives direct control to frame Parameters.
ZSTD_resetCStream_internal() now points into it.
2017-04-17 18:29:06 -07:00
Yann Collet c47c68f6ca proper evaluation of Huffman CTable size 2017-04-17 16:14:21 -07:00
Yann Collet 88009a8ba2 removed srcSize control from CStream
since it's already done from lower bufferless API level
2017-04-12 00:51:24 -07:00
Yann Collet 20d5e03893 content size is controlled at bufferless level
so it's active for all entry points

Also : added relevant test (wrong content size) in fuzzer
2017-04-11 18:34:02 -07:00
Yann Collet 4ee6b15dac force contentSizeFlag=0 when using ZSTD_initCStream_usingCDict()
because by definition srcSize is not known when using this prototype.
added relevant test

Note : this use was already working, because at a later stage
(both ZSTD_compressBegin_usingCDict() and ZSTD_copyCCtx())
pledgedSrcSize=0 is translated into "unknown", no matter the frame parameter.
This is not correct, but of little importance,
as the medium term plan is to no longer set fParams within CDict
2017-04-11 11:59:44 -07:00
Yann Collet ab9162ebb4 simplified call graph
by calling ZSTD_compressBegin_internal() instead of ZSTD_compressBegin_advanced()
2017-04-11 10:46:20 -07:00
Yann Collet e88034fe26 simplified ZSTD_initCStream*() flow
all variants converge towards ZSTD_initCStream_stage2()
2017-04-10 22:24:02 -07:00
Yann Collet 4b987ad8ce Introduce ZSTD_initCStream_internal()
This is now the regroup point for ZSTD_initCStream*() functions

ZSTD_initCStream_advanced() now properly checks for parameters validity.

Also : added <assert.h> usage inside zstd_compress.c
Needs ZSTD_DEBUG=1 macro to be triggered.
Will be triggered by default from `tests` directory
2017-04-10 17:50:44 -07:00
Yann Collet 0181fef545 ensure cctx internal buffer is correctly sized in case of memory error 2017-04-06 01:25:26 -07:00
Yann Collet 02d37aa1c1 ensure correct size of internal buffers in case of error 2017-04-05 14:53:51 -07:00
Nick Terrell 405d2a1027 Explicitly convert scratchBuffer to unsigned* 2017-04-04 16:35:31 -07:00
Nick Terrell 16a739cab0 Switch call of FSE_count() to FSE_count_wksp() 2017-04-04 16:17:21 -07:00
Yann Collet 7cf78f1be7 Protects ZSTD_compressBegin_usingCDict() vs NULL cdict dereference
Will issue an error (GENERIC) is cdict==NULL
2017-04-04 12:38:14 -07:00
Nick Terrell 26b046a7c4 Remove unnecessary dictID store 2017-04-03 21:46:28 -07:00
Nick Terrell 39a6cc5172 Make ZSTD_compress_usingCDict() respect contentSizeFlag 2017-04-03 21:09:55 -07:00
Nick Terrell 62ecad3819 Fix ZSTD_initCStream_usingCDict() to use dictionary 2017-04-03 21:05:59 -07:00
Yann Collet 30c7698970 optimize ZSTDMT_compress() memory usage
does no longer allocate temporary buffers
when there is enough room in dstBuffer to decompress directly there.
(previous method would skip that for 1st chunk only).

Also : fix ZSTD_compressBound() for small srcSize
2017-03-31 18:27:03 -07:00
Yann Collet 3f75d52527 Changed ZSTD_compressBound()
required so that if Total = A+B
compressBound(Total) <= compressBound(A) + compressBound(B)
under condition of a minimum size for A and B

Will help for ZSTDMT_compress() memory allocation
2017-03-31 17:11:38 -07:00
Yann Collet eea7858e2b fixed minor warnings in debug code 2017-03-30 16:47:19 -07:00
Yann Collet 34cc487d05 overlap at full windowSize for max compression level
as it provides max compression ratio
2017-03-30 16:23:22 -07:00
Yann Collet 458e955c23 improved ZSTDMT_compress()
Use a bit more threads by default.
Uses overlap segments to boost compression ratio (like the streaming variant)
2017-03-30 15:51:58 -07:00
Yann Collet 6476c51b86 Merge pull request #637 from facebook/zstdmt
Zstdmt
2017-03-30 14:18:37 -07:00
Nick Terrell 5152fb2cb2 Convert all tabs to spaces 2017-03-29 18:51:58 -07:00
Yann Collet ca5a8bbe36 re-added patch ... 2017-03-29 17:15:27 -07:00
Yann Collet 2e2e78de47 removed unnecessary restriction on minmatchLength
it's now transparently translated to nearest value when unsupported
(7->6) (3->4)
2017-03-29 16:02:47 -07:00
Yann Collet 933ce4a1dd fix : minmatch 7 conversion
minmatch 7 now converted to minmatch 6 for strategies which do not support 7
Used to folded into "default", which applied minmatch 4
2017-03-29 14:35:38 -07:00
Yann Collet 2238870eb6 Merge pull request #625 from facebook/loadCDict
limited CDict acceptation criteria to be the same as DDict
2017-03-24 16:06:20 -07:00