Sen Huang
d88c1d95ce
Remove inlining for opt
2021-09-01 16:58:57 -04:00
Nick Terrell
46f2710562
[HUF] Improve Huffman encoding speed
...
Improve Huffman encoding speed by 20% for gcc and 10% for clang.
| Compiler | Benchmark | Config | Dataset | Ratio | Speed MB/s (dev) | Speed MB/s (huf-cspeed) | Speed MB/s (huf-cspeed - dev) |
|----------|-------------------|---------|-------------|-------|------------------|-------------------------|-------------------------------|
| gcc | compress | level_1 | enwik7 | 2.43 | 253.70 | 258.72 | 2.0% |
| gcc | compress | level_1 | silesia | 2.88 | 341.90 | 348.15 | 1.8% |
| gcc | compress_literals | level_1 | enwik7 | 1.49 | 761.83 | 912.76 | 19.8% |
| gcc | compress_literals | level_1 | silesia | 1.28 | 754.83 | 902.37 | 19.5% |
| gcc | compress_literals | level_7 | enwik7 | 1.29 | 502.81 | 552.79 | 9.9% |
| gcc | compress_literals | level_7 | silesia | 1.11 | 675.97 | 776.44 | 14.9% |
| clang | compress | level_1 | enwik7 | 2.43 | 277.54 | 280.98 | 1.2% |
| clang | compress | level_1 | silesia | 2.88 | 369.98 | 375.46 | 1.5% |
| clang | compress_literals | level_1 | enwik7 | 1.49 | 828.83 | 918.41 | 10.8% |
| clang | compress_literals | level_1 | silesia | 1.28 | 815.81 | 905.41 | 11.0% |
| clang | compress_literals | level_7 | enwik7 | 1.29 | 533.13 | 553.30 | 3.8% |
| clang | compress_literals | level_7 | silesia | 1.11 | 714.52 | 775.38 | 8.5% |
2021-07-27 15:10:35 -07:00
Nick Terrell
91c9a247b6
[lib] Fix determinism bug in the optimal parser
...
`ZSTD_insertBt1()` has a speed optimization that skips the prefix of
very long matches.
40def70387/lib/compress/zstd_opt.c (L476)
This optimization is based off the length longest match found. However,
when indices are reset, we only ensure that we can reference the whole
window starting from `ip`. If the previous block ended with a long match
then `nextToUpdate` could be much less than `ip`. It might be far enough
back that `nextToUpdate < maxDist`, so it doesn't have a full window of
data to reference. This can cause non-determinism bugs, because we may
find a match that is beyond `ip - maxDist`, and may sometimes be
un-referencable, and that match triggers the speed optimization.
The fix is to base the `windowLow` off of the `target` of
`ZSTD_updateTree_internal()`, because anything below that value will be
obsolete by the time `ZSTD_updateTree_internal()` completes.
2021-05-13 17:05:59 -07:00
Nick Terrell
a494308ae9
[copyright][license] Switch to yearless copyright and some cleanup in the linux-kernel files
...
* Switch to yearless copyright per FB policy
* Fix up SPDX-License-Identifier lines in `contrib/linux-kernel` sources
* Add zstd copyright/license header to the `contrib/linux-kernel` sources
* Update the `tests/test-license.py` to check for yearless copyright
* Improvements to `tests/test-license.py`
* Check `contrib/linux-kernel` in `tests/test-license.py`
2021-03-30 10:30:43 -07:00
Nick Terrell
66e811d782
[license] Update year to 2021
2021-01-04 17:53:52 -05:00
senhuang42
d6911b86be
Require LDM matches to be strictly greater in length
2020-10-09 12:56:18 -04:00
senhuang42
b9c8033cde
Define kNullRawSeqStore for every file
2020-10-07 19:02:41 -04:00
senhuang42
a6165c1b28
Change matchState_t::ldmSeqStore to pointer
2020-10-07 14:13:57 -04:00
senhuang42
abce708a56
Move posInSequence correction to correct location
2020-10-07 13:56:25 -04:00
senhuang42
0c515590d8
Replace offCode of largest match if ldm's offCode is superior
2020-10-07 13:56:25 -04:00
senhuang42
0fac8e07e1
Refactor usage of ms->ldmSeqStore so that it is not modified during compressBlock(), and simplify skipRawSeqStoreBytes
2020-10-07 13:56:25 -04:00
senhuang42
a5500cf2af
Refactor separate ldm variables all into one struct
2020-10-07 13:56:25 -04:00
senhuang42
0325d878f2
Remove bubbling down matches with longer offCode and same matchLen
2020-10-07 13:56:25 -04:00
senhuang42
ddf8a3f1b9
Enable inclusion of mid-flight LDMs in opt parser
2020-10-07 13:56:25 -04:00
senhuang42
88f72ed942
Correct incorrect offcode calculation
2020-10-07 13:56:25 -04:00
senhuang42
d8b43a4202
Add explicit conversion of size_t to U32
2020-10-07 13:56:25 -04:00
senhuang42
b8bfc4e63d
Add cSize regression test to fuzzer.c
2020-10-07 13:56:25 -04:00
senhuang42
c87d2e5866
Prefix new static ldm helpers with ZSTD_opt
2020-10-07 13:56:25 -04:00
senhuang42
429dec4f42
Add DEBUGLOG() calls in ldm helpers
2020-10-07 13:56:25 -04:00
senhuang42
10647924f1
Make function descriptions more accurate
2020-10-07 13:56:25 -04:00
senhuang42
37617e23d7
Correct matchLength calculation and remove unnecessary functions
2020-10-07 13:56:25 -04:00
senhuang42
7dee62c287
Reset ldmSeqStore after initStats_ultra() pass for btultra2
2020-10-07 13:56:25 -04:00
senhuang42
0718aa70df
Refactor existing functions to use posInSequence
2020-10-07 13:56:25 -04:00
senhuang42
7348b40a87
Adjustments to ldm_calculateMatchRange() to calculate bounds correctly
2020-10-07 13:56:25 -04:00
senhuang42
a1ef2db5b2
Add ldm_calculateMatchRange() function
2020-10-07 13:56:25 -04:00
senhuang42
4793ae3b84
Prevent duplicate LDMs from being inserted
2020-10-07 13:56:25 -04:00
senhuang42
65f9cfeeec
Add extra bounds check to prevent heap access after free ASAN error
2020-10-07 13:56:25 -04:00
senhuang42
bff5785fd5
Address mixed variables C90 warning
2020-10-07 13:56:25 -04:00
senhuang42
724b94ed18
ldm_getNextMatch fixed return values
2020-10-07 13:56:25 -04:00
senhuang42
ea92fb3a68
Cleanups, add comments and explanations
2020-10-07 13:56:25 -04:00
senhuang42
78da2e1808
Fixed sifting algorithm
2020-10-07 13:56:25 -04:00
senhuang42
6ccd97fc96
Fixed end of match boundary update issues
2020-10-07 13:56:25 -04:00
senhuang42
28394b64f2
Add proper bounds check on adding ldms
2020-10-07 13:56:25 -04:00
senhuang42
a2f2b58d04
Add a function ldm_voidSequences()
2020-10-07 13:56:25 -04:00
senhuang42
9c3c7cd20e
Fix function argument to getNextMatch()
2020-10-07 13:56:25 -04:00
senhuang42
c8b8572b38
Adjustments to no longer segfault on nci
2020-10-07 13:56:25 -04:00
senhuang42
5df9b5e05f
Add initial getNextMatch() in opt parser
2020-10-07 13:56:25 -04:00
senhuang42
f8ce7cabc3
Added more debugging
2020-10-07 13:56:25 -04:00
senhuang42
84009a076a
Add re-copying of ldmSeqStore after processing
2020-10-07 13:56:25 -04:00
senhuang42
42395a70c2
Add debug statements, flesh out functions
2020-10-07 13:56:25 -04:00
senhuang42
dd3dd199bb
Get zstd to build with new functions and callsites, fix arguments
2020-10-07 13:56:25 -04:00
senhuang42
766c4a8c28
Implement part of ldm_maybeAddLdm()
2020-10-07 13:56:25 -04:00
senhuang42
84777059d2
Implement ldm_getNextMatch()
2020-10-07 13:56:24 -04:00
senhuang42
28c74bf591
Implement basic splitSequence and skipSequence functions
2020-10-07 13:56:24 -04:00
senhuang42
634ab7830d
Flesh out required args for ldm_handleLdm()
2020-10-07 13:56:24 -04:00
senhuang42
db70761032
Add callsites to appropriate locations in ..opt_generic()
2020-10-07 13:56:24 -04:00
senhuang42
aea61e3c91
Add ldm helper function declarations into opt parser
2020-10-07 13:56:24 -04:00
Nick Terrell
f91ed5c766
[lib] s/current/curr because it collides with Linux Kernel macro
2020-09-09 14:35:39 -07:00
Nick Terrell
c465f24457
ZSTD_ prefix mem{cpy,move,set},malloc,calloc,free
2020-08-26 12:26:03 -07:00
Nick Terrell
6d687a8816
[lib] Fix dictionary + repcodes + optimal parser
2020-05-12 10:36:53 -07:00