[Webkit-unassigned] [Bug 79700] Old JIT's style of JSVALUE64 strict equality is subtly wrong
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Mon Feb 27 14:11:32 PST 2012
https://bugs.webkit.org/show_bug.cgi?id=79700
--- Comment #1 from Filip Pizlo <fpizlo at apple.com> 2012-02-27 14:11:32 PST ---
Effect of fixing this bug on the old JIT alone:
Benchmark report for SunSpider, V8, and Kraken on bigmac (MacPro5,1).
VMs tested:
"TipOfTree" at /Volumes/Data/pizlo/secondary/OpenSource/WebKitBuild/Release/jsc (r109012)
"FixStrictEq" at /Volumes/Data/pizlo/tertiary/OpenSource/WebKitBuild/Release/jsc (r109012)
Collected 12 samples per benchmark/VM, with 4 VM invocations per benchmark. Emitted a call to gc() between sample
measurements. Used 1 benchmark iteration per VM invocation for warm-up. Used the jsc-specific preciseTime() function
to get microsecond-level timing. Reporting benchmark execution times with 95% confidence intervals in milliseconds.
TipOfTree FixStrictEq
SunSpider:
3d-cube 7.8731+-0.0866 7.7743+-0.0273 might be 1.0127x faster
3d-morph 9.0184+-0.0711 ? 9.0695+-0.0844 ?
3d-raytrace 8.8989+-0.0353 ? 8.9743+-0.0489 ?
access-binary-trees 2.1351+-0.0202 ? 2.1829+-0.0302 ? might be 1.0224x slower
access-fannkuch 12.8129+-0.0538 ? 12.8549+-0.0613 ?
access-nbody 7.5549+-0.0149 ? 7.5795+-0.0325 ?
access-nsieve 4.0846+-0.0496 4.0702+-0.0296
bitops-3bit-bits-in-byte 2.5402+-0.0143 ? 2.5597+-0.0254 ?
bitops-bits-in-byte 6.0092+-0.0462 ? 6.0612+-0.0879 ?
bitops-bitwise-and 4.1884+-0.0428 4.1626+-0.0284
bitops-nsieve-bits 5.8416+-0.0245 5.8223+-0.0116
controlflow-recursive 2.1768+-0.0101 2.1765+-0.0107
crypto-aes 6.7658+-0.0526 ? 6.7666+-0.0462 ?
crypto-md5 3.0531+-0.0219 ? 3.0901+-0.0252 ? might be 1.0121x slower
crypto-sha1 2.5216+-0.0110 ? 2.5669+-0.0407 ? might be 1.0180x slower
date-format-tofte 11.5518+-0.0745 ? 11.6777+-0.0880 ? might be 1.0109x slower
date-format-xparb 10.0422+-0.0722 10.0357+-0.4060
math-cordic 7.3075+-0.0304 ? 7.3452+-0.0550 ?
math-partial-sums 10.4582+-0.0430 10.4473+-0.0368
math-spectral-norm 4.4460+-0.0106 4.4335+-0.0117
regexp-dna 8.9352+-0.0591 ! 9.1242+-0.1154 ! definitely 1.0212x slower
string-base64 4.4846+-0.0407 4.4737+-0.0360
string-fasta 7.4731+-0.0683 ? 7.5027+-0.0554 ?
string-tagcloud 12.9983+-0.0806 12.9008+-0.0748
string-unpack-code 20.5740+-0.1799 ? 20.7335+-0.2388 ?
string-validate-input 5.7883+-0.0450 ? 5.7972+-0.0675 ?
<arithmetic> * 7.2898+-0.0150 ? 7.3147+-0.0276 ? might be 1.0034x slower
<geometric> 6.2071+-0.0124 ? 6.2302+-0.0186 ? might be 1.0037x slower
<harmonic> 5.2143+-0.0125 ? 5.2415+-0.0167 ? might be 1.0052x slower
TipOfTree FixStrictEq
V8:
crypto 200.5909+-0.8850 200.5259+-2.1154
deltablue 273.9431+-1.0710 ? 275.8511+-2.2189 ?
earley-boyer 122.0667+-0.4705 ! 125.0908+-0.9036 ! definitely 1.0248x slower
raytrace 72.4627+-1.3499 70.5994+-0.7281 might be 1.0264x faster
regexp 100.1878+-0.2697 ? 100.2717+-0.5699 ?
richards 273.5731+-1.6529 271.8100+-2.5250
splay 78.6417+-0.3525 ? 78.8001+-0.3159 ?
<arithmetic> 160.2094+-0.3694 ? 160.4213+-0.8186 ? might be 1.0013x slower
<geometric> * 139.8682+-0.4297 ? 139.8975+-0.5637 ? might be 1.0002x slower
<harmonic> 122.8290+-0.5907 122.5509+-0.4585 might be 1.0023x faster
TipOfTree FixStrictEq
Kraken:
ai-astar 2066.426+-23.896 ? 2082.081+-32.564 ?
audio-beat-detection 565.301+-8.300 559.745+-2.729
audio-dft 454.652+-1.785 ! 457.614+-0.894 ! definitely 1.0065x slower
audio-fft 427.331+-1.656 ! 438.617+-5.425 ! definitely 1.0264x slower
audio-oscillator 408.272+-8.005 403.664+-4.012 might be 1.0114x faster
imaging-darkroom 589.170+-13.623 581.885+-9.494 might be 1.0125x faster
imaging-desaturate 620.326+-2.433 ? 620.356+-2.651 ?
imaging-gaussian-blur 2161.544+-6.454 2160.878+-3.860
json-parse-financial 63.443+-0.335 63.031+-0.349
json-stringify-tinderbox 76.483+-0.301 ? 76.857+-0.358 ?
stanford-crypto-aes 150.218+-0.553 ^ 136.480+-0.459 ^ definitely 1.1007x faster
stanford-crypto-ccm 111.246+-0.449 ^ 107.975+-0.204 ^ definitely 1.0303x faster
stanford-crypto-pbkdf2 389.989+-0.585 ? 392.650+-3.546 ?
stanford-crypto-sha256-iterative 153.509+-0.718 ^ 148.666+-0.397 ^ definitely 1.0326x faster
<arithmetic> * 588.422+-1.960 587.893+-2.361 might be 1.0009x faster
<geometric> 346.167+-0.806 ^ 342.577+-0.834 ^ definitely 1.0105x faster
<harmonic> 210.267+-0.541 ^ 206.656+-0.469 ^ definitely 1.0175x faster
TipOfTree FixStrictEq
All benchmarks:
<arithmetic> 203.1683+-0.5834 203.0560+-0.7756 might be 1.0006x faster
<geometric> 32.7020+-0.0484 32.6687+-0.0813 might be 1.0010x faster
<harmonic> 9.1979+-0.0216 ? 9.2423+-0.0290 ? might be 1.0048x slower
TipOfTree FixStrictEq
Geomean of preferred means:
<scaled-result> 39.1475+-0.0646 ? 39.1831+-0.1085 ? might be 1.0009x slower
--
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
More information about the webkit-unassigned
mailing list