[Webkit-unassigned] [Bug 69235] DFG should speculate more aggressively on obvious cases on polymorphic get_by_id
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Sun Oct 2 23:08:51 PDT 2011
https://bugs.webkit.org/show_bug.cgi?id=69235
--- Comment #2 from Filip Pizlo <fpizlo at apple.com> 2011-10-02 23:08:51 PST ---
Updated numbers.
Benchmark report for SunSpider, V8, and Kraken.
VMs tested:
"TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc
"PolyGetById" at /Volumes/Data/pizlo/senary/OpenSource/WebKitBuild/Release/jsc
Collected 30 samples per benchmark/VM, with 10 VM invocations per benchmark. 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 PolyGetById
SunSpider:
3d-cube 7.4617+-0.1034 7.4526+-0.1151
3d-morph 7.4606+-0.0887 7.4365+-0.0834
3d-raytrace 8.2420+-0.1147 ^ 7.7820+-0.1234 ^ definitely 1.0591x faster
access-binary-trees 1.7895+-0.0448 ? 1.7961+-0.0455 ?
access-fannkuch 6.3124+-0.0547 ? 6.3763+-0.0769 ? might be 1.0101x slower
access-nbody 3.5895+-0.0703 3.5023+-0.0547 might be 1.0249x faster
access-nsieve 2.6121+-0.0409 ? 2.6302+-0.0403 ?
bitops-3bit-bits-in-byte 1.7307+-0.0198 ? 1.7413+-0.0154 ?
bitops-bits-in-byte 2.7374+-0.0427 ? 2.7884+-0.0551 ? might be 1.0187x slower
bitops-bitwise-and 3.2574+-0.0601 3.2269+-0.0450
bitops-nsieve-bits 5.4171+-0.0626 ? 5.4872+-0.0837 ? might be 1.0129x slower
controlflow-recursive 2.0842+-0.0298 ? 2.0907+-0.0324 ?
crypto-aes 6.8417+-0.1346 6.7743+-0.1371
crypto-md5 2.8669+-0.0543 2.8525+-0.0602
crypto-sha1 2.5060+-0.0458 2.5034+-0.0418
date-format-tofte 10.1911+-0.1981 9.9102+-0.1504 might be 1.0284x faster
date-format-xparb 9.4119+-0.1436 9.3752+-0.1469
math-cordic 6.2322+-0.0542 ? 6.2370+-0.0557 ?
math-partial-sums 7.5658+-0.0878 7.5025+-0.0732
math-spectral-norm 2.8764+-0.0392 2.8682+-0.0580
regexp-dna 10.9170+-0.0949 ? 10.9202+-0.1689 ?
string-base64 5.6590+-0.0913 5.6038+-0.1394
string-fasta 6.6458+-0.0872 ? 6.6492+-0.1068 ?
string-tagcloud 11.8912+-0.1948 ? 12.0330+-0.2048 ? might be 1.0119x slower
string-unpack-code 21.3281+-0.2824 21.1507+-0.2839
string-validate-input 6.3290+-0.1368 6.3204+-0.0888
<arithmetic> * 6.3060+-0.0137 6.2697+-0.0253
<geometric> 5.1665+-0.0136 5.1465+-0.0210
<harmonic> 4.2350+-0.0197 4.2296+-0.0251
TipOfTree PolyGetById
V8:
crypto 72.1259+-0.2271 ? 72.2060+-0.3149 ?
deltablue 225.0479+-1.0014 ? 226.4046+-1.0914 ?
earley-boyer 87.5991+-0.1713 ? 87.6375+-0.2942 ?
raytrace 61.4064+-0.2985 61.3754+-0.3465
regexp 104.0808+-0.4639 103.7140+-0.3796
richards 186.9374+-0.5667 186.4882+-0.5270
splay 91.2943+-0.3209 91.2434+-0.3609
<arithmetic> 118.3560+-0.1797 ? 118.4385+-0.2442 ?
<geometric> * 106.4654+-0.1328 ? 106.4734+-0.2054 ?
<harmonic> 97.3225+-0.1318 97.3019+-0.2024
TipOfTree PolyGetById
Kraken:
ai-astar 492.7137+-2.1612 ^ 488.8511+-1.3972 ^ definitely 1.0079x faster
audio-beat-detection 192.8215+-1.4299 ^ 190.6301+-0.5796 ^ definitely 1.0115x faster
audio-dft 278.8327+-1.5564 ^ 271.9584+-1.7589 ^ definitely 1.0253x faster
audio-fft 128.4893+-0.7137 ? 128.5923+-0.6978 ?
audio-oscillator 257.6952+-1.4953 ^ 246.6424+-0.9898 ^ definitely 1.0448x faster
imaging-darkroom 420.0356+-1.3004 419.9649+-1.3683
imaging-desaturate 225.0034+-0.9091 223.9795+-0.4624
imaging-gaussian-blur 582.8256+-1.7895 581.5096+-1.4356
json-parse-financial 49.2039+-0.2322 48.9254+-0.2208
json-stringify-tinderbox 68.1326+-0.2636 ! 68.9949+-0.2137 ! definitely 1.0127x slower
stanford-crypto-aes 130.5265+-1.0122 130.4759+-1.3555
stanford-crypto-ccm 102.2631+-0.5399 ^ 100.8593+-0.3557 ^ definitely 1.0139x faster
stanford-crypto-pbkdf2 195.9558+-1.1230 193.9616+-1.0843 might be 1.0103x faster
stanford-crypto-sha256-iterative 77.5329+-0.3205 77.4402+-0.5185
<arithmetic> * 228.7166+-0.3183 ^ 226.6275+-0.3175 ^ definitely 1.0092x faster
<geometric> 178.2956+-0.2542 ^ 176.8639+-0.3042 ^ definitely 1.0081x faster
<harmonic> 137.9784+-0.2156 ^ 137.3059+-0.2782 ^ definitely 1.0049x faster
TipOfTree PolyGetById
All benchmarks:
<arithmetic> 89.2443+-0.0906 ^ 88.6142+-0.1108 ^ definitely 1.0071x faster
<geometric> 23.2815+-0.0377 ^ 23.1759+-0.0602 ^ definitely 1.0046x faster
<harmonic> 7.4452+-0.0338 7.4355+-0.0430
TipOfTree PolyGetById
Geomean of preferred means:
<scaled-result> 53.5489+-0.0465 ^ 53.2835+-0.0919 ^ definitely 1.0050x faster
--
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