[Webkit-unassigned] [Bug 69868] DFG virtual register allocator should be more aggressive in reusing temporary slots
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Tue Oct 11 15:38:18 PDT 2011
https://bugs.webkit.org/show_bug.cgi?id=69868
--- Comment #6 from Filip Pizlo <fpizlo at apple.com> 2011-10-11 15:38:18 PST ---
Updated numbers for all benchmarks:
Benchmark report for SunSpider, V8, and Kraken.
VMs tested:
"TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc
"BetterVRAlloc" at /Volumes/Data/pizlo/septenary/OpenSource/WebKitBuild/Release/jsc
Collected 12 samples per benchmark/VM, with 4 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 BetterVRAlloc
SunSpider:
3d-cube 7.3791+-0.1438 7.3422+-0.2024
3d-morph 7.6559+-0.1532 ? 7.6660+-0.1369 ?
3d-raytrace 7.5526+-0.1831 ? 7.6005+-0.1643 ?
access-binary-trees 1.6997+-0.0453 ? 1.7195+-0.0596 ? might be 1.0117x slower
access-fannkuch 6.4543+-0.1042 6.4251+-0.0914
access-nbody 3.3612+-0.0839 3.3161+-0.0804 might be 1.0136x faster
access-nsieve 2.5566+-0.0616 ? 2.6244+-0.0842 ? might be 1.0265x slower
bitops-3bit-bits-in-byte 1.7225+-0.0330 ? 1.7510+-0.0330 ? might be 1.0165x slower
bitops-bits-in-byte 2.7106+-0.0671 ? 2.7899+-0.0492 ? might be 1.0293x slower
bitops-bitwise-and 3.3565+-0.0647 3.3551+-0.1026
bitops-nsieve-bits 5.4728+-0.1269 ? 5.5231+-0.1031 ?
controlflow-recursive 2.0593+-0.0302 ? 2.0820+-0.0449 ? might be 1.0110x slower
crypto-aes 6.7189+-0.1746 ? 6.7529+-0.1620 ?
crypto-md5 2.8141+-0.0487 ? 2.8817+-0.0505 ? might be 1.0240x slower
crypto-sha1 2.4218+-0.0621 ? 2.5025+-0.0589 ? might be 1.0333x slower
date-format-tofte 10.0579+-0.1816 10.0549+-0.1680
date-format-xparb 9.4010+-0.4992 ? 9.6260+-0.4280 ? might be 1.0239x slower
math-cordic 6.5915+-0.1215 6.5062+-0.1217 might be 1.0131x faster
math-partial-sums 7.6552+-0.1328 ? 7.7737+-0.1838 ? might be 1.0155x slower
math-spectral-norm 2.8287+-0.0347 ? 2.8479+-0.0720 ?
regexp-dna 10.7310+-0.1783 10.7082+-0.1497
string-base64 5.3418+-0.1218 5.3307+-0.1068
string-fasta 6.4322+-0.1368 ? 6.4788+-0.2142 ?
string-tagcloud 11.3594+-0.1888 ? 11.3987+-0.2564 ?
string-unpack-code 20.4972+-0.2142 ? 20.6382+-0.3662 ?
string-validate-input 6.2973+-0.1110 ? 6.3350+-0.1157 ?
<arithmetic> * 6.1973+-0.0333 ? 6.2319+-0.0329 ?
<geometric> 5.0876+-0.0263 ? 5.1265+-0.0274 ?
<harmonic> 4.1650+-0.0314 ? 4.2101+-0.0321 ? might be 1.0108x slower
TipOfTree BetterVRAlloc
V8:
crypto 73.4915+-0.5072 ^ 72.5717+-0.3088 ^ definitely 1.0127x faster
deltablue 229.3694+-2.9698 ? 230.5546+-1.8132 ?
earley-boyer 91.6589+-0.2657 ? 92.8696+-1.2849 ? might be 1.0132x slower
raytrace 58.4779+-0.3150 ? 59.0484+-0.5281 ?
regexp 104.5795+-0.3130 104.2271+-0.4320
richards 207.9325+-1.2140 ^ 189.4965+-1.3250 ^ definitely 1.0973x faster
splay 96.3336+-0.5974 95.9125+-0.5882
<arithmetic> 123.1205+-0.4639 ^ 120.6686+-0.3824 ^ definitely 1.0203x faster
<geometric> * 109.5373+-0.2146 ^ 108.2105+-0.3030 ^ definitely 1.0123x faster
<harmonic> 98.9614+-0.1761 ^ 98.4156+-0.2803 ^ definitely 1.0055x faster
TipOfTree BetterVRAlloc
Kraken:
ai-astar 503.0404+-2.6460 501.6886+-4.6412
audio-beat-detection 194.5291+-1.4049 ^ 191.3495+-0.9404 ^ definitely 1.0166x faster
audio-dft 272.6498+-2.3990 ? 279.1601+-5.1765 ? might be 1.0239x slower
audio-fft 125.3488+-0.6055 ? 125.3593+-0.5099 ?
audio-oscillator 253.9799+-2.2824 253.1516+-1.7852
imaging-darkroom 423.5625+-5.1567 ? 425.0566+-8.0295 ?
imaging-desaturate 233.2266+-1.3487 232.7642+-1.3199
imaging-gaussian-blur 586.1470+-0.9626 ? 587.3914+-1.0103 ?
json-parse-financial 55.5272+-0.1996 ! 56.1811+-0.3381 ! definitely 1.0118x slower
json-stringify-tinderbox 69.5921+-1.4267 68.9636+-0.4248
stanford-crypto-aes 133.6110+-1.4540 133.2715+-1.5036
stanford-crypto-ccm 102.0949+-0.4570 101.1988+-0.6843
stanford-crypto-pbkdf2 194.6565+-1.3833 ^ 192.1962+-0.7880 ^ definitely 1.0128x faster
stanford-crypto-sha256-iterative 72.4658+-0.2888 ^ 71.5203+-0.3084 ^ definitely 1.0132x faster
<arithmetic> * 230.0308+-0.5548 229.9466+-0.9315
<geometric> 179.6701+-0.5064 179.2885+-0.6065
<harmonic> 140.4526+-0.5347 140.0540+-0.3782
TipOfTree BetterVRAlloc
All benchmarks:
<arithmetic> 90.2852+-0.1993 89.9141+-0.2787
<geometric> 23.2351+-0.0815 ? 23.2761+-0.0696 ?
<harmonic> 7.3290+-0.0539 ? 7.4054+-0.0549 ? might be 1.0104x slower
TipOfTree BetterVRAlloc
Geomean of preferred means:
<scaled-result> 53.8492+-0.1318 53.7238+-0.1173
--
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