[webkit-reviews] review requested: [Bug 64493] GC allocation fast path has too many operations : [Attachment 100735] the patch
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Wed Jul 13 17:35:29 PDT 2011
Filip Pizlo <fpizlo at apple.com> has asked for review:
Bug 64493: GC allocation fast path has too many operations
https://bugs.webkit.org/show_bug.cgi?id=64493
Attachment 100735: the patch
https://bugs.webkit.org/attachment.cgi?id=100735&action=review
------- Additional Comments from Filip Pizlo <fpizlo at apple.com>
This patch results in the following performance wins:
[pizlo at minime PerformanceTests] ../Tools/Scripts/sunspider-compare-results --v8
v8-v4-results/sunspider-results-2011-07-13-16.43.10.js
v8-v4-results/sunspider-results-2011-07-13-16.38.09.js
TEST COMPARISON FROM TO
DETAILS
=============================================================================
** TOTAL **: 1.019x as fast 1205.1ms +/- 0.4% 1182.6ms +/- 0.2%
significant
=============================================================================
v8: 1.019x as fast 1205.1ms +/- 0.4% 1182.6ms +/- 0.2%
significant
crypto: - 196.3ms +/- 0.5% 195.6ms +/- 0.7%
deltablue: 1.026x as fast 252.7ms +/- 1.6% 246.3ms +/- 1.0%
significant
earley-boyer: 1.028x as fast 136.6ms +/- 0.4% 132.9ms +/- 0.4%
significant
raytrace: 1.013x as fast 75.4ms +/- 0.8% 74.4ms +/- 0.5%
significant
regexp: 1.061x as fast 114.4ms +/- 1.0% 107.8ms +/- 0.4%
significant
richards: - 227.8ms +/- 1.1% 226.2ms +/- 0.7%
splay: 1.013x as fast 201.9ms +/- 1.0% 199.4ms +/- 0.8%
significant
[pizlo at minime PerformanceTests] ../Tools/Scripts/sunspider-compare-results
sunspider-1.0-results/sunspider-results-2011-07-13-16.42.42.js
sunspider-1.0-results/sunspider-results-2011-07-13-16.39.19.js
TEST COMPARISON FROM TO
DETAILS
=============================================================================
** TOTAL **: 1.006x as fast 177.1ms +/- 0.2% 176.0ms +/- 0.2%
significant
=============================================================================
3d: - 25.3ms +/- 0.8% 25.1ms +/- 0.4%
cube: - 9.0ms +/- 0.4% 9.0ms +/- 0.8%
morph: - 7.2ms +/- 2.6% 7.1ms +/- 1.1%
raytrace: - 9.0ms +/- 0.4% 9.0ms +/- 0.0%
access: - 22.4ms +/- 0.7% 22.3ms +/- 0.6%
binary-trees: - 2.0ms +/- 2.8% 2.0ms +/- 2.0%
fannkuch: ?? 11.1ms +/- 0.7% 11.2ms +/- 1.0%
not conclusive: might be *1.009x as slow*
nbody: - 6.0ms +/- 0.0% 6.0ms +/- 0.0%
nsieve: 1.071x as fast 3.3ms +/- 4.0% 3.1ms +/- 2.8%
significant
bitops: ?? 15.5ms +/- 1.0% 15.6ms +/- 1.1%
not conclusive: might be *1.001x as slow*
3bit-bits-in-byte: - 2.0ms +/- 0.0% 2.0ms +/- 0.0%
bits-in-byte: ?? 5.5ms +/- 2.6% 5.5ms +/- 2.8%
not conclusive: might be *1.004x as slow*
bitwise-and: - 3.0ms +/- 1.3% 3.0ms +/- 1.3%
nsieve-bits: - 5.0ms +/- 0.0% 5.0ms +/- 0.0%
controlflow: ?? 1.1ms +/- 8.7% 1.2ms +/- 9.1%
not conclusive: might be *1.018x as slow*
recursive: ?? 1.1ms +/- 8.7% 1.2ms +/- 9.1%
not conclusive: might be *1.018x as slow*
crypto: - 11.1ms +/- 0.8% 11.1ms +/- 0.9%
aes: - 7.0ms +/- 0.6% 7.0ms +/- 0.6%
md5: ?? 2.1ms +/- 3.7% 2.1ms +/- 4.1%
not conclusive: might be *1.010x as slow*
sha1: - 2.0ms +/- 0.0% 2.0ms +/- 0.0%
date: 1.033x as fast 22.6ms +/- 0.9% 21.9ms +/- 0.7%
significant
format-tofte: 1.019x as fast 14.0ms +/- 0.7% 13.8ms +/- 0.9%
significant
format-xparb: 1.057x as fast 8.6ms +/- 1.7% 8.1ms +/- 1.1%
significant
math: ?? 16.0ms +/- 0.3% 16.0ms +/- 0.3%
not conclusive: might be *1.003x as slow*
cordic: ?? 6.0ms +/- 0.7% 6.0ms +/- 0.0%
not conclusive: might be *1.003x as slow*
partial-sums: ?? 7.0ms +/- 0.0% 7.0ms +/- 0.6%
not conclusive: might be *1.003x as slow*
spectral-norm: - 3.0ms +/- 0.0% 3.0ms +/- 0.0%
regexp: - 10.0ms +/- 0.4% 10.0ms +/- 0.0%
dna: - 10.0ms +/- 0.4% 10.0ms +/- 0.0%
string: - 53.0ms +/- 0.4% 53.0ms +/- 0.5%
base64: 1.020x as fast 6.0ms +/- 0.0% 5.9ms +/- 1.6%
significant
fasta: - 7.0ms +/- 0.6% 7.0ms +/- 0.0%
tagcloud: - 13.1ms +/- 0.5% 13.0ms +/- 0.4%
unpack-code: ?? 20.6ms +/- 0.8% 20.8ms +/- 0.7%
not conclusive: might be *1.010x as slow*
validate-input: - 6.4ms +/- 2.3% 6.3ms +/- 2.4%
[pizlo at minime PerformanceTests]
More information about the webkit-reviews
mailing list