[Webkit-unassigned] [Bug 70995] The GC should be parallel
bugzilla-daemon at webkit.org
bugzilla-daemon at webkit.org
Thu Oct 27 00:41:51 PDT 2011
https://bugs.webkit.org/show_bug.cgi?id=70995
--- Comment #1 from Filip Pizlo <fpizlo at apple.com> 2011-10-27 00:41:51 PST ---
Created an attachment (id=112647)
--> (https://bugs.webkit.org/attachment.cgi?id=112647&action=review)
work in progress
This works, and produces speed-ups, but still needs clean-ups. Here's the performance using my harness tuned to measure GC behavior.
Benchmark report for V8.
VMs tested:
"TipOfTree" at /Volumes/Data/pizlo/quinary/OpenSource/WebKitBuild/Release/jsc
"ParallelGC" at /Volumes/Data/pizlo/senary/OpenSource/WebKitBuild/Release/jsc
Collected 8 samples per benchmark/VM, with 4 VM invocations per benchmark. Ran 5 benchmark
iterations, and measured the total time of those iterations, for each sample. Emitted a call
to gc() between sample measurements. Used 5 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 ParallelGC
crypto 369.6983+-2.7221 ? 370.0403+-1.5868 ?
deltablue 903.7007+-8.3329 ? 911.5094+-5.1716 ?
earley-boyer 462.6758+-1.3969 461.6687+-1.3534
raytrace 315.5462+-1.9768 ? 318.6208+-1.9752 ?
regexp 549.5400+-4.7576 ^ 541.9117+-1.7295 ^ definitely 1.0141x faster
richards 632.6940+-2.0461 ? 634.8623+-2.2790 ?
splay 414.3850+-2.1458 ^ 325.4297+-17.9843 ^ definitely 1.2733x faster
<arithmetic> 521.1771+-1.3120 ^ 509.1490+-2.7390 ^ definitely 1.0236x faster
<geometric> * 492.4826+-1.0591 ^ 476.1071+-3.8258 ^ definitely 1.0344x faster
<harmonic> 468.0389+-1.1209 ^ 448.5087+-4.9235 ^ definitely 1.0435x 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