[Webkit-unassigned] [Bug 158619] [JSC] Inline JSC::toInt32 to improve kraken

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=158619

--- Comment #1 from Yusuke Suzuki <utatane.tea at gmail.com> ---
Seems good. I'll check Octane, SunSpider, ASMBench and upload the patch.

Benchmark report for Kraken on hanayamata.

VMs tested:
"baseline" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf-master/Release/bin/jsc
"patched" at /home/yusukesuzuki/dev/WebKit/WebKitBuild/perf/Release/bin/jsc

Collected 100 samples per benchmark/VM, with 100 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.

                                           baseline                  patched                                      

ai-astar                                96.084+-0.548             96.015+-0.523         
audio-beat-detection                    48.621+-1.138             48.357+-1.020         
audio-dft                              123.803+-0.373            123.549+-0.109         
audio-fft                               37.399+-0.050             37.381+-0.039         
audio-oscillator                        53.682+-0.071      ?      53.718+-0.082         ?
imaging-darkroom                        87.673+-0.063      ?      87.724+-0.060         ?
imaging-desaturate                      56.400+-0.064             56.333+-0.041         
imaging-gaussian-blur                   79.772+-1.849      ?      79.936+-1.394         ?
json-parse-financial                    44.649+-0.208      ?      44.997+-0.302         ?
json-stringify-tinderbox                27.110+-0.056             26.812+-0.428           might be 1.0111x faster
stanford-crypto-aes                     43.989+-0.157      ^      43.001+-0.078         ^ definitely 1.0230x faster
stanford-crypto-ccm                     44.338+-0.688      ?      44.588+-0.686         ?
stanford-crypto-pbkdf2                 108.420+-1.044      ^     106.231+-0.437         ^ definitely 1.0206x faster
stanford-crypto-sha256-iterative        35.538+-0.077      ^      35.314+-0.037         ^ definitely 1.0063x faster

<arithmetic>                            63.391+-0.168             63.140+-0.140           might be 1.0040x faster

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160610/d5005e4b/attachment-0001.html>