[Webkit-unassigned] [Bug 69868] DFG virtual register allocator should be more aggressive in reusing temporary slots

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=69868





--- Comment #5 from Filip Pizlo <fpizlo at apple.com>  2011-10-11 15:28:31 PST ---
It turns out that my previous attempt at a virtual register allocator was a total bust.  I fixed it, and now we have real numbers:



Benchmark report for V8.

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                                  

crypto              72.9077+-0.4529    ?    73.1820+-0.3834       ?
deltablue          233.4981+-2.2990    ^   229.2700+-1.5657       ^ definitely 1.0184x faster
earley-boyer        91.4999+-0.3203    ?    91.6503+-0.4134       ?
raytrace            58.7518+-0.4181         58.5479+-0.4437       
regexp             104.2604+-0.2749        104.1496+-0.4870       
richards           207.6681+-1.7385    ^   188.0565+-0.9151       ^ definitely 1.1043x faster
splay               95.9854+-0.3932         95.5227+-0.5338       

<arithmetic>       123.5102+-0.5281    ^   120.0541+-0.3077       ^ definitely 1.0288x faster
<geometric> *      109.6144+-0.2834    ^   107.7297+-0.1889       ^ definitely 1.0175x faster
<harmonic>          98.9001+-0.1905    ^    98.0251+-0.1539       ^ definitely 1.0089x faster

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.