[Webkit-unassigned] [Bug 102686] [V8] toFastV8 for non-Nodes

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=102686





--- Comment #15 from Dan Carney <dcarney at google.com>  2012-12-04 07:19:50 PST ---
(In reply to comment #10)
> Can you run dom-traverse to make sure we're not regressing?  I suspect calling GetIsolate one extra time will show up on that perf test.  (It should be fixable as described above.)

Your comment no longer applies, but here are the numbers, run twice each with interwoven runs. Looks like a small perf bump.

with patch:

avg 3069.1348651348653 runs/s
median 0 runs/s
stdev 21.196854488648402 runs/s
min 3009.99000999001 runs/s
max 3088.81018981019 runs/s

avg 3192.8747252747253 runs/s
median 0 runs/s
stdev 12.296375015512377 runs/s
min 3163.9750249750246 runs/s
max 3213.18981018981 runs/s

without patch:

avg 2963.217982017982 runs/s
median 0 runs/s
stdev 16.350015231822884 runs/s
min 2920.834165834166 runs/s
max 2987.306693306693 runs/s

avg 2925.672727272727 runs/s
median 0 runs/s
stdev 6.087018068830624 runs/s
min 2908.5304695304694 runs/s
max 2936.0009990009994 runs/s

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