[Webkit-unassigned] [Bug 72878] Strength reduction for Mul and Mod operations for known constants in DFG

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=72878





--- Comment #14 from Filip Pizlo <fpizlo at apple.com>  2011-11-23 11:33:46 PST ---
(In reply to comment #13)
> (In reply to comment #8)
> > >     format-xparb:      1.091x as fast     12.0ms +/- 0.0%    11.0ms +/- 0.0%     significant
> > 
> > Something to keep in mind: date-format-xparb fluctuates by 10% every time I breathe on the DFG, for any reason.  So I tend to ignore it.  That's not to say that this isn't a true progression.  But taking progressions on xparb seriously is bad luck: next time you make a patch you might find xparb regressing by 10%, for no reason at all! ;-)
> 
> Sounds like the integral milliseconds truncation of bug 71801.

This pathology shows up even when doing microsecond timing with Tools/Scripts/bencher.

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