[Webkit-unassigned] [Bug 68593] DFG JIT should infer which uses of a variable are not aliased

Tue Sep 27 14:53:29 PDT 2011

https://bugs.webkit.org/show_bug.cgi?id=68593

Filip Pizlo <fpizlo at apple.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #108794|0                           |1
        is obsolete|                            |
 Attachment #108904|                            |review?
               Flag|                            |

--- Comment #6 from Filip Pizlo <fpizlo at apple.com>  2011-09-27 14:53:29 PST ---
Created an attachment (id=108904)
 --> (https://bugs.webkit.org/attachment.cgi?id=108904&action=review)
the patch

This undoes a 2% speed-up that I had seen on my laptop, but that was not seen on arewefastyet.com.  I investigated this to death, and found that the DFG does exactly the same thing, in exactly the same order, on the benchmark that has the 14% fluctuation.  Compile time is definitely not an issue.  I'm guessing that the fluctuation that I had seen (14% win after getting rid of static predictions, and a 14% slow-down with this patch) was just some kind of systematic noise and I'm inclined to ignore it.

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