The fix outlined in bug 465484 causes a performance regression for tamarin in misc/primes
- sending to qrb for target milestone, this is marked as a blocker for the TM/TR nanojit merge so needs to be targeted from this release - will be able to track perf change using mentioned performance micro benchmark: test/performance/misc/primes.as
Assignee: nobody → rreitmai
Status: NEW → ASSIGNED
Flags: flashplayer-qrb? → flashplayer-qrb+
Priority: -- → P2
Target Milestone: --- → flash10.1
Summary: Need to come up with a better fix for bug 465484 → Need to come up with a better fix for x86 evict bug 465484
(In reply to comment #0) > The fix outlined in bug 465484 causes a performance regression for tamarin in > misc/primes ...but surely only on non-SSE2 machines, right?
Summary: Need to come up with a better fix for x86 evict bug 465484 → TM merge: Need to come up with a better fix for x86 evict bug 465484
Status: ASSIGNED → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → DUPLICATE
Duplicate of bug: 491084
re-opening to track the work of improving the good-enough fix that is in place now. The existing fix causes a 5% slowdown in misc/primes.as, but negligible slowdown in the rest of the tests.
No longer blocks: 503556
Status: VERIFIED → REOPENED
Priority: P2 → P3
Resolution: DUPLICATE → ---
Assignee: rreitmai → nobody
this is probably low hanging fruit for x86
Retargeting to Future.
Summary: TM merge: Need to come up with a better fix for x86 evict bug 465484 → Need more faster fix for x86 evict bug 465484
Target Milestone: Q4 11 - Anza → Future
Removing Andre blocker, as Dan assigned it two future.
No longer depends on: 641055
You need to log in before you can comment on or make changes to this bug.