NEW 213687
findGCObjectPointersForMarking may have optimization opportunities 
https://bugs.webkit.org/show_bug.cgi?id=213687
Summary findGCObjectPointersForMarking may have optimization opportunities
Keith Miller
Reported 2020-06-27 16:47:07 PDT
https://bugs.webkit.org/show_bug.cgi?id=213686 indicated it was a 2% perf regression on Speedo2 and JetStream2 to not return early from findGCObjectPointersForMarking after finding a live pointer to some HeapKind::JSCell. Likely, that regression comes from trying to determine if the pointer was live. It's possible most of this is from checking if the different pointers are live. Some improvements could be specializing findGCObjectPointersForMarking on the HeapKind or passing all values simultaneously to MarkedBlock::isLive.
Attachments
Add attachment proposed patch, testcase, etc.
Note You need to log in before you can comment on or make changes to this bug.