Confirmed users
367
edits
No edit summary |
|||
Line 27: | Line 27: | ||
** Break up super-large functions into disjoint regions that can be regalloc'd and optimized separately | ** Break up super-large functions into disjoint regions that can be regalloc'd and optimized separately | ||
** Investigate GVN quadratic behavior : {{bug|845068}} | ** Investigate GVN quadratic behavior : {{bug|845068}} | ||
* Further improve heap access performance : {{bug|870743}}, {{bug|897425}}, {{bug|915157}} | * Further improve heap access performance : {{bug|870743}}, {{bug|897425}}, {{bug|915157}}, {{bug|983131}}, {{bug|865523}} | ||
* For asm.js->asm.js calls on x86, pass the return value via xmm register, not fp stack | * For asm.js->asm.js calls on x86, pass the return value via xmm register, not fp stack | ||
* Extend the asm.js signal handler tricks to: | * Extend the asm.js signal handler tricks to: |