This optimization turns out to have a negative effect on i5 processors
throughout the board. This needs further investigation to figure out
the underlying reason but will be rolled out for now.
R=danno@chromium.org
Review URL: https://codereview.chromium.org/
12388016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@13772
ce2b1a6d-e550-0410-aec6-
3dcde31c8c00
} else {
__ jmp(entry, RelocInfo::RUNTIME_ENTRY);
}
- } else if (!needs_lazy_deopt && frame_is_built_) {
- // Optimization for ia32 only that skips the indirection through a
- // jump table entry for conditional deopts if possible.
- ASSERT(cc != no_condition);
- __ j(cc, entry, RelocInfo::RUNTIME_ENTRY);
} else {
// We often have several deopts to the same entry, reuse the last
// jump entry if this is the case.