[Webkit-unassigned] [Bug 196240] New: [bmalloc] bmalloc should handle realloc for larger size well

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Mar 25 21:00:56 PDT 2019


https://bugs.webkit.org/show_bug.cgi?id=196240

            Bug ID: 196240
           Summary: [bmalloc] bmalloc should handle realloc for larger
                    size well
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: bmalloc
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: ysuzuki at apple.com
                CC: ggaren at apple.com

We should explore the chance to use existing contiguous regions when "realloc" is called for large objects.
It can avoid deallocations & allocations. It is particularly nice for current and peak memory footprint.

1. For current memory footprint, "realloc" avoids deallocations. If deallocations happen, we do not see memory progression immediately. It will be delayed by asynchronous scavenger. So avoiding deallocations gets pure win in terms of "current" memory footprint.
2. For peak memory footprint, we do not have "new" and "old" memory region at the same time. So we can make peak memory footprint small.


In system malloc, using "realloc" aggressively in JSC Butterfly allocations shows 3% progression in RAMification non-JIT tests.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20190326/60f5eefe/attachment.html>


More information about the webkit-unassigned mailing list