Chromium Code Reviews
Help | Chromium Project | Sign in
(297)

Issue 11151005: Improve FastCloneShallowArray/ObjectStubs with VFP copying on ARM (Closed)

Can't Edit
Can't Publish+Mail
Start Review
Created:
1 year, 6 months ago by Jay Conrod
Modified:
5 months, 1 week ago
CC:
v8-dev_googlegroups.com
Visibility:
Public.

Description

Improve FastCloneShallowArray/ObjectStubs with VFP copying on ARM
BUG=none
TEST=none

Patch Set 1 #

Total comments: 3

Patch Set 2 : #

Total comments: 1
Unified diffs Side-by-side diffs Delta from patch set Stats (+82 lines, -19 lines) Lint Patch
M src/arm/assembler-arm.h View 1 1 chunk +4 lines, -0 lines 0 comments 0 errors Download
M src/arm/code-stubs-arm.cc View 1 3 chunks +47 lines, -19 lines 0 comments 0 errors Download
M src/arm/macro-assembler-arm.h View 1 1 chunk +9 lines, -0 lines 0 comments 0 errors Download
M src/arm/macro-assembler-arm.cc View 1 1 chunk +22 lines, -0 lines 1 comment 0 errors Download
Trybot results:
Commit:

Messages

Total messages: 6
danno
http://codereview.chromium.org/11151005/diff/1/src/arm/code-stubs-arm.cc File src/arm/code-stubs-arm.cc (right): http://codereview.chromium.org/11151005/diff/1/src/arm/code-stubs-arm.cc#newcode334 src/arm/code-stubs-arm.cc:334: return CpuFeatures::IsSupported(VFP2) ? 128 : 8; By increasing this ...
1 year, 5 months ago #1
Jay Conrod
Thanks for the comments. This new revision should address everything.
1 year, 5 months ago #2
danno
LGTM with a nit. I'll land this for you when the feature freeze is lifted ...
1 year, 5 months ago #3
m.m.capewell
On Cortex-A9 using svn revision 12837, we see a performance decrease of ~6% on Octane/Box2D ...
1 year, 5 months ago #4
Jay Conrod
I need to borrow a Cortex A9 device from someone here to confirm this. Will ...
1 year, 5 months ago #5
danno
1 year, 5 months ago #6
I have been able to confirm a regression on a nexus prime (A9), although the
regression isn't 100% reproducible. I additionally see a 3.5% regression on
Octane/PDF.js, which does seem reproducible. I'll wait to land this until the
regressions can be addressed.
Sign in to reply to this message.

Powered by Google App Engine
RSS Feeds Recent Issues | This issue
This is Rietveld 1280:2d3e6564b7b6