Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(345)

Issue 10538104: Optimization of some packed array cases. (Closed)

Created:
8 years, 6 months ago by danno
Modified:
8 years, 6 months ago
Reviewers:
Jakob Kummerow
CC:
v8-dev
Visibility:
Public.

Description

Optimization of some packed array cases. R=jkummerow@chromium.org TEST=slight improvement in 3d-morph Committed: https://code.google.com/p/v8/source/detail?r=11777

Patch Set 1 #

Patch Set 2 : Turn off flag" #

Patch Set 3 : Remove debugging changes #

Patch Set 4 : Ready to review #

Total comments: 8
Unified diffs Side-by-side diffs Delta from patch set Stats (+128 lines, -9 lines) Patch
M src/elements.cc View 1 2 3 9 chunks +116 lines, -9 lines 6 comments Download
M src/objects.cc View 2 chunks +12 lines, -0 lines 2 comments Download

Messages

Total messages: 3 (0 generated)
danno
8 years, 6 months ago (2012-06-12 11:20:41 UTC) #1
Jakob Kummerow
LGTM. I don't like how CopySmiToDoubleElements(), CopyPackedSmiToDoubleElements() and CopyObjectToDoubleElements() share what feels like 90% of ...
8 years, 6 months ago (2012-06-12 12:50:43 UTC) #2
danno
8 years, 6 months ago (2012-06-12 15:38:14 UTC) #3
landed

http://codereview.chromium.org/10538104/diff/5001/src/elements.cc
File src/elements.cc (right):

http://codereview.chromium.org/10538104/diff/5001/src/elements.cc#newcode379
src/elements.cc:379: to_end = to_start + copy_size;
On 2012/06/12 12:50:44, Jakob wrote:
> I have a suspicion that the Windows compiler might wish for a
> static_cast<uint32_t>(copy_size) here (and again below).

Done.

http://codereview.chromium.org/10538104/diff/5001/src/elements.cc#newcode425
src/elements.cc:425: if (hole_or_object == the_hole) {
Yes :-) 
On 2012/06/12 12:50:44, Jakob wrote:
> Why this change? Is it faster?

http://codereview.chromium.org/10538104/diff/5001/src/elements.cc#newcode625
src/elements.cc:625: bool is_packed = FLAG_packed_arrays &&
On 2012/06/12 12:50:44, Jakob wrote:
> Another unnecessary FLAG check.

Done.

http://codereview.chromium.org/10538104/diff/5001/src/objects.cc
File src/objects.cc (right):

http://codereview.chromium.org/10538104/diff/5001/src/objects.cc#newcode10026
src/objects.cc:10026: if (FLAG_packed_arrays && IsJSArray()) {
On 2012/06/12 12:50:44, Jakob wrote:
> This flag check isn't necessary, as we can't have PACKED elements kinds
without
> the flag anyway. Same below.

Done.

Powered by Google App Engine
This is Rietveld 408576698