Chromium Code Reviews
chromiumcodereview-hr@appspot.gserviceaccount.com (chromiumcodereview-hr) | Please choose your nickname with Settings | Help | Chromium Project | Gerrit Changes | Sign out
(304)

Issue 1257573002: Configurable vector capacity in blink::DistributedNodes

Created:
5 years, 5 months ago by Mikhail
Modified:
5 years, 4 months ago
Reviewers:
hayato
CC:
blink-reviews, webcomponents-bugzilla_chromium.org, sof, eae+blinkwatch, blink-reviews-dom_chromium.org, dglazkov+blink, rwlbuis, sigbjorn
Base URL:
https://chromium.googlesource.com/chromium/blink.git@master
Target Ref:
refs/heads/master
Project:
blink
Visibility:
Public.

Description

Configurable vector capacity in blink::DistributedNodes This patch allows setting initial vector capacity in blink::DistributedNodes. It improves the results of the 'LargeDistributionWithoutLayout.html' performance test approximately by 4% (Linux desktop x64). before: avg 299.6348742769932 runs/s median 299.50975629463005 runs/s stdev 2.05745824437092 runs/s min 295.2029520295203 runs/s max 303.323285749493 runs/s after: avg 310.3259225820019 runs/s median 312.8024733902208 runs/s stdev 6.222720323148261 runs/s min 298.0681458298404 runs/s max 317.88500509940576 runs/s

Patch Set 1 #

Unified diffs Side-by-side diffs Delta from patch set Stats (+2 lines, -1 line) Patch
M Source/core/dom/shadow/DistributedNodes.h View 1 chunk +1 line, -0 lines 0 comments Download
M Source/core/dom/shadow/ElementShadow.cpp View 1 chunk +1 line, -1 line 0 comments Download

Messages

Total messages: 4 (1 generated)
Mikhail
PTAL
5 years, 5 months ago (2015-07-24 10:30:26 UTC) #2
hayato
I'm afraid that we don't get meaningful benefits from this CL. LargeDistributionWithoutLayout was introduced as ...
5 years, 4 months ago (2015-07-27 08:03:29 UTC) #3
Mikhail
5 years, 4 months ago (2015-07-28 13:28:00 UTC) #4
On 2015/07/27 08:03:29, hayato wrote:
> I'm afraid that we don't get meaningful benefits from this CL.
> LargeDistributionWithoutLayout was introduced as a micro bench mark to catch a
> regression. We don't need to improve the result of this benchmark.
> 
> Do we have any improvement other than the result of this benchmark?

Not really. Thanks for having a look.

Powered by Google App Engine
This is Rietveld 408576698