New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 801141 link

Starred by 1 user
Status: WontFix
Owner:
r...@chromium.org
Closed: Mar 2018
Cc:
rsesek@chromium.org
█ mge...@chromium.org
h...@chromium.org
rnk@google.com
thakis@chromium.org
█ hjd@chromium.org
r...@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 2
Type: Bug-Regression



Sign in to add a comment

1.6% regression in rasterize_and_record_micro.top_25 at 526726:526820

Project Member Reported by hjd@chromium.org, Jan 11 2018 
See the link to graphs below.
Project Member

Comment 1 by 42576172...@developer.gserviceaccount.com, Jan 11 2018 

All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=801141

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=dbfe68bf9754910f9bbde8a96622340b16c6b6685d1e938d2bd06851d8b20b94


Bot(s) for this bug's original alert(s):

linux-release
Project Member

Comment 3 by 42576172...@developer.gserviceaccount.com, Jan 11 2018

Cc: rnk@google.com mge...@chromium.org r...@chromium.org rsesek@chromium.org h...@chromium.org thakis@chromium.org mmenke@chromium.org
Owner: mmenke@chromium.org
Status: Assigned (was: Untriaged)
📍 Found significant differences after each of 2 commits.
https://pinpoint-dot-chromeperf.appspot.com/job/14bc86af040000

Roll clang 318667:321529.
By rnk@google.com · Wed Jan 03 18:32:36 2018
chromium @ 0d72bb1494c94523ae7bd944476a495e8b281bcc

Remove virtual from DnsConfig's destructor
By mmenke@chromium.org · Wed Jan 03 20:26:32 2018
chromium @ 04dee1f1b44ed7bac935a9fa4ccbe37f2a7dc87e

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Comment 4 by mmenke@chromium.org, Jan 11 2018

Owner: ----
Status: Available (was: Assigned)
I removed the virtual keyword from a destructor.  If that made any difference at all, you should be talking to whomever works on clang.  Seems much more likely this was related to the clang change, if it's a real regression.

Comment 5 by r...@chromium.org, Jan 12 2018 

This is probably the clang roll. Perhaps it's the same underlying code change as in  https://crbug.com/801134 .

Comment 6 by sullivan@chromium.org, Jan 25 2018

Owner: r...@chromium.org
Status: Assigned (was: Available)
rnk: Is this something that can be fixed? If not, can you close?

Comment 7 by mmenke@chromium.org, Jan 25 2018

Cc: -mmenke@chromium.org

Comment 8 by vmi...@chromium.org, Feb 5 2018

Components: Internals>GPU>Metrics

Comment 9 by r...@chromium.org, Mar 28 2018

Status: WontFix (was: Assigned)
Based on conversations with our team, I think we're not going to investigate performance regressions smaller than 2% caused by compiler updates. Sometimes arbitrary heuristic changes can benefit some benchmarks and hurt others by small percentages.
►

Sign in to add a comment