All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=622332

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgICgwtidtgoM


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

Google Chrome Linux x64

As per https://bugs.chromium.org/p/chromium/issues/detail?id=608067
"ChromiumPerf/linux/sizes / chrome-stripped / stripped" is the thing we should care about.

https://chromeperf.appspot.com/report?sid=a514f920dd8efd47095d638a55303444b0c85ad57d7b9308c58a37e4dcdf810e&start_rev=401220&end_rev=401230

Then, the graph also shows a regression around rev401226.
This is v8 auto roll.
https://chromium.googlesource.com/chromium/src/+log/f791fc5aab4fd9a3c984591c366edc9f1a0145a1%5E..f791fc5aab4fd9a3c984591c366edc9f1a0145a1?pretty=fuller

Here are v8 changes introduced in this roll.
https://chromium.googlesource.com/v8/v8/+log/7774be71..4f1b295d

And following build config change looks the most suspicious one.
https://chromium.googlesource.com/v8/v8/+/365e32b1302d7d9b97761d5037050b2200d1a323

dpranke, could you check if this size change is reasonable?

That v8 change is definitely expected to increase size, and that order of magnitude change wouldn't surprise me too much. I'll do some local builds to double-check.

Did we see any perf improvements in the same timeframe?

We've only notified on regressions and didn't check improvements.
But, if you have any idea which tests can be improved, you can check it at https://chromeperf.appspot.com/ though the page is down for over quota now.

This won't make M-53, so I'm clearing the label.

I can confirm that you get a ~450k increase from setting that flag on v8. Because we're doing an LTO build on Linux now, I don't know if it's possible to directly compare the sizes of the v8 components, but unless LTO is doing something weird, I don't know what else it could be.

I'm bouncing this to machenbach@ to confirm that this is the correct size/speed tradeoff.

Adding v8 perf sheriffs.

Ping!

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/bbb61d8aea732457e1d5a0c9abdd79eacf1622ae

commit bbb61d8aea732457e1d5a0c9abdd79eacf1622ae
Author: machenbach <machenbach@chromium.org>
Date: Tue Jul 12 08:36:13 2016

[gn] Experiment: Reset -O3 to -O2.

Temporary commit to see performance data. Will be reverted
shortly after.

BUG= chromium:622332 
TBR=vogelheim, jochen
NOTRY=true

Review-Url: https://codereview.chromium.org/2135313002
Cr-Commit-Position: refs/heads/master@{#37667}

[modify] https://crrev.com/bbb61d8aea732457e1d5a0c9abdd79eacf1622ae/gni/v8.gni

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/bbb61d8aea732457e1d5a0c9abdd79eacf1622ae

commit bbb61d8aea732457e1d5a0c9abdd79eacf1622ae
Author: machenbach <machenbach@chromium.org>
Date: Tue Jul 12 08:36:13 2016

[gn] Experiment: Reset -O3 to -O2.

Temporary commit to see performance data. Will be reverted
shortly after.

BUG= chromium:622332 
TBR=vogelheim, jochen
NOTRY=true

Review-Url: https://codereview.chromium.org/2135313002
Cr-Commit-Position: refs/heads/master@{#37667}

[modify] https://crrev.com/bbb61d8aea732457e1d5a0c9abdd79eacf1622ae/gni/v8.gni

The following revision refers to this bug:
  https://chromium.googlesource.com/v8/v8.git/+/ce5265016bfb25d0f0eb9e6a4dfb390cd6dbc8ac

commit ce5265016bfb25d0f0eb9e6a4dfb390cd6dbc8ac
Author: machenbach <machenbach@chromium.org>
Date: Tue Jul 12 18:18:42 2016

Revert of [gn] Experiment: Reset -O3 to -O2. (patchset #1 id:1 of https://codereview.chromium.org/2135313002/ )

Reason for revert:
Data collected

Original issue's description:
> [gn] Experiment: Reset -O3 to -O2.
>
> Temporary commit to see performance data. Will be reverted
> shortly after.
>
> BUG= chromium:622332 
> TBR=vogelheim, jochen
> NOTRY=true
>
> Committed: https://crrev.com/bbb61d8aea732457e1d5a0c9abdd79eacf1622ae
> Cr-Commit-Position: refs/heads/master@{#37667}

TBR=vogelheim@chromium.org,jochen@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= chromium:622332 

Review-Url: https://codereview.chromium.org/2141113002
Cr-Commit-Position: refs/heads/master@{#37690}

[modify] https://crrev.com/ce5265016bfb25d0f0eb9e6a4dfb390cd6dbc8ac/gni/v8.gni

I temporarily switched -O3 back to -O2 and collected the regressions in the internal issue 627583. There are several 3-4% regressions (we don't alert under 3%, so there might be more).

I think the speed win compensates for the sizes regression. Though I don't know if we have a formula how much speed conterbalaces how much space in general.

Setting therefore to wontfix. Please reopen if you think this is unjustified.