base/allocator: Hardcode TCMalloc debug optimization level to 2 on Linux.

base/allocator: Hardcode TCMalloc debug optimization level to 2 on Linux.

TCMalloc is very slow in debug mode. This may have a substantial improvement
on test cycle time. Theoretically, we shouldn't need to debug TCMalloc
very often.

BUG=388949
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=287226

[willchan no longer on Chromium, 2014-08-02 00:06:25]

Hey Nico, I'm thinking of landing this as an experiment, assuming it doesn't
break anything. But I'll probably end up reverting it. I'd like to examine the
test cycle times before and after. Does this look kosher for an experiment?

[Nico, 2014-08-02 00:10:49]

lgtm

https://codereview.chromium.org/438963007/diff/1/base/allocator/allocator.gyp
File base/allocator/allocator.gyp (right):

https://codereview.chromium.org/438963007/diff/1/base/allocator/allocator.gyp...
base/allocator/allocator.gyp:309: 'debug_optimize': 2,
Add a comment along the lines of "this code gets called a lot and is debugged
rarely, so it should be fast by default" and point to the bug in the comment

[willchan no longer on Chromium, 2014-08-02 00:19:27]

Cool, thanks. I've updated the CL accordingly.

[DaleCurtis, 2014-08-02 00:54:31]

FYI: there are a lot more of these you can set, but I don't know if they're
helpful in this case:

https://code.google.com/p/chromium/codesearch#chromium/src/third_party/ffmpeg...

[willchan no longer on Chromium, 2014-08-02 18:36:35]

The CQ bit was checked by willchan@chromium.org

[commit-bot: I haz the power, 2014-08-02 18:37:38]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/willchan@chromium.org/438963007/20001

[willchan no longer on Chromium, 2014-08-02 18:40:01]

Thanks Dale. If this experiment goes well, I'll followup with a CL to do it in
more cases.

[commit-bot: I haz the power, 2014-08-02 22:21:58]

Change committed as 287226