Fix CPU and memory perf PyAuto test running on av_perf.

Fix CPU perf PyAuto test (called ui_perf_test_measure_thread) running on av_perf.

This test measures CPU and memory performance of Chrome renderer process while playing a video file.

Since this test and fps perf tests run on same media files, we combined both tests into one.

BUG=115602
TEST=ran on local bot.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=126015

[DaleCurtis, 2012-02-24 17:02:24]

+scherkus since he reviewed the last one like this too.

LGTM, but have you considered instead merging this into the FPS test since they
use the same videos? Is there any reason we can't do that?

https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
File chrome/test/functional/media/media_cpu_perf.py (right):

https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
chrome/test/functional/media/media_cpu_perf.py:33: class
MediaFPSPerfTest(pyauto.PyUITest):
CPUTest?

[shadi, 2012-02-24 17:39:06]

On 2012/02/24 17:02:24, DaleCurtis wrote:
> +scherkus since he reviewed the last one like this too.
> 
> LGTM, but have you considered instead merging this into the FPS test since
they
> use the same videos? Is there any reason we can't do that?
> 
>
https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
> File chrome/test/functional/media/media_cpu_perf.py (right):
> 
>
https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
> chrome/test/functional/media/media_cpu_perf.py:33: class
> MediaFPSPerfTest(pyauto.PyUITest):
> CPUTest?

There is no reason why not to merge this with FPS, but I wouldn't use the same
fps_html file to measure cpu and memory to avoid any cpu/memory overhead in the
fps test.

So, they would be merged but we still need to play each file twice.

[DaleCurtis, 2012-02-24 17:44:43]

On 2012/02/24 17:39:06, shadi wrote:
> On 2012/02/24 17:02:24, DaleCurtis wrote:
> > +scherkus since he reviewed the last one like this too.
> > 
> > LGTM, but have you considered instead merging this into the FPS test since
> they
> > use the same videos? Is there any reason we can't do that?
> > 
> >
>
https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
> > File chrome/test/functional/media/media_cpu_perf.py (right):
> > 
> >
>
https://chromiumcodereview.appspot.com/9464004/diff/1005/chrome/test/function...
> > chrome/test/functional/media/media_cpu_perf.py:33: class
> > MediaFPSPerfTest(pyauto.PyUITest):
> > CPUTest?
> 
> There is no reason why not to merge this with FPS, but I wouldn't use the same
> fps_html file to measure cpu and memory to avoid any cpu/memory overhead in
the
> fps test.
> 
> So, they would be merged but we still need to play each file twice.

Why couldn't you use the same HTML file? The FPS one just plays and records FPS
data while this one just plays.

[shadi, 2012-02-24 21:24:47]

I combined cpu and fps tests together.

The fps CL is not yet in, so I cannot delete those files yet.

I will do that as well as deleting old bot tests in a separate CL.

[shadi, 2012-02-29 21:51:37]

Dale,

I fixed the rebase. After getting LGTM I will rename the files before the final
commit.

[DaleCurtis, 2012-02-29 22:39:11]

http://codereview.chromium.org/9464004/diff/9001/chrome/test/data/media/html/...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/9001/chrome/test/data/media/html/...
chrome/test/data/media/html/media_fps_perf.html:6: <title>CPU, Memomry, and FPS
Perf Test</title>
Memory

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:6: """CPU, Memomry, and FPS
performance test for <video>.
Memory

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:51: We do not calculate the CPU
and memory stats while running the FPS test to
Did you observe differences that made you decide this was necessary? Otherwise
it just seems to complicate the code.

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:66:
self.CallJavascriptFunc('startTest', [file_url, True], tab_index=1))
I thought tab_index wasn't consistent, does this always work? Ditto everywhere
you use tab_index=1 below.

[shadi, 2012-03-01 02:40:53]

Dale can you please review?

http://codereview.chromium.org/9464004/diff/9001/chrome/test/data/media/html/...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/9001/chrome/test/data/media/html/...
chrome/test/data/media/html/media_fps_perf.html:6: <title>CPU, Memomry, and FPS
Perf Test</title>
On 2012/02/29 22:39:11, DaleCurtis wrote:
> Memory

Done.

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:6: """CPU, Memomry, and FPS
performance test for <video>.
On 2012/02/29 22:39:11, DaleCurtis wrote:
> Memory

Done.

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:51: We do not calculate the CPU
and memory stats while running the FPS test to
After offline discussion, we decided that even though there is a small overhead
(5~8MB) with 2160p tests, the numbers are consistent and is not worth running
separate tests to avoid this.
On 2012/02/29 22:39:11, DaleCurtis wrote:
> Did you observe differences that made you decide this was necessary? Otherwise
> it just seems to complicate the code.

http://codereview.chromium.org/9464004/diff/9001/chrome/test/functional/media...
chrome/test/functional/media/media_fps_perf.py:66:
self.CallJavascriptFunc('startTest', [file_url, True], tab_index=1))
It appears to work. :-)
On 2012/02/29 22:39:11, DaleCurtis wrote:
> I thought tab_index wasn't consistent, does this always work? Ditto everywhere
> you use tab_index=1 below.

[DaleCurtis, 2012-03-01 02:47:10]

LGTM % nits.

http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
chrome/test/data/media/html/media_fps_perf.html:18: // Set to true to record FPS
stats.
Is this necessary anymore?

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:50:
os.path.join(_TEST_MEDIA_PATH, '%s' % file_name))
Just file_name without %

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:73:
pyauto_utils.PrintPerfResult('fps', file_name, decoded_fps, 'fps')
Will we need to clean up the old graphs on the bot or is it smart enough to take
care of that after this rename?

[shadi, 2012-03-01 19:17:33]

Scherkus, can you please take a final look?

http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
chrome/test/data/media/html/media_fps_perf.html:18: // Set to true to record FPS
stats.
On 2012/03/01 02:47:11, DaleCurtis wrote:
> Is this necessary anymore?

Done.

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:50:
os.path.join(_TEST_MEDIA_PATH, '%s' % file_name))
On 2012/03/01 02:47:11, DaleCurtis wrote:
> Just file_name without %

Done.

http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:73:
pyauto_utils.PrintPerfResult('fps', file_name, decoded_fps, 'fps')
On 2012/03/01 02:47:11, DaleCurtis wrote:
> Will we need to clean up the old graphs on the bot or is it smart enough to
take
> care of that after this rename?
I think they are automatically cleaned up after when no data is available to
display. Anyway, I will verify that.

[DaleCurtis, 2012-03-01 21:12:49]

I just noticed media_fps_perf was missing the executable bit in my local
checkout, can you set +x before you commit?

On 2012/03/01 19:17:33, shadi wrote:
> Scherkus, can you please take a final look?
> 
>
http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
> File chrome/test/data/media/html/media_fps_perf.html (right):
> 
>
http://codereview.chromium.org/9464004/diff/15001/chrome/test/data/media/html...
> chrome/test/data/media/html/media_fps_perf.html:18: // Set to true to record
FPS
> stats.
> On 2012/03/01 02:47:11, DaleCurtis wrote:
> > Is this necessary anymore?
> 
> Done.
> 
>
http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
> File chrome/test/functional/media/media_fps_perf.py (right):
> 
>
http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
> chrome/test/functional/media/media_fps_perf.py:50:
> os.path.join(_TEST_MEDIA_PATH, '%s' % file_name))
> On 2012/03/01 02:47:11, DaleCurtis wrote:
> > Just file_name without %
> 
> Done.
> 
>
http://codereview.chromium.org/9464004/diff/15001/chrome/test/functional/medi...
> chrome/test/functional/media/media_fps_perf.py:73:
> pyauto_utils.PrintPerfResult('fps', file_name, decoded_fps, 'fps')
> On 2012/03/01 02:47:11, DaleCurtis wrote:
> > Will we need to clean up the old graphs on the bot or is it smart enough to
> take
> > care of that after this rename?
> I think they are automatically cleaned up after when no data is available to
> display. Anyway, I will verify that.

[scherkus (not reviewing), 2012-03-03 01:34:16]

LGTM w/ nits

http://codereview.chromium.org/9464004/diff/17001/chrome/test/data/media/html...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/17001/chrome/test/data/media/html...
chrome/test/data/media/html/media_fps_perf.html:1: <!-- Used by media_stat_perf
to calculate <video> performance statistics. -->
is this supposed to match the python name? (i.e., media_fps_perf.py)

http://codereview.chromium.org/9464004/diff/17001/chrome/test/functional/medi...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/17001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:61: mem_usage =
renderer_process.get_memory_info()[0] / 1024
what unit is this in?

perhaps affix the units? mem_usage -> mem_usage_mb, etc.

[shadi, 2012-03-08 21:52:28]

http://codereview.chromium.org/9464004/diff/17001/chrome/test/data/media/html...
File chrome/test/data/media/html/media_fps_perf.html (right):

http://codereview.chromium.org/9464004/diff/17001/chrome/test/data/media/html...
chrome/test/data/media/html/media_fps_perf.html:1: <!-- Used by media_stat_perf
to calculate <video> performance statistics. -->
Yes, it does. However, the file names will change since it is not just fps
tests.

I didn't want to change them now so that I can rebase.
On 2012/03/03 01:34:16, scherkus wrote:
> is this supposed to match the python name? (i.e., media_fps_perf.py)

http://codereview.chromium.org/9464004/diff/17001/chrome/test/functional/medi...
File chrome/test/functional/media/media_fps_perf.py (right):

http://codereview.chromium.org/9464004/diff/17001/chrome/test/functional/medi...
chrome/test/functional/media/media_fps_perf.py:61: mem_usage =
renderer_process.get_memory_info()[0] / 1024
On 2012/03/03 01:34:16, scherkus wrote:
> what unit is this in?
> 
> perhaps affix the units? mem_usage -> mem_usage_mb, etc.

Done.

[commit-bot: I haz the power, 2012-03-08 21:52:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shadi@chromium.org/9464004/26001

[commit-bot: I haz the power, 2012-03-09 00:43:40]

Try job failure for 9464004-26001 (retry) on win_rel for step "browser_tests".
It's a second try, previously, step "browser_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2012-03-09 18:54:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shadi@chromium.org/9464004/26001

[commit-bot: I haz the power, 2012-03-09 21:15:14]

Try job failure for 9464004-26001 (retry) on win_rel for step "browser_tests".
It's a second try, previously, step "browser_tests" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2012-03-09 22:30:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shadi@chromium.org/9464004/26001

[commit-bot: I haz the power, 2012-03-10 01:22:07]

Try job failure for 9464004-26001 (retry) on win_rel for step "browser_tests".
It's a second try, previously, steps "browser_tests, installer_util_unittests"
failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2012-03-10 04:16:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/shadi@chromium.org/9464004/26001

[commit-bot: I haz the power, 2012-03-10 07:09:37]

Change committed as 126015