gpu: Add support for gpu_perftests for android.

gpu: Add support for gpu_perftests for android.

This cl adds a gpu_perftests_apk target to build an apk for android.
It also adds a check that detects if the timings are valid and discards them if they're not.
The number of valid runs (where we could get timings) is printed out along the other perf results as "sample_runs".

BUG=423481
Committed: https://crrev.com/7396d547d657957fe50469cfbb08619baed406b8
Cr-Commit-Position: refs/heads/master@{#317756}

[Daniele Castagna, 2015-02-21 00:11:58]

dcastagna@chromium.org changed reviewers:
+ reveman@chromium.org

[reveman, 2015-02-23 16:48:29]

lgtm with nits

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
File gpu/perftests/texture_upload_perftest.cc (right):

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:271: if (i >= kUploadPerfWarmupRuns) {
nit: maybe "if (i < kUpload...) continue" to reduce levels of nesting and make
the code a bit easier to read..

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:272: if (run.size() > 0) {
Nit: if (!run.size()) continue;

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:274: for (const Measurement& m : run) {
nit: s/m/measurement/

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:275: auto& agg = aggregates[m.name];
nit: s/agg/aggregate/

[Daniele Castagna, 2015-02-23 16:58:17]

New patchsets have been uploaded after l-g-t-m from reveman@chromium.org

[Daniele Castagna, 2015-02-23 16:59:25]

dcastagna@chromium.org changed reviewers:
+ piman@chromium.org

[Daniele Castagna, 2015-02-23 16:59:26]

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
File gpu/perftests/texture_upload_perftest.cc (right):

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:271: if (i >= kUploadPerfWarmupRuns) {
On 2015/02/23 at 16:48:29, reveman wrote:
> nit: maybe "if (i < kUpload...) continue" to reduce levels of nesting and make
the code a bit easier to read..

Done.

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:272: if (run.size() > 0) {
On 2015/02/23 at 16:48:29, reveman wrote:
> Nit: if (!run.size()) continue;

I moved this with the previous check.

Done.

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:274: for (const Measurement& m : run) {
On 2015/02/23 at 16:48:29, reveman wrote:
> nit: s/m/measurement/

Done.

https://codereview.chromium.org/944073005/diff/1/gpu/perftests/texture_upload...
gpu/perftests/texture_upload_perftest.cc:275: auto& agg = aggregates[m.name];
On 2015/02/23 at 16:48:29, reveman wrote:
> nit: s/agg/aggregate/

Done.

[piman, 2015-02-24 01:11:54]

lgtm

https://codereview.chromium.org/944073005/diff/60001/gpu/perftests/texture_up...
File gpu/perftests/texture_upload_perftest.cc (right):

https://codereview.chromium.org/944073005/diff/60001/gpu/perftests/texture_up...
gpu/perftests/texture_upload_perftest.cc:283: const auto m =
entry.second.Divide(successful_runs);
nit: may want to defend against successful_runs==0, to avoid divide-by-0.

[Daniele Castagna, 2015-02-24 02:42:01]

Thank you for the review.

https://codereview.chromium.org/944073005/diff/60001/gpu/perftests/texture_up...
File gpu/perftests/texture_upload_perftest.cc (right):

https://codereview.chromium.org/944073005/diff/60001/gpu/perftests/texture_up...
gpu/perftests/texture_upload_perftest.cc:283: const auto m =
entry.second.Divide(successful_runs);
On 2015/02/24 at 01:11:53, piman (Very slow to review) wrote:
> nit: may want to defend against successful_runs==0, to avoid divide-by-0.

Done.

[Daniele Castagna, 2015-02-24 02:43:30]

The CQ bit was checked by dcastagna@chromium.org

[Daniele Castagna, 2015-02-24 02:43:30]

The patchset sent to the CQ was uploaded after l-g-t-m from
reveman@chromium.org, piman@chromium.org
Link to the patchset: https://codereview.chromium.org/944073005/#ps80001 (title:
"Defend against divide by 0.")

[commit-bot: I haz the power, 2015-02-24 02:44:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/944073005/80001

[commit-bot: I haz the power, 2015-02-24 05:24:55]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2015-02-24 05:25:40]

Patchset 5 (id:??) landed as
https://crrev.com/7396d547d657957fe50469cfbb08619baed406b8
Cr-Commit-Position: refs/heads/master@{#317756}