Add some statistics to the monsoon profile run

Add some statistics to the monsoon profile run

Add some statistics to the monsoon profile run that are output on the command
line.  The statistics include the total power, the average power += variance, the
peak power and the amount of time that was measured.

BUG=None

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=256926

[atreat, 2014-03-03 22:07:59]

Adding a few useful statistics.

[dtu, 2014-03-03 23:24:47]

https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
File tools/telemetry/telemetry/core/platform/profiler/monsoon.py (left):

https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
tools/telemetry/telemetry/core/platform/profiler/monsoon.py:213:
out.append((sample, main_voltage_v))
I've got another patch for this: https://codereview.chromium.org/185643005

https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
File tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py
(right):

https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py:8: from
scipy.integrate import simps, trapz
Sorry, I don't think we have numpy or scipy. I'm hesitant to take on this
dependency.

[atreat, 2014-03-05 19:54:26]

On 2014/03/03 23:24:47, dtu wrote:
>
https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
> File tools/telemetry/telemetry/core/platform/profiler/monsoon.py (left):
> 
>
https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
> tools/telemetry/telemetry/core/platform/profiler/monsoon.py:213:
> out.append((sample, main_voltage_v))
> I've got another patch for this: https://codereview.chromium.org/185643005
> 
>
https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
> File tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py
> (right):
> 
>
https://chromiumcodereview.appspot.com/185953004/diff/20001/tools/telemetry/t...
> tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py:8: from
> scipy.integrate import simps, trapz
> Sorry, I don't think we have numpy or scipy. I'm hesitant to take on this
> dependency.

Should I calculate this all by hand or is there another python library I should
be using?

[tonyg, 2014-03-08 17:30:03]

https://codereview.chromium.org/185953004/diff/20001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/profiler/monsoon.py (left):

https://codereview.chromium.org/185953004/diff/20001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiler/monsoon.py:213:
out.append((sample, main_voltage_v))
On 2014/03/03 23:24:48, dtu wrote:
> I've got another patch for this: https://codereview.chromium.org/185643005

If you rebase, the changes in this file should no longer be needed.

https://codereview.chromium.org/185953004/diff/20001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py
(right):

https://codereview.chromium.org/185953004/diff/20001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py:8: from
scipy.integrate import simps, trapz
On 2014/03/03 23:24:48, dtu wrote:
> Sorry, I don't think we have numpy or scipy. I'm hesitant to take on this
> dependency.

Looks like you are just calling pretty straightforward methods like average, std
and amax.

We should either have equivalents, or should implement equivalents here:
https://code.google.com/p/chromium/codesearch#chromium/src/tools/perf/metrics...

The catch is that tools/telemetry cannot depend on tools/perf. Dave may have
opinions here, but I'd be in favor of moving statistics.py into
tools/telemetry/telemetry/util/statistics.py . I bet there are other cases in
Telemetry results where we've re-implemented some of these things.

[atreat, 2014-03-11 18:06:15]

On 2014/03/08 17:30:03, tonyg wrote:
> The catch is that tools/telemetry cannot depend on tools/perf. Dave may have
> opinions here, but I'd be in favor of moving statistics.py into
> tools/telemetry/telemetry/util/statistics.py . I bet there are other cases in
> Telemetry results where we've re-implemented some of these things.

I've done as you suggested and wrote my own statistics functions that were
missing.

[atreat, 2014-03-12 14:39:58]

On 2014/03/11 18:06:15, atreat wrote:
> On 2014/03/08 17:30:03, tonyg wrote:
> > The catch is that tools/telemetry cannot depend on tools/perf. Dave may have
> > opinions here, but I'd be in favor of moving statistics.py into
> > tools/telemetry/telemetry/util/statistics.py . I bet there are other cases
in
> > Telemetry results where we've re-implemented some of these things.
> 
> I've done as you suggested and wrote my own statistics functions that were
> missing.

Can someone review?

[tonyg, 2014-03-12 15:28:38]

This is much better, thank you.

Just some minor style nits and then this should be ready to go.

https://codereview.chromium.org/185953004/diff/40001/tools/perf/metrics/stati...
File tools/perf/metrics/statistics_unittest.py (right):

https://codereview.chromium.org/185953004/diff/40001/tools/perf/metrics/stati...
tools/perf/metrics/statistics_unittest.py:1: # Copyright 2013 The Chromium
Authors. All rights reserved.
This file needs to move along with statistics.py.

Also, please add some basic tests for the new methods you've added. Be sure to
include the edge cases like 0, empty, etc.

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py
(right):

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py:54:
power.append(sample.amps * sample.volts)
Neat pythonism:

power_samples = [s.amps * s.volts for s in samples]

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/core/platform/profiler/monsoon_profiler.py:56: print
('Power total=%f' % statistics.TrapezoidalRule(power, 1/5000.) +
I'm wondering whether the labels would be better as:

Energy consumption=...
Average power=...
Peak power=...
Profile duration=...

Also, recommend line breaks between each.

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
File tools/telemetry/telemetry/util/statistics.py (right):

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/util/statistics.py:165: Both numerator and denominator
can be given as either individual
Nit: update the bit about the denominator.

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/util/statistics.py:170: denominator: A quantity that
represents a count of the number of things.
Nit: remove this line

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/util/statistics.py:180: def StandardDeviation(data):
Even though these are really common concepts, I recommend pydoc style comments
like we have in ArithmeticMean().

https://codereview.chromium.org/185953004/diff/40001/tools/telemetry/telemetr...
tools/telemetry/telemetry/util/statistics.py:206: """
The content of this comment is perfect, but please format it according to:
http://google-styleguide.googlecode.com/svn/trunk/pyguide.html#Comments

[atreat, 2014-03-12 19:43:21]

On 2014/03/12 15:28:38, tonyg wrote:
> This is much better, thank you.
> 
> Just some minor style nits and then this should be ready to go.

Thanks for the review!  Sorry I wasn't able to get to this till this afternoon. 
Lunch+meetings :P

[tonyg, 2014-03-12 19:56:35]

lgtm

[atreat, 2014-03-12 20:11:32]

The CQ bit was checked by adam.treat@samsung.com

[commit-bot: I haz the power, 2014-03-12 20:15:40]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/adam.treat@samsung.com/185953004/60001

[commit-bot: I haz the power, 2014-03-13 02:30:09]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-13 02:30:09]

Retried try job too often on win for step(s) compile
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win&number...

[tonyg, 2014-03-13 13:50:34]

The CQ bit was checked by tonyg@chromium.org

[commit-bot: I haz the power, 2014-03-13 13:51:20]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/adam.treat@samsung.com/185953004/60001

[commit-bot: I haz the power, 2014-03-13 13:51:29]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-13 13:51:30]

Failed to apply patch for tools/perf/metrics/smoothness.py:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  patching file tools/perf/metrics/smoothness.py
  Hunk #1 FAILED at 4.
  Hunk #2 FAILED at 76.
  Hunk #3 FAILED at 87.
  Hunk #4 FAILED at 98.
  Hunk #5 succeeded at 70 (offset -42 lines).
  4 out of 5 hunks FAILED -- saving rejects to file
tools/perf/metrics/smoothness.py.rej

Patch:       tools/perf/metrics/smoothness.py
Index: tools/perf/metrics/smoothness.py
diff --git a/tools/perf/metrics/smoothness.py b/tools/perf/metrics/smoothness.py
index
4041900847f315f1e93df7bd235774371028c167..d638da917e3f2ed2250b8503b63671347404e407
100644
--- a/tools/perf/metrics/smoothness.py
+++ b/tools/perf/metrics/smoothness.py
@@ -4,10 +4,10 @@
 
 from metrics import Metric
 from metrics import rendering_stats
-from metrics import statistics
 from telemetry.page import page_measurement
 from telemetry.page.perf_tests_helper import FlattenList
 from telemetry.core.timeline.model import TimelineModel
+from telemetry.util import statistics
 
 TIMELINE_MARKER = 'Smoothness'
 
@@ -76,8 +76,7 @@ class SmoothnessMetric(Metric):
   def AddResults(self, tab, results):
     if self._stats.mouse_wheel_scroll_latency:
       mean_mouse_wheel_scroll_latency = statistics.ArithmeticMean(
-          self._stats.mouse_wheel_scroll_latency,
-          len(self._stats.mouse_wheel_scroll_latency))
+          self._stats.mouse_wheel_scroll_latency)
       mouse_wheel_scroll_latency_discrepancy = statistics.DurationsDiscrepancy(
           self._stats.mouse_wheel_scroll_latency)
       results.Add('mean_mouse_wheel_scroll_latency', 'ms',
@@ -87,8 +86,7 @@ class SmoothnessMetric(Metric):
 
     if self._stats.touch_scroll_latency:
       mean_touch_scroll_latency = statistics.ArithmeticMean(
-          self._stats.touch_scroll_latency,
-          len(self._stats.touch_scroll_latency))
+          self._stats.touch_scroll_latency)
       touch_scroll_latency_discrepancy = statistics.DurationsDiscrepancy(
           self._stats.touch_scroll_latency)
       results.Add('mean_touch_scroll_latency', 'ms',
@@ -98,8 +96,7 @@ class SmoothnessMetric(Metric):
 
     if self._stats.js_touch_scroll_latency:
       mean_js_touch_scroll_latency = statistics.ArithmeticMean(
-          self._stats.js_touch_scroll_latency,
-          len(self._stats.js_touch_scroll_latency))
+          self._stats.js_touch_scroll_latency)
       js_touch_scroll_latency_discrepancy = statistics.DurationsDiscrepancy(
           self._stats.js_touch_scroll_latency)
       results.Add('mean_js_touch_scroll_latency', 'ms',
@@ -112,8 +109,7 @@ class SmoothnessMetric(Metric):
     results.Add('frame_times', 'ms', frame_times)
 
     # Arithmetic mean of frame times.
-    mean_frame_time = statistics.ArithmeticMean(frame_times,
-                                                len(frame_times))
+    mean_frame_time = statistics.ArithmeticMean(frame_times)
     results.Add('mean_frame_time', 'ms', round(mean_frame_time, 3))
 
     # Absolute discrepancy of frame time stamps.

[atreat, 2014-03-13 15:02:58]

The CQ bit was checked by adam.treat@samsung.com

[commit-bot: I haz the power, 2014-03-13 15:03:10]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/adam.treat@samsung.com/185953004/80001

[commit-bot: I haz the power, 2014-03-13 21:13:52]

Change committed as 256926