benchmarks/startup_mobile: scaling governor tweaks

Comparing benchmark results between experimental.startup.android.coldish
and experimental.startup.mobile revealed that the latter is a bit more
noisy (and reports slower startup). One difference visible in the log is
the lack of scaling governor adjustment. We do the scaling governor
adjustment for all other benchmarks I know of, hence correcting the
behavior of _MobileStartupSharedState to match it.

In my local testing this tweak did not bring the noise levels to match.
Looking further.

Bug: 760498
Change-Id: Ifb12f227a135d652fd068a28329887dcae35233a
Reviewed-on: https://chromium-review.googlesource.com/1178054Reviewed-by: Juan Antonio Navarro Pérez <perezju@chromium.org>
Reviewed-by: Ned Nguyen <nednguyen@google.com>
Commit-Queue: Egor Pasko <pasko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#584005}

tools/perf/benchmarks/startup_mobile.py

@@ -67,6 +67,7 @@ class _MobileStartupSharedState(story_module.SharedState):
     wpr_mode = wpr_modes.WPR_REPLAY
     if finder_options.use_live_sites:
       wpr_mode = wpr_modes.WPR_OFF
+    self.platform.SetFullPerformanceModeEnabled(True)
     self.platform.network_controller.Open(wpr_mode)
     self._story_set = story_set
 
@@ -76,6 +77,7 @@ class _MobileStartupSharedState(story_module.SharedState):
 
   def TearDownState(self):
     self.platform.network_controller.Close()
+    self.platform.SetFullPerformanceModeEnabled(False)
 
   def LaunchBrowser(self, url):
     self.platform.FlushDnsCache()