Refactor mojo_benchmark to get rid of warm/cold start special cases.

mojo/devtools/common/mojo_benchmark

Index: mojo/devtools/common/mojo_benchmark
diff --git a/mojo/devtools/common/mojo_benchmark b/mojo/devtools/common/mojo_benchmark
index e0ffff0a82925b1de2b96c44d4d6dfa27a57c624..b57d45e206baa61b34fa1aa10f3a36de38a299d0 100755
--- a/mojo/devtools/common/mojo_benchmark
+++ b/mojo/devtools/common/mojo_benchmark
@@ -50,22 +50,44 @@ _BENCHMARK_APP = 'https://core.mojoapps.io/benchmark.mojo'
 _CACHE_SERVICE_URL = 'mojo:url_response_disk_cache'
 _NETWORK_SERVICE_URL = 'mojo:network_service'
 
+_COLD_START_SHELL_ARGS = [
+  '--args-for=%s %s' % (_CACHE_SERVICE_URL, '--clear'),
+  '--args-for=%s %s' % (_NETWORK_SERVICE_URL, '--clear'),
+]
+
 # Additional time in seconds allocated per shell run to accommodate start-up.
 # The shell should terminate before hitting this time out, it is an error if it
 # doesn't.
 _EXTRA_TIMEOUT = 20
 
 
-def _get_output_file(shell, name, cold_start):
-    file_name = 'benchmark-%s-%s-%s.trace' % (
-        name.replace(' ', '_'),
-        'cold_start' if cold_start else 'warm_start',
-        time.strftime('%Y%m%d%H%M%S'))
-    return file_name
+def _rewrite_benchmark_list(benchmark_list):
+  """Rewrites the specification of benchmarks to run, yielding two benchmarks
+  (one for cold and one for warm start) for each benchmark defined in the input
+  list.
+  """
+  result = list()

[viettrungluu, 2015/10/14 21:24:16]

Probably "= []" would look more normal.

[ppi, 2015/10/14 21:47:27]

Done.

+  for benchmark_spec in benchmark_list:

[viettrungluu, 2015/10/14 21:24:16]

I wonder if you shouldn't have the function be: _generate_benchmark_variants(),
and do below:

benchmark_list = []
for benchmark in benchmark_list_params['benchmarks']:
  benchmark += _generate_benchmark_variants(benchmark)

[ppi, 2015/10/14 21:47:27]

Sg, done.

+    # Cold start.
+    result.append({
+      'name': benchmark_spec['name'] + ' (cold start)',
+      'app': benchmark_spec['app'],
+      'duration': benchmark_spec['duration'],
+      'measurements': benchmark_spec['measurements'],
+      'shell-args': benchmark_spec.get('shell-args',
+                                       []) + _COLD_START_SHELL_ARGS})
+    # Warm start.
+    result.append({
+      'name': benchmark_spec['name'] + ' (warm start)',
+      'app': benchmark_spec['app'],
+      'duration': benchmark_spec['duration'],
+      'measurements': benchmark_spec['measurements'],
+      'shell-args': benchmark_spec.get('shell-args', [])})
+  return result
 
 
 def _run_benchmark(shell, shell_args, name, app, duration_seconds, measurements,
-                   cold_start, verbose, android, save_traces):
+                   verbose, android, save_traces):
   """Runs `benchmark.mojo` in shell with correct arguments, parses and
   presents the benchmark results.
   """
@@ -77,7 +99,8 @@ def _run_benchmark(shell, shell_args, name, app, duration_seconds, measurements,
   output_file = None
   device_output_file = None
   if save_traces:
-    output_file = _get_output_file(shell, name, cold_start)
+    output_file = 'benchmark-%s-%s.trace' % (name.replace(' ', '_'),
+        time.strftime('%Y%m%d%H%M%S'))
     if android:
       device_output_file = os.path.join(shell.get_tmp_dir_path(), output_file)
       benchmark_args.append('--trace-output=' + device_output_file)
@@ -93,13 +116,9 @@ def _run_benchmark(shell, shell_args, name, app, duration_seconds, measurements,
   shell_args.append('--args-for=%s %s' % (_BENCHMARK_APP,
                                           ' '.join(benchmark_args)))
 
-  if cold_start:
-    shell_args.append('--args-for=%s %s' % (_CACHE_SERVICE_URL, '--clear'))
-    shell_args.append('--args-for=%s %s' % (_NETWORK_SERVICE_URL, '--clear'))
-
   if verbose:
     print 'shell arguments: ' + str(shell_args)
-  print '[ %s ] %s' % (name, 'cold start' if cold_start else 'warm start')
+  print '[ %s ]' % name
   return_code, output, did_time_out = shell.run_and_get_output(
       shell_args, timeout=timeout)
   output_lines = [line.strip() for line in output.split('\n')]
@@ -146,7 +165,7 @@ def main():
   target_os = 'android' if script_args.android else 'linux'
   benchmark_list_params = {"target_os": target_os}
   exec script_args.benchmark_list_file in benchmark_list_params
-  benchmark_list = benchmark_list_params['benchmarks']
+  benchmark_list = _rewrite_benchmark_list(benchmark_list_params['benchmarks'])
 
   succeeded = True
   for benchmark_spec in benchmark_list:
@@ -156,12 +175,7 @@ def main():
     shell_args = benchmark_spec.get('shell-args', []) + common_shell_args
     measurements = benchmark_spec['measurements']
     _run_benchmark(shell, shell_args, name, app, duration, measurements,
-                   cold_start=True, verbose=script_args.verbose,
-                   android=script_args.android,
-                   save_traces=script_args.save_traces)
-    _run_benchmark(shell, shell_args, name, app, duration, measurements,
-                   cold_start=False, verbose=script_args.verbose,
-                   android=script_args.android,
+                   verbose=script_args.verbose, android=script_args.android,

[viettrungluu, 2015/10/14 21:24:16]

I find it odd that you give these things as keyword arguments, when they're
required positional arguments (though this is a pre-existing defect).

[ppi, 2015/10/14 21:47:27]

Done.

                    save_traces=script_args.save_traces)
 
   return 0 if succeeded else 1