Skip to content

Commit

Permalink
Fix bad spelling of work default used as default runid
Browse files Browse the repository at this point in the history 
Add new column to output report, insts / sec. This provides
more info than just the absolute times. Changed the
act size column to refelect the avg real size of the xml
instances and added column for overall size of xml.

Add output at start of run that identifies params of run. This
was just because I do not like a program that gives no indication
that it started.

Remove the global XML_DICTIONARY. Remove the variable and move
the creation to just before we execute. The only reason for the
dictionary was to remove things from the code profiled before
we isolated the profile code to just the test code. It was just
wasting memory building multiple xmls before the tests.
  • Loading branch information
@KSchopmeyer @andy-maier
KSchopmeyer authored and andy-maier committed Mar 2, 2019
1 parent a2bdbdd commit 1fa4b4e
Showing 1 changed file with 42 additions and 45 deletions.
87 changes: 42 additions & 45 deletions tests/manualtest/run_response_performance.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,9 +29,8 @@
DEFAULT_RESPONSE_SIZE = [100, 1000]
DEFAULT_RESPONSE_COUNT = [100, 1000]
DEFAULT_TOP_N_ROWS = 20
DEFAULT_RUNID = 'default'

# Dictionary
XML_DICTIONARY = {}

PROFILE_OUT_PREFIX = 'perf'
PROFILE_DUMP_SUFFIX = 'profile'
Expand Down Expand Up @@ -110,7 +109,7 @@ class _PywbemCustomFormatter(_SmartFormatter,
https://stackoverflow.com/questions/18462610/argumentparser-epilog-and-description-formatting-in-conjunction-with-argumentdef
"""
pass
pass # pylint: disable=unnecessary-pass


TEMPLATE = u"""\n<VALUE.NAMEDINSTANCE>
Expand Down Expand Up @@ -156,6 +155,9 @@ def build_xml_obj(number, size):
return TEMPLATE.format(number, pattern, s1_value, a1_value)


AVG_RESPONSE_SIZE = 0


def create_xml_objs(count, size):
"""
Create the set of CIMInstances XML that will go into the test XML
Expand All @@ -166,9 +168,8 @@ def create_xml_objs(count, size):
obj = build_xml_obj(number, size)
obj_size.append(len(obj))
objstr += obj

global AVG_OBJ_SIZE
AVG_OBJ_SIZE = sum(obj_size) / len(obj_size)
global AVG_RESPONSE_SIZE
AVG_RESPONSE_SIZE = sum(obj_size) / len(obj_size)

return objstr

Expand All @@ -186,12 +187,14 @@ def create_xml(count=1000, size=100):

xml += '\n</IRETURNVALUE></IMETHODRESPONSE></SIMPLERSP></MESSAGE></CIM>'

return xml
return xml, AVG_RESPONSE_SIZE


def execute_test_code(xml_string, profiler):
"""
The test code to be executed.
The test code to be executed. If a profiler is defined it is enabled
just before the test code is executed and disabled just after the
code is executed.
"""
if profiler:
if isinstance(profiler, cProfile.Profile):
Expand All @@ -216,7 +219,7 @@ def execute_with_time(xml_string, profiler):
"""
Start time measurement and execute the test code.
desc reserved for future tests.
Returns the execution time
"""
start_time = time.time()

Expand All @@ -238,7 +241,7 @@ def __init__(self, args):
self.profiler = args.profiler
self.verbose = args.verbose
self.log = args.log
self.runid = args.runid or 'default'
self.runid = args.runid or DEFAULT_RUNID
# insure no blanks in the text
self.runid.replace(" ", "-")

Expand Down Expand Up @@ -266,6 +269,7 @@ def __init__(self, args):
# number of lines to display for cprofile output
self.top_n_rows = args.top_n_rows

# The following describes the columns in the cprofile output.
# ncalls - how many times the function/method has been called (in case
# the same function/method is being called recursively then ncalls has
# two values eg. 120/20, where the first is the true number of calls,
Expand Down Expand Up @@ -300,21 +304,21 @@ def __repr__(self):

def execute_raw_tests(self, profiler=None):
"""
Execute the parse test for all of the input parameters defined in
args. This allows multiple tests to be executed
We want to reduce this to the minimum code since everything here is
profiled
Execute the parse test for all of the input parameters defined in in
self.response_count and self.response_size. This allows multiple tests
to be executed
"""
table_rows = []
for response_size in self.response_size:
for response_count in self.response_count:
key = key = "%s:%s" % (response_count, response_size)
xml = XML_DICTIONARY[key]
xml, avg_resp_size = create_xml(response_count, response_size)
execution_time = execute_with_time(xml, profiler=profiler)
row = (response_size,
int(len(xml) / response_count),
int(avg_resp_size),
response_count,
execution_time)
len(xml),
"{0:5.2f}".format(execution_time),
int(response_count / execution_time))
table_rows.append(row)
return table_rows

Expand All @@ -334,23 +338,21 @@ def execute_cprofile_tests(self):

try:
ps.dump_stats(self.dumpfilename)
except Exception as ex:
except Exception as ex: # pylint: disable=broad-except
print('cProfiler dump stats exception %s %s. Ignored' %
(ex.__class__.__name__, ex))

# modify stats for display
# modify stats for display and output to stdout
ps.strip_dirs()
ps.sort_stats(*self.cprofilesort)

# output profile stats to stdout
ps.print_stats(self.top_n_rows)

# if dest defined, output to file defined by dest
if self.logfile:
with open(self.logfile, 'w') as stream:
ps = pstats.Stats(profiler, stream=stream)
ps.strip_dirs()
ps.sort_stats("tottime", "ncalls")
ps.sort_stats(*self.cprofilesort)
ps.print_stats(self.top_n_rows)

return table_rows
Expand All @@ -363,12 +365,9 @@ def execute_pyinstrument_tests(self):
"""
table_rows = []
profiler = Profiler()
# The start and stop were moved to the execution code
# profiler.start()

table_rows = self.execute_raw_tests(profiler)

# profiler.stop()
_uprint(None, profiler.output_text(unicode=True, color=True))
if self.logfile:
_uprint(self.logfile, profiler.output_text(unicode=True,
Expand All @@ -378,16 +377,6 @@ def execute_pyinstrument_tests(self):
def execute_tests(self):
"""Execute the test associated with profiler input argument."""

# build a dictionary of xml responses for the test. The removes this
# from any possible profile tests.
global XML_DICTIONARY
XML_DICTIONARY = {}
for response_size in self.response_size:
for response_count in self.response_count:
xml = create_xml(response_count, response_size)
key = "%s:%s" % (response_count, response_size)
XML_DICTIONARY[key] = xml

if self.profiler == 'none':
table_rows = self.execute_raw_tests()
elif self.profiler == 'pyinst':
Expand All @@ -399,15 +388,18 @@ def execute_tests(self):
'Should never occur' % self.profiler)

# build and output results report
header = ["Exp Response\nSize Bytes",
"Act Response\nSize (Bytes)",
header = ["Exp inst\nsize",
"Act inst\nsize",
"Response\nCount",
"Parse time\nsec.)"]
"XML\nsize",
"Parse time\nsec.",
"Instances\nper sec"]
title = 'Results: profile={0}, response_counts={1},\n ' \
'response-sizes={2}, {3}'.format(self.profiler,
self.response_count,
self.response_size,
self.file_datetime)
'response-sizes={2}, runid={3} {4}'.format(self.profiler,
self.response_count,
self.response_size,
self.runid,
self.file_datetime)
table = tabulate(table_rows, header, tablefmt="grid")

# print statistics to terminal
Expand Down Expand Up @@ -559,10 +551,10 @@ def parse_args():
tests_arggroup.add_argument(
'-r', '--runid', dest='runid',
metavar='text', type=str,
action='store', default="defailt",
action='store', default=DEFAULT_RUNID,
help='R|A string that is concatenated into each filename\n'
'created by the script.\n'
'Default: %s' % "default")
'Default: %s' % DEFAULT_RUNID)

tests_arggroup.add_argument(
'-i', '--individual', dest='individual',
Expand Down Expand Up @@ -605,6 +597,11 @@ def main():
"""
args = parse_args()

print('Starting profiler={0}, response-count={1} response-size={2} '
'runid={3} log={4}'.format(args.profiler, args.response_count,
args.response_size, args.runid,
args.log))

if args.individual:
execute_individual_tests(args)
else:
Expand Down

0 comments on commit 1fa4b4e

Please sign in to comment.