def main(): global task_groups global summaries parser = argparse.ArgumentParser() # usage: python process.py <logfiles> parser.add_argument(dest='filenames', nargs='+', help='input Legion Prof log filenames') args = parser.parse_args() deserializer = LegionProfASCIIDeserializer(Dummy(), callbacks) if args.filenames is None: print("You must pass in a logfile!") exit(-1) has_matches = False for file_name in args.filenames: matches = deserializer.parse(file_name, True) has_matches = has_matches or matches > 0 if not has_matches: print('No matches found! Exiting...') return # filter only the tasks of interest task_group_ids = set(task_groups.iterkeys()) for i in task_group_ids: if task_groups[i].variant_id != 2: del task_groups[i] for i, grp in task_groups.iteritems(): summaries.append(grp.get_summary()) # filter out the first and last 5% as they are mostly abnormal outliers = len(summaries) / 10 summaries = summaries[outliers:-outliers] print(itv_sum(summaries))
def main(): parser = argparse.ArgumentParser() # usage: python span_example.py <task0> <task1> <logfile0> <logfile1> parser.add_argument( "-t", "--task", nargs=1, type=str, help="The task name you want to find. These are interpreted as regexes" ) parser.add_argument("-a", "--ascii", action="store_true", dest="ascii_parser", help="Use ASCII parser.") parser.add_argument(dest='filenames', nargs='+', help='input Legion Prof log filenames') args = parser.parse_args() if args.ascii_parser: deserializer = LegionProfASCIIDeserializer(Dummy(), callbacks) else: deserializer = LegionProfBinaryDeserializer(None, callbacks) if args.filenames is None: print("You must pass in a logfile!") exit(-1) has_matches = False for file_name in args.filenames: matches = deserializer.parse(file_name, True) has_matches = has_matches or matches > 0 if not has_matches: print('No matches found! Exiting...') return # now search through the spans task = args.task[0] matching_tasks = [] for task_name in task_spans.iterkeys(): if re.search(task, task_name): matching_tasks.append(task_name) err = False if len(matching_tasks) == 0: print("No matching tasks for " + task + " found!") exit(-1) for task_name in matching_tasks: spans = task_spans[task_name] print("Task " + task_name + ":") for span in spans: print(str(span[1] - span[0]) + "us (start: " + \ str(span[0]) + "us, stop: " + str(span[1]) + "us)")
def main(): parser = argparse.ArgumentParser() # usage: python span_example.py <task0> <task1> <logfile0> <logfile1> parser.add_argument("-t", "--task", nargs=1, type=str, help="The task name you want to find. These are interpreted as regexes") parser.add_argument("-a", "--ascii", action="store_true", dest="ascii_parser", help="Use ASCII parser.") parser.add_argument(dest='filenames', nargs='+', help='input Legion Prof log filenames') args = parser.parse_args() if args.ascii_parser: deserializer = LegionProfASCIIDeserializer(Dummy(), callbacks) else: deserializer = LegionProfBinaryDeserializer(None, callbacks) if args.filenames is None: print("You must pass in a logfile!") exit(-1) has_matches = False for file_name in args.filenames: matches = deserializer.parse(file_name, True) has_matches = has_matches or matches > 0 if not has_matches: print('No matches found! Exiting...') return # now search through the spans task = args.task[0] matching_tasks = [] for task_name in task_spans.iterkeys(): if re.search(task, task_name): matching_tasks.append(task_name) err = False if len(matching_tasks) == 0: print("No matching tasks for " + task + " found!") exit(-1) for task_name in matching_tasks: spans = task_spans[task_name] print("Task " + task_name + ":") for span in spans: print(str(span[1] - span[0]) + "us (start: " + \ str(span[0]) + "us, stop: " + str(span[1]) + "us)")
def main(): parser = argparse.ArgumentParser() # usage: python span_example.py <task0> <task1> <logfile0> <logfile1> parser.add_argument("-t", "--tasks", nargs=2, type=str, help="The two task names you want to find the span between. These are interpreted as regexes") parser.add_argument("-s", "--skip", nargs=2, type=int, default=(0, 0), help="The number of task occurrences you want to prune in the span.") parser.add_argument("-a", "--ascii", action="store_true", dest="ascii_parser", help="Use ASCII parser.") parser.add_argument(dest='filenames', nargs='+', help='input Legion Prof log filenames') args = parser.parse_args() if args.ascii_parser: deserializer = LegionProfASCIIDeserializer(Dummy(), callbacks) else: deserializer = LegionProfBinaryDeserializer(None, callbacks) if args.filenames is None: print("You must pass in a logfile!") exit(-1) has_matches = False for file_name in args.filenames: matches = deserializer.parse(file_name, True) has_matches = has_matches or matches > 0 if not has_matches: print('No matches found! Exiting...') return # now search through the spans task0 = args.tasks[0] task1 = args.tasks[1] skip0 = args.skip[0] skip1 = args.skip[1] matching_tasks0 = [] matching_tasks1 = [] for task_name in task_spans.iterkeys(): if re.search(task0, task_name): matching_tasks0.append(task_name) if re.search(task1, task_name): matching_tasks1.append(task_name) err = False for (task, matches) in [(task0, matching_tasks0), (task1, matching_tasks1)]: if len(matches) == 0: print("No matching tasks for " + task + " found!") err = True if len(matches) > 1: print("Task " + task + " had disambiguous matches. Pick the task name you want from the following list:") print(matches) err = True if err: exit(-1) match0 = matching_tasks0[0] match1 = matching_tasks1[0] span0 = sorted(task_spans[match0], key=lambda r: r[0]) span1 = sorted(task_spans[match1], key=lambda r: r[0]) span0 = span0[skip0:] if skip1 > 0: span1 = span1[0:-skip1] start = min([s[0] for s in span0]) stop = max([s[1] for s in span1]) task_span = (start, stop) print("Range between " + task0 + " and " + task1 + " was " + \ str(task_span[1] - task_span[0]) + "us (start: " + \ str(task_span[0]) + "us, stop: " + str(task_span[1]) + "us)")