Add analysis scripts

Author: gjbex

source-code/mpi4py/analyze_mpifitness_data.py

@@ -0,0 +1,44 @@
+#!/usr/bin/env python
+
+from argparse import ArgumentParser, FileType
+import numpy as np
+import re
+import sys
+
+
+def accumulate(data_file):
+    regex = re.compile(r'^([a-z]+).+?:\s+(\d+(?:\.\d+)(?:e[+-]\d+)?)')
+    timings = dict()
+    for line in data_file:
+        match = regex.search(line)
+        if match:
+            test = match.group(1)
+            time = float(match.group(2))
+            if test not in timings:
+                timings[test] = list()
+            timings[test].append(time)
+    return timings
+
+
+def print_stats(test, data_list):
+    print(f'{test}:')
+    data = np.array(data_list)
+    print(f'    min:    {data.min()}')
+    print(f'    median: {np.median(data)}')
+    print(f'    mean:   {data.mean()}')
+    print(f'    max:    {data.max()}')
+    print(f'    n:      {len(data)}')
+
+
+def main():
+    arg_parser = ArgumentParser(description='analyze mpifitness data')
+    arg_parser.add_argument('file', type=FileType('r'), help='file to analyse')
+    options = arg_parser.parse_args()
+    timings = accumulate(options.file)
+    for test, data_list in timings.items():
+        print_stats(test, data_list)
+    return 0
+
+
+if __name__ == '__main__':
+    sys.exit(main())

source-code/mpi4py/plot_time_distribution.py

@@ -0,0 +1,32 @@
+#!/usr/bin/env python
+
+from argparse import ArgumentParser, FileType
+import matplotlib.pyplot as plt
+import seaborn as sns
+import sys
+from analyze_mpifitness_data import accumulate
+
+
+def main():
+    arg_parser = ArgumentParser(description='plot MPI time distirbution')
+    arg_parser.add_argument('--file', required=True, type=FileType('r'),
+                            help='file to plot data from')
+    arg_parser.add_argument('--test', required=True,
+                            choices=['pingpong', 'bcast', 'scatter', 'gather',
+                                     'alltoall', 'reduce'],
+                            help='test to visualize')
+    arg_parser.add_argument('--bins', type=int, default=5,
+                            help='number of bins in histogram')
+    arg_parser.add_argument('--rug', action='store_true', help='show rug')
+    options = arg_parser.parse_args()
+    timings = accumulate(options.file)
+    if options.rug:
+        sns.distplot(timings[options.test], rug=True, hist=False)
+    else:
+        sns.distplot(timings[options.test], bins=options.bins)
+    plt.show()
+    return 0
+
+
+if __name__ == '__main__':
+    sys.exit(main())