diff --git a/benchmark/benchmark.py b/benchmark/benchmark.py
old mode 100644
new mode 100755
index c44d88a..5667e36
--- a/benchmark/benchmark.py
+++ b/benchmark/benchmark.py
@@ -1,33 +1,59 @@
+#!/usr/bin/env python
 '''
 Created on 04.10.2019
 
 @author: gocht
 '''
+import argparse
 import sys
 import benchmark_helper
 import pickle
 
+# Available tests
 tests = ["bm_baseline.py", "bm_simplefunc.py"]
 
+# Available instrumenters
 instrumenters = ["profile", "trace", "dummy", "None"]
 if sys.version_info.major >= 3:
     instrumenters.extend(["cProfile", "cTrace"])
 
-# How many times the instrumented code is run during 1 test run
+# Default values for: How many times the instrumented code is run during 1 test run
 reps_x = {
     "bm_baseline.py": ["1000000", "2000000", "3000000", "4000000", "5000000"],
     "bm_simplefunc.py": ["100000", "200000", "300000", "400000", "500000"],
 }
-# How many times a test invocation is repeated (number of timings per test instance)
-test_repetitions = 51
 
-bench = benchmark_helper.BenchmarkEnv(repetitions=test_repetitions)
+
+def str_to_int(s):
+    return int(float(s))
+
+
+parser = argparse.ArgumentParser(description='Benchmark the instrumenters.',
+                                 formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+parser.add_argument('--test', '-t', metavar='TEST', nargs='+', default=tests,
+                    choices=tests, help='Which test(s) to run')
+parser.add_argument('--repetitions', '-r', default=51, type=str_to_int,
+                    help='How many times a test invocation is repeated (number of timings per test instance)')
+parser.add_argument('--loop-count', '-l', type=str_to_int, nargs='+',
+                    help=('How many times the instrumented code is run during 1 test run. '
+                          'Can be repeated and will create 1 test instance per argument'))
+parser.add_argument('--instrumenter', '-i', metavar='INST', nargs='+', default=instrumenters,
+                    choices=instrumenters, help='The instrumenter(s) to use')
+parser.add_argument('--output', '-o', default='results.pkl', help='Output file for the results')
+parser.add_argument('--dry-run', action='store_true', help='Print parsed arguments and exit')
+args = parser.parse_args()
+
+if args.dry_run:
+    print(args)
+    sys.exit(0)
+
+bench = benchmark_helper.BenchmarkEnv(repetitions=args.repetitions)
 results = {}
 
-for test in tests:
+for test in args.test:
     results[test] = {}
 
-    for instrumenter in instrumenters:
+    for instrumenter in args.instrumenter:
         results[test][instrumenter] = {}
 
         if instrumenter == "None":
@@ -40,12 +66,13 @@
         print("#########")
         print("{}: {}".format(test, scorep_settings))
         print("#########")
-        for reps in reps_x[test]:
-            times = bench.call(test, [reps],
+        loop_counts = args.loop_count if args.loop_count else reps_x[test]
+        for reps in loop_counts:
+            times = bench.call(test, [str(reps)],
                                enable_scorep,
                                scorep_settings=scorep_settings)
             print("{:<8}: {}".format(reps, times))
             results[test][instrumenter][reps] = times
 
-with open("results.pkl", "wb") as f:
+with open(args.output, "wb") as f:
     pickle.dump(results, f)