1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
|
import sys
import numpy as np
import ipyparallel as ipp
def echo(delay=0):
def inner_echo(x, **kwargs):
import time
if delay:
time.sleep(delay)
return x
return inner_echo
def profile_many_empty_tasks(lview, n, block=True):
lview.map(echo(0), [None] * n, block=block)
# def profile_echo_many_arguments(lview, number_of_arguments):
# lview.map(
# lambda x: echo_many_arguments(*x),
# [
# tuple(np.empty(1, dtype=np.int8) for n in range(number_of_arguments))
# for x in range(16)
# ],
# block=False,
# )
def profile_tasks_with_large_data(lview, num_bytes):
for _ in range(10):
for i in range(10):
lview.apply_sync(echo(0), np.array([0] * num_bytes, dtype=np.int8))
def run_profiling(selected_profiling_task, selected_view):
client = ipp.Client(profile='asv')
# add it to path on the engines
# client[:].apply_sync(add_to_path, master_project_parent)
print('profiling task: ', selected_profiling_task)
print('profiling view: ', selected_view)
if selected_view == 'direct':
view = client[:]
elif selected_view == 'spanning_tree':
view = client.spanning_tree_view()
else:
view = client.load_balanced_view()
# view = client[:] if selected_view == 'direct' else client.load_balanced_view()
if selected_profiling_task == 'many_empty_tasks':
for x in range(1, 5):
profile_many_empty_tasks(view, 10**x)
elif selected_profiling_task == 'many_empty_tasks_non_blocking':
for x in range(1, 5):
profile_many_empty_tasks(view, 10**x, block=False)
elif selected_profiling_task == 'tasks_with_large_data':
for x in range(1, 8):
profile_tasks_with_large_data(view, 10**x)
# elif selected_profiling_task == 'echo_many_arguments':
# for i in range(100):
# for number_of_arguments in ((2 ** x) - 1 for x in range(1, 9)):
# profile_echo_many_arguments(view, number_of_arguments)
if __name__ == "__main__":
run_profiling(sys.argv[1], sys.argv[2])
|
