Python benchmark library
Installation
pip install penchmark
pip install penchmark[charts]
Example
Classes Callee, InData
from penchmark import benchmark_and_print, Callee, InData
callees = (
Callee(callee_name='mul', callee=lambda x: x[0] * x[1]),
Callee(callee_name='nop', callee=lambda x: x)
)
dataset = (
InData(name='small-data', data=(2, 1), count_of_call=100000),
InData(name='big-data', data=(200, 10), count_of_call=1000),
InData(name='skipped-data', data=(1, 1), count_of_call=0)
)
benchmark_and_print(callees, dataset)
Tuples
from penchmark import benchmark_and_print
callees = (
('mul', lambda x: x[0] * x[1]),
('nop', lambda x: x)
)
dataset = (
('small-data', (2, 1), 100000),
('big-data', (200, 10), 1000),
('skipped-data', (1, 1), 0)
)
benchmark_and_print(callees, dataset)
Auto generate names of callees
from penchmark import benchmark_and_print
def mul(x): return x[0] * x[1]
def nop(x): return x
dataset = (
('small-data', (2, 1), 100000),
('big-data', (200, 10), 1000),
('skipped-data', (1, 1), 0)
)
benchmark_and_print((mul, nop), dataset)
Markdown result
small-data
| callee_name |
elapsed |
ratio |
| nop |
0.0050 |
1.0000 |
| mul |
0.0080 |
1.5842 |
big-data
| callee_name |
elapsed |
ratio |
| nop |
0.0000 |
1.0000 |
| mul |
0.0001 |
1.7201 |
Summary
| callee_name |
mean |
median |
| nop |
1.0000 |
1.0000 |
| mul |
1.6521 |
1.6521 |
Console mode result
...
benchmark_and_print((mul, nop), dataset, markdown=False)
SMALL-DATA
callee_name elapsed ratio
------------- --------- -------
nop 0.0050 1.0000
mul 0.0079 1.5944
BIG-DATA
callee_name elapsed ratio
------------- --------- -------
nop 0.0001 1.0000
mul 0.0001 1.7565
SUMMARY
callee_name mean median
------------- ------ --------
nop 1.0000 1.0000
mul 1.6754 1.6754
Callee with exceptions
from penchmark import benchmark_and_print
def callee_with_exceptions(x):
if not x:
raise Exception()
callees = (
('callee-with-exceptions', callee_with_exceptions),
('callee-without-exceptions', lambda x: None)
)
dataset = (
('valid-data', True, 10),
('invalid-data', False, 10),
)
benchmark_and_print(callees, dataset)
valid-data
| callee_name |
elapsed |
ratio |
| callee-without-exceptions |
0.00000 |
1 |
| callee-with-exceptions |
0.00000 |
2.36735 |
invalid-data
| callee_name |
elapsed |
ratio |
| callee-without-exceptions |
0.00000 |
1 |
| callee-with-exceptions |
ERROR |
|
Summary
| callee_name |
mean |
median |
| callee-without-exceptions |
1 |
1 |
Expected using InData
from penchmark import benchmark_and_print, InData
def mul2_1(sequ): return [x * 2 for x in sequ]
def mul2_2(sequ): return [x + x for x in sequ]
dataset = (
InData(name='small-data', data=(2, 1), count_of_call=100000, expected=[4, 2]),
InData(name='big-data', data=(200, 10), count_of_call=1000, expected=[400, 20]),
InData(name='skipped-data', data=(1, 1), count_of_call=0, expected=[2, 2])
)
benchmark_and_print((mul2_1, mul2_2), dataset)
Expected using tuples
from penchmark import benchmark_and_print
def mul2_1(sequ): return [x * 2 for x in sequ]
def mul2_2(sequ): return [x + x for x in sequ]
dataset = (
('small-data', (2, 1), 100000, [4, 2]),
('big-data', (200, 10), 1000, [400, 20]),
('skipped-data', (1, 1), 0, [2, 2])
)
benchmark_and_print((mul2_1, mul2_2), dataset)
small-data
| callee_name |
elapsed |
ratio |
| mul2_2 |
0.02114 |
1 |
| mul2_1 |
0.02317 |
1.09603 |
big-data
| callee_name |
elapsed |
ratio |
| mul2_2 |
0.00021 |
1 |
| mul2_1 |
0.00022 |
1.01147 |
Summary
| callee_name |
mean |
median |
| mul2_2 |
1 |
1 |
| mul2_1 |
1.05375 |
1.05375 |
