Skip to content

Instantly share code, notes, and snippets.

@Kento75

Kento75/file1.txt

Last active September 11, 2019 09:08
Show Gist options
  • Save Kento75/2a4490a478dbb19c2d16e8665d7c4d88 to your computer and use it in GitHub Desktop.
Save Kento75/2a4490a478dbb19c2d16e8665d7c4d88 to your computer and use it in GitHub Desktop.
Download ZIP
【Pythonメモ】モジュールのプロファイリング ref: https://qiita.com/Kento75/items/8b72d64884d7072eece2
Raw
file1.txt
$ python -m cProfile -s cumulative julia_set.py
Length of x: 1000
Total elements: 1000000
calculate_z_serial_purepython took 23.838048934936523 seconds
36221995 function calls in 25.485 seconds
Ordered by: cumulative time
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.000 0.000 25.485 25.485 {built-in method builtins.exec}
1 0.060 0.060 25.485 25.485 julia_set.py:1(<module>)
1 1.270 1.270 25.425 25.425 julia_set.py:8(calc_pure_python)
1 16.907 16.907 23.839 23.839 julia_set.py:44(calculate_z_serial_purepython)
34219980 6.932 0.000 6.932 0.000 {built-in method builtins.abs}
2002000 0.300 0.000 0.300 0.000 {method 'append' of 'list' objects}
1 0.015 0.015 0.015 0.015 {built-in method builtins.sum}
3 0.001 0.000 0.001 0.000 {built-in method builtins.print}
2 0.000 0.000 0.000 0.000 {built-in method time.time}
4 0.000 0.000 0.000 0.000 {built-in method builtins.len}
1 0.000 0.000 0.000 0.000 {method 'disable' of '_lsprof.Profiler' objects}
（以下略）
Raw
file10.txt
$ python line_profiler/kernprof.py -l -v julia_set.py
Length of x: 1000
Total elements: 1000000
calculate_z_serial_purepython took 277.837988615036 seconds
Wrote profile results to julia_set.py.lprof
Timer unit: 3.9474e-07 s
Total time: 173.619 s
File: julia_set.py
Function: calculate_z_serial_purepython at line 45
Line # Hits Time Per Hit % Time Line Contents
==============================================================
45 @profile
46 def calculate_z_serial_purepython(maxiter, zs, cs):
47 """
48 ジュリア漸化式を使用して output リストを計算する
49 """
50 1 18915.0 18915.0 0.0 output = [0] * len(zs)
51 1000001 2707922.0 2.7 0.6 for i in range(len(zs)):
52 1000000 2632718.0 2.6 0.6 n = 0
53 1000000 3088312.0 3.1 0.7 z = zs[i]
54 1000000 2982926.0 3.0 0.7 c = cs[i]
55 34219980 177968268.0 5.2 40.5 while abs(z) < 2 and n < maxiter:
56 33219980 142884632.0 4.3 32.5 z = z * z + c
57 33219980 104282683.0 3.1 23.7 n += 1
58
59 1000000 3264717.0 3.3 0.7 output[i] = n
60
61 1 5.0 5.0 0.0 return output
Raw
file2.txt
# profile.stats → ファイル名
$ python -m cProfile -o profile.stats julia_set.py
Raw
file3.txt
$ python
>>> import pstats
>>> p = pstats.Stats('profile.stats')
>>> p.sort_stats('cumulative')
>>> <pstats.Stats object at 0x00000299728D9278>
>>> p.print_stats()
Mon Jul 16 08:44:00 2018 profile.stats
36221995 function calls in 29.504 seconds
Ordered by: cumulative time
ncalls tottime percall cumtime percall filename:lineno(function)
1 0.000 0.000 29.504 29.504 {built-in method builtins.exec}
1 0.062 0.062 29.504 29.504 julia_set.py:1(<module>)
1 1.272 1.272 29.442 29.442 julia_set.py:8(calc_pure_python)
1 19.520 19.520 27.850 27.850 julia_set.py:44(calculate_z_serial_purepython)
34219980 8.331 0.000 8.331 0.000 {built-in method builtins.abs}
2002000 0.299 0.000 0.299 0.000 {method 'append' of 'list' objects}
1 0.021 0.021 0.021 0.021 {built-in method builtins.sum}
3 0.001 0.000 0.001 0.000 {built-in method builtins.print}
2 0.000 0.000 0.000 0.000 {built-in method time.time}
4 0.000 0.000 0.000 0.000 {built-in method builtins.len}
1 0.000 0.000 0.000 0.000 {method 'disable' of '_lsprof.Profiler' objects}
<pstats.Stats object at 0x000001FA58817588>
Raw
file4.txt
> p.print_callers()
Ordered by: cumulative time
Function was called by...
ncalls tottime cumtime
{built-in method builtins.exec} <-
before_julia_set_1.py:4(<module>) <- 1 0.051 27.094 {built-in method builtins.exec}
before_julia_set_1.py:24(calc_pure_python) <- 1 1.218 27.044 before_julia_set_1.py:4(<module>)
before_julia_set_1.py:14(measure_time) <- 1 0.000 25.514 before_julia_set_1.py:24(calc_pure_python)
before_julia_set_1.py:64(calculate_z_serial_purepython) <- 1 18.297 25.514 before_julia_set_1.py:14(measure_time)
{built-in method builtins.abs} <- 34219980 7.217 7.217 before_julia_set_1.py:64(calculate_z_serial_purepython)
{method 'append' of 'list' objects} <- 2002000 0.296 0.296 before_julia_set_1.py:24(calc_pure_python)
{built-in method builtins.sum} <- 1 0.015 0.015 before_julia_set_1.py:24(calc_pure_python)
{built-in method builtins.print} <- 1 0.000 0.000 before_julia_set_1.py:14(measure_time)
3 0.000 0.000 before_julia_set_1.py:24(calc_pure_python)
before_julia_set_1.py:13(timefn) <- 1 0.000 0.000 before_julia_set_1.py:4(<module>)
C:\Users\kento\AppData\Local\Programs\Python\Python36\lib\functools.py:44(update_wrapper) <- 1 0.000 0.000 before_julia_set_1.py:13(timefn)
{built-in method time.time} <- 2 0.000 0.000 before_julia_set_1.py:14(measure_time)
2 0.000 0.000 before_julia_set_1.py:24(calc_pure_python)
<frozen importlib._bootstrap>:997(_handle_fromlist) <- 1 0.000 0.000 before_julia_set_1.py:4(<module>)
{built-in method builtins.len} <- 2 0.000 0.000 before_julia_set_1.py:24(calc_pure_python)
2 0.000 0.000 before_julia_set_1.py:64(calculate_z_serial_purepython)
{built-in method builtins.getattr} <- 7 0.000 0.000 C:\Users\kento\AppData\Local\Programs\Python\Python36\lib\functools.py:44(update_wrap
per)
以下略
Raw
file5.txt
>>> p.print_callees()
Ordered by: cumulative time
Function was called by...
ncalls tottime cumtime
{built-in method builtins.exec} <-
julia_set.py:1(<module>) <- 1 0.062 29.504 {built-in method builtins.exec}
julia_set.py:8(calc_pure_python) <- 1 1.272 29.442 julia_set.py:1(<module>)
julia_set.py:44(calculate_z_serial_purepython) <- 1 19.520 27.850 julia_set.py:8(calc_pure_python)
{built-in method builtins.abs} <- 34219980 8.331 8.331 julia_set.py:44(calculate_z_serial_purepython)
{method 'append' of 'list' objects} <- 2002000 0.299 0.299 julia_set.py:8(calc_pure_python)
{built-in method builtins.sum} <- 1 0.021 0.021 julia_set.py:8(calc_pure_python)
{built-in method builtins.print} <- 3 0.001 0.001 julia_set.py:8(calc_pure_python)
{built-in method time.time} <- 2 0.000 0.000 julia_set.py:8(calc_pure_python)
{built-in method builtins.len} <- 2 0.000 0.000 julia_set.py:8(calc_pure_python)
2 0.000 0.000 julia_set.py:44(calculate_z_serial_purepython)
{method 'disable' of '_lsprof.Profiler' objects} <-
<pstats.Stats object at 0x000001FA58817588>
Raw
file6.txt
$ pip install snakeviz
Raw
file7.txt
# profile.stats → ファイル名
$ snakeviz profile.stats
Raw
file8.txt
# エラーが出るがインストールはできる。
$ pip install line_profiler
$ git clone https://github.com/rkern/line_profiler.git
Raw
julia_set.py
import time
# グローバル定数
x1, x2, y1, y2 = -1.8, 1.8, -1.8, 1.8
c_real, c_imag = -0.62772, -0.42193
def calc_pure_python(desired_width, max_iterations):
x_step = (float(x2 - x1) / float(desired_width))
y_step = (float(y1 - y2) / float(desired_width))
x = []
y = []
ycoord = y2
while ycoord > y1:
y.append(ycoord)
ycoord += y_step
xcoord = x1
while xcoord < x2:
x.append(xcoord)
xcoord += x_step
# 座標リストを作り、各点を初期条件にする
zs = []
cs = []
for ycoord in y:
for xcoord in x:
zs.append(complex(xcoord, ycoord))
cs.append(complex(c_real, c_imag))
print('Length of x: ', len(x))
print('Total elements: ', len(zs))
start_time = time.time()
output = calculate_z_serial_purepython(max_iterations, zs, cs)
end_time = time.time()
secs = end_time - start_time
print(f'{ calculate_z_serial_purepython.__name__ } took { secs } seconds')
# 固定値 1000~2のグリッドを、繰り返し上限300回としたときの値
# 正当性チェック
assert sum(output) == 33219980
# CPU バウンドの計算関数
# ↓ に @profileを追加して line_profilerの計測対象にする
@profile
def calculate_z_serial_purepython(maxiter, zs, cs):
"""
ジュリア漸化式を使用して output リストを計算する
"""
output = [0] * len(zs)
for i in range(len(zs)):
n = 0
z = zs[i]
c = cs[i]
while abs(z) < 2 and n < maxiter:
z = z * z + c
n += 1
output[i] = n
return output
if __name__ == '__main__':
# ジュリア集合を計算する
calc_pure_python(desired_width=1000, max_iterations=300)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment