Skip to content

Instantly share code, notes, and snippets.

@jnothman

jnothman/bench_working_memory.csv

Last active January 9, 2018 21:27
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save jnothman/2394b3e199438a500ed9c8f8c9e17500 to your computer and use it in GitHub Desktop.
Save jnothman/2394b3e199438a500ed9c8f8c9e17500 to your computer and use it in GitHub Desktop.
Download ZIP
Benchmarks for working_memory, i.e. https://github.com/scikit-learn/scikit-learn/pull/10280
Raw
bench_working_memory.csv
elapsed n_features n_samples working_memory
0 259.7408814430237 100 40000 1
1 263.6345293521881 100 40000 1
2 263.4755172729492 100 40000 1
3 32.214531898498535 100 20000 1
4 32.38340449333191 100 20000 1
5 32.6368203163147 100 20000 1
6 4.7826080322265625 100 10000 1
7 4.722550392150879 100 10000 1
8 4.735676050186157 100 10000 1
9 0.8438136577606201 100 5000 1
10 0.8434607982635498 100 5000 1
11 0.8423585891723633 100 5000 1
12 0.027704477310180664 100 1000 1
13 0.027656078338623047 100 1000 1
14 0.02758312225341797 100 1000 1
15 0.0010852813720703125 100 100 1
16 0.0006594657897949219 100 100 1
17 0.0006537437438964844 100 100 1
18 55.213985443115234 100 40000 10
19 58.77374887466431 100 40000 10
20 55.19669032096863 100 40000 10
21 11.520464181900024 100 20000 10
22 11.497724294662476 100 20000 10
23 11.49088716506958 100 20000 10
24 2.747018337249756 100 10000 10
25 2.733360767364502 100 10000 10
26 2.8228633403778076 100 10000 10
27 0.675994873046875 100 5000 10
28 0.6749563217163086 100 5000 10
29 0.6764612197875977 100 5000 10
30 0.023439884185791016 100 1000 10
31 0.02269458770751953 100 1000 10
32 0.023115873336791992 100 1000 10
33 0.0008337497711181641 100 100 10
34 0.0006406307220458984 100 100 10
35 0.0006313323974609375 100 100 10
36 47.73020386695862 100 40000 22
37 47.69992017745972 100 40000 22
38 47.62230563163757 100 40000 22
39 11.639677047729492 100 20000 22
40 11.566004514694214 100 20000 22
41 11.810020446777344 100 20000 22
42 2.889223575592041 100 10000 22
43 2.749232292175293 100 10000 22
44 2.7498748302459717 100 10000 22
45 0.6866409778594971 100 5000 22
46 0.6885111331939697 100 5000 22
47 0.690028190612793 100 5000 22
48 0.022968053817749023 100 1000 22
49 0.02248072624206543 100 1000 22
50 0.02284836769104004 100 1000 22
51 0.0007929801940917969 100 100 22
52 0.0006601810455322266 100 100 22
53 0.0006287097930908203 100 100 22
54 47.31692409515381 100 40000 46
55 49.07538294792175 100 40000 46
56 48.277475118637085 100 40000 46
57 11.227022647857666 100 20000 46
58 11.562239646911621 100 20000 46
59 11.235973119735718 100 20000 46
60 2.7667436599731445 100 10000 46
61 2.679640769958496 100 10000 46
62 2.6795284748077393 100 10000 46
63 0.6811165809631348 100 5000 46
64 0.6973702907562256 100 5000 46
65 0.7564349174499512 100 5000 46
66 0.025141000747680664 100 1000 46
67 0.02521491050720215 100 1000 46
68 0.024368762969970703 100 1000 46
69 0.0007998943328857422 100 100 46
70 0.0006432533264160156 100 100 46
71 0.0006325244903564453 100 100 46
72 46.916279554367065 100 40000 100
73 46.784446716308594 100 40000 100
74 47.71704292297363 100 40000 100
75 11.405823945999146 100 20000 100
76 11.669658899307251 100 20000 100
77 11.365206956863403 100 20000 100
78 2.992124319076538 100 10000 100
79 2.9002745151519775 100 10000 100
80 2.6818411350250244 100 10000 100
81 0.681471586227417 100 5000 100
82 0.7413830757141113 100 5000 100
83 0.6922602653503418 100 5000 100
84 0.023858070373535156 100 1000 100
85 0.023314237594604492 100 1000 100
86 0.02337813377380371 100 1000 100
87 0.0008080005645751953 100 100 100
88 0.0006475448608398438 100 100 100
89 0.0006434917449951172 100 100 100
90 48.19805312156677 100 40000 215
91 45.96044921875 100 40000 215
92 47.590593338012695 100 40000 215
93 12.487635135650635 100 20000 215
94 11.264180183410645 100 20000 215
95 11.47057843208313 100 20000 215
96 2.7389330863952637 100 10000 215
97 2.9447615146636963 100 10000 215
98 3.197394371032715 100 10000 215
99 0.7132112979888916 100 5000 215
100 0.714383602142334 100 5000 215
101 0.6667075157165527 100 5000 215
102 0.025145769119262695 100 1000 215
103 0.024105310440063477 100 1000 215
104 0.02427816390991211 100 1000 215
105 0.001071929931640625 100 100 215
106 0.0009653568267822266 100 100 215
107 0.0008819103240966797 100 100 215
108 46.257248401641846 100 40000 464
109 45.165510177612305 100 40000 464
110 45.441216230392456 100 40000 464
111 10.508992910385132 100 20000 464
112 11.447564363479614 100 20000 464
113 11.426690816879272 100 20000 464
114 3.0052504539489746 100 10000 464
115 2.8551981449127197 100 10000 464
116 2.807879686355591 100 10000 464
117 0.855536937713623 100 5000 464
118 0.6697769165039062 100 5000 464
119 0.7944178581237793 100 5000 464
120 0.024903297424316406 100 1000 464
121 0.024701595306396484 100 1000 464
122 0.02453017234802246 100 1000 464
123 0.0008106231689453125 100 100 464
124 0.00064849853515625 100 100 464
125 0.0006377696990966797 100 100 464
126 44.7512321472168 100 40000 1000
127 43.10793614387512 100 40000 1000
128 43.913618326187134 100 40000 1000
129 11.027376651763916 100 20000 1000
130 10.156231164932251 100 20000 1000
131 10.327084064483643 100 20000 1000
132 2.692312240600586 100 10000 1000
133 2.640241861343384 100 10000 1000
134 2.6469826698303223 100 10000 1000
135 0.668205738067627 100 5000 1000
136 0.6639533042907715 100 5000 1000
137 0.663642168045044 100 5000 1000
138 0.025493383407592773 100 1000 1000
139 0.02452874183654785 100 1000 1000
140 0.025629758834838867 100 1000 1000
141 0.0009765625 100 100 1000
142 0.0008301734924316406 100 100 1000
143 0.0008120536804199219 100 100 1000
144 42.89431428909302 100 40000 10000
145 42.75436544418335 100 40000 10000
146 39.34532022476196 100 40000 10000
147 10.498741149902344 100 20000 10000
148 10.209395170211792 100 20000 10000
149 10.23120379447937 100 20000 10000
150 2.631664752960205 100 10000 10000
151 2.6318283081054688 100 10000 10000
152 2.548217535018921 100 10000 10000
153 0.6596729755401611 100 5000 10000
154 0.7769567966461182 100 5000 10000
155 0.8544619083404541 100 5000 10000
156 0.034839630126953125 100 1000 10000
157 0.03863096237182617 100 1000 10000
158 0.0359342098236084 100 1000 10000
159 0.001064300537109375 100 100 10000
160 0.0009489059448242188 100 100 10000
161 0.0009412765502929688 100 100 10000
Raw
bench_working_memory.py
import time
import numpy as np
import pandas as pd
import sklearn
from sklearn.metrics.pairwise import pairwise_distances_argmin_min
n_features = 100
results = []
for working_memory in [1, 10, 22, 46, 100, 215, 464, 1000, 10000]:
with sklearn.config_context(working_memory=working_memory):
for n_samples in [100, 1000, 5000, 10000, 20000, 40000][::-1]:
for i in range(3):
X = np.random.rand(n_samples, n_features)
start = time.time()
pairwise_distances_argmin_min(X, X)
elapsed = time.time() - start
results.append({'n_samples': n_samples,
'n_features': n_features,
'working_memory': working_memory,
'elapsed': elapsed})
print(results[-1])
pd.DataFrame(results).to_csv('bench_working_memory.csv')
pd.DataFrame(results).to_json('bench_working_memory.json', orient='records')
Raw
plot_bench.py
import matplotlib.pyplot as plt
import pandas as pd
bench = pd.read_csv('bench_working_memory.csv')
piv = bench.pivot_table('elapsed', ['working_memory'], ['n_samples', 'n_features'])
piv.plot(logy=True, logx=True)
plt.xlabel('working_memory')
plt.ylabel('time (s)')
plt.show()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment