API tools faq
paste
Advertisement
Guest User

Untitled

a guest
Feb 8th, 2017
168
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 6.63 KB | None | 0 0
raw download clone embed print report
  1. (sklearn-env) [ec2-user@ip-172-31-11-89 ~]$ ipython pairwise_distances_benchmark.py False cosine
  2. ================================================================================
  3.  
  4. # sparse=False, n_x=100000, n_y=1000, n_dim=1000
  5. # X array: 0.8 GB, Y array 0.008 GB, result array 0.8 GB
  6. # metric = cosine
  7. ================================================================================
  8.  
  9. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  10. n_jobs= 1 => 1 loop, best of 1: 2.55 s per loop
  11. n_jobs= 2 => 1 loop, best of 1: 7.82 s per loop
  12. n_jobs= 4 => 1 loop, best of 1: 9.4 s per loop
  13. n_jobs= 8 => 1 loop, best of 1: 13.4 s per loop
  14. n_jobs= 16 => 1 loop, best of 1: 23.2 s per loop
  15.  
  16. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  17. n_jobs= 1 => 1 loop, best of 1: 3.13 s per loop
  18. n_jobs= 2 => 1 loop, best of 1: 8.22 s per loop
  19. n_jobs= 4 => 1 loop, best of 1: 12.8 s per loop
  20. n_jobs= 8 => 1 loop, best of 1: 22.8 s per loop
  21. n_jobs= 16 => 1 loop, best of 1: 42.9 s per loop
  22.  
  23. ## {'backend': 'threading', 'MKL_NUM_THREADS': 8}
  24. n_jobs= 1 => 1 loop, best of 1: 3.12 s per loop
  25. n_jobs= 2 => 1 loop, best of 1: 3.08 s per loop
  26. n_jobs= 4 => 1 loop, best of 1: 3.21 s per loop
  27. n_jobs= 8 => 1 loop, best of 1: 4.07 s per loop
  28. n_jobs= 16 => 1 loop, best of 1: 6.3 s per loop
  29.  
  30. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  31. n_jobs= 1 => 1 loop, best of 1: 3.17 s per loop
  32. n_jobs= 2 => 1 loop, best of 1: 8.05 s per loop
  33. n_jobs= 4 => 1 loop, best of 1: 9.48 s per loop
  34. n_jobs= 8 => 1 loop, best of 1: 13.5 s per loop
  35. n_jobs= 16 => 1 loop, best of 1: 23.3 s per loop
  36.  
  37. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  38. n_jobs= 1 => 1 loop, best of 1: 3.2 s per loop
  39. n_jobs= 2 => 1 loop, best of 1: 8.25 s per loop
  40. n_jobs= 4 => 1 loop, best of 1: 12.9 s per loop
  41. n_jobs= 8 => 1 loop, best of 1: 22.7 s per loop
  42. n_jobs= 16 => 1 loop, best of 1: 42.7 s per loop
  43.  
  44. ## {'backend': 'threading', 'MKL_NUM_THREADS': 1}
  45. n_jobs= 1 => 1 loop, best of 1: 3.17 s per loop
  46. n_jobs= 2 => 1 loop, best of 1: 3.22 s per loop
  47. n_jobs= 4 => 1 loop, best of 1: 3.28 s per loop
  48. n_jobs= 8 => 1 loop, best of 1: 4.12 s per loop
  49. n_jobs= 16 => 1 loop, best of 1: 6.2 s per loop
  50.  
  51. ================================================================================
  52.  
  53. # sparse=False, n_x=10000, n_y=10000, n_dim=1000
  54. # X array: 0.08 GB, Y array 0.08 GB, result array 0.8 GB
  55. # metric = cosine
  56. ================================================================================
  57.  
  58. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  59. n_jobs= 1 => 1 loop, best of 1: 2.15 s per loop
  60. n_jobs= 2 => 1 loop, best of 1: 5.18 s per loop
  61. n_jobs= 4 => 1 loop, best of 1: 4.66 s per loop
  62. n_jobs= 8 => 1 loop, best of 1: 4.42 s per loop
  63. n_jobs= 16 => 1 loop, best of 1: 4.53 s per loop
  64.  
  65. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  66. n_jobs= 1 => 1 loop, best of 1: 2.15 s per loop
  67. n_jobs= 2 => 1 loop, best of 1: 4.65 s per loop
  68. n_jobs= 4 => 1 loop, best of 1: 7.27 s per loop
  69. n_jobs= 8 => 1 loop, best of 1: 13 s per loop
  70. n_jobs= 16 => 1 loop, best of 1: 25 s per loop
  71.  
  72. ## {'backend': 'threading', 'MKL_NUM_THREADS': 8}
  73. n_jobs= 1 => 1 loop, best of 1: 2.15 s per loop
  74. n_jobs= 2 => 1 loop, best of 1: 1.82 s per loop
  75. n_jobs= 4 => 1 loop, best of 1: 1.92 s per loop
  76. n_jobs= 8 => 1 loop, best of 1: 2.02 s per loop
  77. n_jobs= 16 => 1 loop, best of 1: 2.48 s per loop
  78.  
  79. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  80. n_jobs= 1 => 1 loop, best of 1: 2.14 s per loop
  81. n_jobs= 2 => 1 loop, best of 1: 5.28 s per loop
  82. n_jobs= 4 => 1 loop, best of 1: 4.73 s per loop
  83. n_jobs= 8 => 1 loop, best of 1: 4.63 s per loop
  84. n_jobs= 16 => 1 loop, best of 1: 4.57 s per loop
  85.  
  86. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  87. n_jobs= 1 => 1 loop, best of 1: 2.14 s per loop
  88. n_jobs= 2 => 1 loop, best of 1: 4.67 s per loop
  89. n_jobs= 4 => 1 loop, best of 1: 7.3 s per loop
  90. n_jobs= 8 => 1 loop, best of 1: 13.1 s per loop
  91. n_jobs= 16 => 1 loop, best of 1: 25 s per loop
  92.  
  93. ## {'backend': 'threading', 'MKL_NUM_THREADS': 1}
  94. n_jobs= 1 => 1 loop, best of 1: 2.13 s per loop
  95. n_jobs= 2 => 1 loop, best of 1: 1.88 s per loop
  96. n_jobs= 4 => 1 loop, best of 1: 1.98 s per loop
  97. n_jobs= 8 => 1 loop, best of 1: 2.02 s per loop
  98. n_jobs= 16 => 1 loop, best of 1: 2.29 s per loop
  99. ================================================================================
  100.  
  101. # sparse=False, n_x=10000, n_y=10000, n_dim=10
  102. # X array: 0.0008 GB, Y array 0.0008 GB, result array 0.8 GB
  103. # metric = cosine
  104. ================================================================================
  105.  
  106. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  107. n_jobs= 1 => 1 loop, best of 1: 726 ms per loop
  108. n_jobs= 2 => 1 loop, best of 1: 4.16 s per loop
  109. n_jobs= 4 => 1 loop, best of 1: 3.8 s per loop
  110. n_jobs= 8 => 1 loop, best of 1: 3.65 s per loop
  111. n_jobs= 16 => 1 loop, best of 1: 3.78 s per loop
  112.  
  113. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 8}
  114. n_jobs= 1 => 1 loop, best of 1: 724 ms per loop
  115. n_jobs= 2 => 1 loop, best of 1: 3.54 s per loop
  116. n_jobs= 4 => 1 loop, best of 1: 5.86 s per loop
  117. n_jobs= 8 => 1 loop, best of 1: 10.7 s per loop
  118. n_jobs= 16 => 1 loop, best of 1: 20.7 s per loop
  119.  
  120. ## {'backend': 'threading', 'MKL_NUM_THREADS': 8}
  121. n_jobs= 1 => 1 loop, best of 1: 719 ms per loop
  122. n_jobs= 2 => 1 loop, best of 1: 1.12 s per loop
  123. n_jobs= 4 => 1 loop, best of 1: 1.01 s per loop
  124. n_jobs= 8 => 1 loop, best of 1: 1.19 s per loop
  125. n_jobs= 16 => 1 loop, best of 1: 1.4 s per loop
  126.  
  127. ## {'mmap_result': False, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  128. n_jobs= 1 => 1 loop, best of 1: 728 ms per loop
  129. n_jobs= 2 => 1 loop, best of 1: 4.21 s per loop
  130. n_jobs= 4 => 1 loop, best of 1: 3.86 s per loop
  131. n_jobs= 8 => 1 loop, best of 1: 3.66 s per loop
  132. n_jobs= 16 => 1 loop, best of 1: 3.74 s per loop
  133.  
  134. ## {'mmap_result': True, 'backend': 'multiprocessing', 'MKL_NUM_THREADS': 1}
  135. n_jobs= 1 => 1 loop, best of 1: 715 ms per loop
  136. n_jobs= 2 => 1 loop, best of 1: 3.64 s per loop
  137. n_jobs= 4 => 1 loop, best of 1: 5.87 s per loop
  138. n_jobs= 8 => 1 loop, best of 1: 10.7 s per loop
  139. n_jobs= 16 => 1 loop, best of 1: 20.8 s per loop
  140.  
  141. ## {'backend': 'threading', 'MKL_NUM_THREADS': 1}
  142. n_jobs= 1 => 1 loop, best of 1: 723 ms per loop
  143. n_jobs= 2 => 1 loop, best of 1: 1.09 s per loop
  144. n_jobs= 4 => 1 loop, best of 1: 1.1 s per loop
  145. n_jobs= 8 => 1 loop, best of 1: 1.23 s per loop
  146. n_jobs= 16 => 1 loop, best of 1: 1.36 s per loop
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement
Public Pastes
Advertisement
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!