Skip to content

Instantly share code, notes, and snippets.

@yuxuanzhuang

yuxuanzhuang/benchmark_on_thread_effect_of_numpy.ipynb Secret

Last active October 14, 2020 15:27
Show Gist options
  • Save yuxuanzhuang/82e1e7b57d0cda80ac964d1cd138f618 to your computer and use it in GitHub Desktop.
Save yuxuanzhuang/82e1e7b57d0cda80ac964d1cd138f618 to your computer and use it in GitHub Desktop.
Download ZIP
Display the source blob
Display the rendered blob
Raw
benchmark_on_thread_effect_of_numpy.ipynb
{
"cells": [
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import timeit\n",
"%matplotlib inline"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [],
"source": [
"from threadpoolctl import threadpool_limits "
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [],
"source": [
"import time"
]
},
{
"cell_type": "code",
"execution_count": 43,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The line_profiler extension is already loaded. To reload it, use:\n",
" %reload_ext line_profiler\n"
]
}
],
"source": [
"%load_ext line_profiler"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Without np.dot"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Time profile"
]
},
{
"cell_type": "code",
"execution_count": 55,
"metadata": {},
"outputs": [],
"source": [
"def thread_effect_on_numpy(size, n_thread):\n",
" with threadpool_limits(n_thread):\n",
" for i in range(20):\n",
" ts = np.random.rand(size,3)\n",
"# _ = np.dot(ts, np.random.rand(3,3))\n",
" _ = ts + np.random.rand(size,3)"
]
},
{
"cell_type": "code",
"execution_count": 56,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"70.7 ms ± 811 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 1)"
]
},
{
"cell_type": "code",
"execution_count": 57,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"70.4 ms ± 383 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 6)"
]
},
{
"cell_type": "code",
"execution_count": 58,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"70.2 ms ± 442 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 12)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Line profile"
]
},
{
"cell_type": "code",
"execution_count": 62,
"metadata": {},
"outputs": [],
"source": [
"def thread_effect_on_numpy(size, n_thread):\n",
" with threadpool_limits(n_thread):\n",
" for i in range(20):\n",
" ts = np.random.rand(size,3)\n",
"# _ = np.dot(ts, np.random.rand(3,3))\n",
" _ = ts + np.random.rand(size,3)\n",
" time.sleep(0.5) # a buffer time between two iter"
]
},
{
"cell_type": "code",
"execution_count": 63,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.0906 s\n",
"File: <ipython-input-62-69068f548b95>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5194.0 5194.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 139.0 6.6 0.0 for i in range(20):\n",
" 4 20 35716.0 1785.8 0.4 ts = np.random.rand(size,3)\n",
" 5 # _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 37070.0 1853.5 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10012475.0 500623.8 99.2 time.sleep(0.5)"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 1)"
]
},
{
"cell_type": "code",
"execution_count": 64,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.0891 s\n",
"File: <ipython-input-62-69068f548b95>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 4964.0 4964.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 135.0 6.4 0.0 for i in range(20):\n",
" 4 20 34964.0 1748.2 0.3 ts = np.random.rand(size,3)\n",
" 5 # _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 37132.0 1856.6 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10011892.0 500594.6 99.2 time.sleep(0.5)"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 2)"
]
},
{
"cell_type": "code",
"execution_count": 65,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.0892 s\n",
"File: <ipython-input-62-69068f548b95>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5189.0 5189.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 134.0 6.4 0.0 for i in range(20):\n",
" 4 20 34625.0 1731.2 0.3 ts = np.random.rand(size,3)\n",
" 5 # _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 37139.0 1857.0 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10012092.0 500604.6 99.2 time.sleep(0.5)"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 6)"
]
},
{
"cell_type": "code",
"execution_count": 66,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.0891 s\n",
"File: <ipython-input-62-69068f548b95>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5152.0 5152.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 134.0 6.4 0.0 for i in range(20):\n",
" 4 20 34760.0 1738.0 0.3 ts = np.random.rand(size,3)\n",
" 5 # _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 37205.0 1860.2 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10011816.0 500590.8 99.2 time.sleep(0.5)"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 12)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# With np.dot"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Time profile"
]
},
{
"cell_type": "code",
"execution_count": 51,
"metadata": {},
"outputs": [],
"source": [
"def thread_effect_on_numpy(size, n_thread):\n",
" with threadpool_limits(n_thread):\n",
" for i in range(20):\n",
" ts = np.random.rand(size,3)\n",
" _ = np.dot(ts, np.random.rand(3,3))\n",
" _ = ts + np.random.rand(size,3)"
]
},
{
"cell_type": "code",
"execution_count": 52,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"88.1 ms ± 516 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 1)"
]
},
{
"cell_type": "code",
"execution_count": 53,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"81.8 ms ± 349 µs per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 6)"
]
},
{
"cell_type": "code",
"execution_count": 54,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"162 ms ± 3.74 ms per loop (mean ± std. dev. of 7 runs, 10 loops each)\n"
]
}
],
"source": [
"%timeit thread_effect_on_numpy(100000, 12)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Line profile"
]
},
{
"cell_type": "code",
"execution_count": 67,
"metadata": {},
"outputs": [],
"source": [
"def thread_effect_on_numpy(size, n_thread):\n",
" with threadpool_limits(n_thread):\n",
" for i in range(20):\n",
" ts = np.random.rand(size,3)\n",
" _ = np.dot(ts, np.random.rand(3,3))\n",
" _ = ts + np.random.rand(size,3)\n",
" time.sleep(0.5) # a buffer time between two iter"
]
},
{
"cell_type": "code",
"execution_count": 68,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.1036 s\n",
"File: <ipython-input-67-7d449ec863ee>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5224.0 5224.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 136.0 6.5 0.0 for i in range(20):\n",
" 4 20 34411.0 1720.5 0.3 ts = np.random.rand(size,3)\n",
" 5 20 14159.0 708.0 0.1 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 37360.0 1868.0 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10012273.0 500613.7 99.1 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 1)"
]
},
{
"cell_type": "code",
"execution_count": 69,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.1006 s\n",
"File: <ipython-input-67-7d449ec863ee>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5138.0 5138.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 138.0 6.6 0.0 for i in range(20):\n",
" 4 20 34689.0 1734.5 0.3 ts = np.random.rand(size,3)\n",
" 5 20 9780.0 489.0 0.1 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 38269.0 1913.5 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10012557.0 500627.8 99.1 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 2)"
]
},
{
"cell_type": "code",
"execution_count": 70,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.0969 s\n",
"File: <ipython-input-67-7d449ec863ee>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5083.0 5083.0 0.1 with threadpool_limits(n_thread):\n",
" 3 21 131.0 6.2 0.0 for i in range(20):\n",
" 4 20 34686.0 1734.3 0.3 ts = np.random.rand(size,3)\n",
" 5 20 6795.0 339.8 0.1 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 38038.0 1901.9 0.4 _ = ts + np.random.rand(size,3)\n",
" 7 20 10012139.0 500607.0 99.2 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 6)"
]
},
{
"cell_type": "code",
"execution_count": 71,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.1305 s\n",
"File: <ipython-input-67-7d449ec863ee>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 4809.0 4809.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 136.0 6.5 0.0 for i in range(20):\n",
" 4 20 34591.0 1729.5 0.3 ts = np.random.rand(size,3)\n",
" 5 20 6837.0 341.9 0.1 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 72670.0 3633.5 0.7 _ = ts + np.random.rand(size,3)\n",
" 7 20 10011480.0 500574.0 98.8 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 12)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Non-numpy Operation after np.dot"
]
},
{
"cell_type": "code",
"execution_count": 77,
"metadata": {},
"outputs": [],
"source": [
"def thread_effect_on_numpy(size, n_thread):\n",
" with threadpool_limits(n_thread):\n",
" for i in range(20):\n",
" ts = np.random.rand(size,3)\n",
" _ = np.dot(ts, np.random.rand(3,3))\n",
" _ = [i * i for i in range(100000)]\n",
" time.sleep(0.5) # a buffer time between two iter"
]
},
{
"cell_type": "code",
"execution_count": 78,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.2795 s\n",
"File: <ipython-input-77-1bc3cbed1859>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 4985.0 4985.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 148.0 7.0 0.0 for i in range(20):\n",
" 4 20 35276.0 1763.8 0.3 ts = np.random.rand(size,3)\n",
" 5 20 26085.0 1304.2 0.3 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 201422.0 10071.1 2.0 _ = [i * i for i in range(100000)]\n",
" 7 20 10011595.0 500579.8 97.4 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 1)"
]
},
{
"cell_type": "code",
"execution_count": 79,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.2745 s\n",
"File: <ipython-input-77-1bc3cbed1859>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 4776.0 4776.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 143.0 6.8 0.0 for i in range(20):\n",
" 4 20 34844.0 1742.2 0.3 ts = np.random.rand(size,3)\n",
" 5 20 21775.0 1088.8 0.2 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 200999.0 10050.0 2.0 _ = [i * i for i in range(100000)]\n",
" 7 20 10011933.0 500596.7 97.4 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 2)"
]
},
{
"cell_type": "code",
"execution_count": 82,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.5088 s\n",
"File: <ipython-input-77-1bc3cbed1859>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 4823.0 4823.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 135.0 6.4 0.0 for i in range(20):\n",
" 4 20 34576.0 1728.8 0.3 ts = np.random.rand(size,3)\n",
" 5 20 29308.0 1465.4 0.3 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 428546.0 21427.3 4.1 _ = [i * i for i in range(100000)]\n",
" 7 20 10011364.0 500568.2 95.3 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 6)"
]
},
{
"cell_type": "code",
"execution_count": 81,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Timer unit: 1e-06 s\n",
"\n",
"Total time: 10.5215 s\n",
"File: <ipython-input-77-1bc3cbed1859>\n",
"Function: thread_effect_on_numpy at line 1\n",
"\n",
"Line # Hits Time Per Hit % Time Line Contents\n",
"==============================================================\n",
" 1 def thread_effect_on_numpy(size, n_thread):\n",
" 2 1 5186.0 5186.0 0.0 with threadpool_limits(n_thread):\n",
" 3 21 131.0 6.2 0.0 for i in range(20):\n",
" 4 20 34568.0 1728.4 0.3 ts = np.random.rand(size,3)\n",
" 5 20 28471.0 1423.5 0.3 _ = np.dot(ts, np.random.rand(3,3))\n",
" 6 20 440734.0 22036.7 4.2 _ = [i * i for i in range(100000)]\n",
" 7 20 10012452.0 500622.6 95.2 time.sleep(0.5) # a buffer time between two iter"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"%lprun -f thread_effect_on_numpy thread_effect_on_numpy(100000, 12)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "gsoc",
"language": "python",
"name": "gsoc"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.3"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
@yuxuanzhuang
Copy link
Author

yuxuanzhuang commented Oct 14, 2020 •
edited
Loading

As a sum up for the code performance after numpy.dot (Line 6)

Threads Time
version_1 version_2 version_3
without numpy.dot 1 37070 37606 36344
2 37132 37527 36547
6 37139 37747 36638
12 37205 37493 36684 (not really using 12 threads with MKL)
with numpy.dot 1 37360 36285 36369
2 38269 36373 37200
6 38038 39142 36863
12 72670 67931 36739 (not really using 12 threads with MKL)
non numpy operation after numpy.dot 1 201422 208773 213470
2 200999 208948 212878
6 428546 216324 213259
12 440734 463529 213133 (not really using 12 threads with MKL)

Another interesting observation is that non-numpy operation is already performing pretty bad with 6 threads.

The test was down with

  • 6-core CPU (hyperthreading enabled)

version_1

  • numpy 1.19.1
  • python 3.8
threadpool_info()
[{'filepath': '/home/scottzhuang/anaconda3/envs/gsoc/lib/python3.8/site-packages/numpy.libs/libopenblasp-r0-ae94cfde.3.9.dev.so',
  'prefix': 'libopenblas',
  'user_api': 'blas',
  'internal_api': 'openblas',
  'version': '0.3.9.dev',
  'num_threads': 12,
  'threading_layer': 'pthreads'},
 {'filepath': '/home/scottzhuang/anaconda3/envs/gsoc/lib/libgomp.so.1.0.0',
  'prefix': 'libgomp',
  'user_api': 'openmp',
  'internal_api': 'openmp',
  'version': None,
  'num_threads': 12},
 {'filepath': '/home/scottzhuang/anaconda3/envs/gsoc/lib/libmkl_rt.so',
  'prefix': 'libmkl_rt',
  'user_api': 'blas',
  'internal_api': 'mkl',
  'version': '2020.0.1',
  'num_threads': 6,
  'threading_layer': 'intel'},
 {'filepath': '/home/scottzhuang/anaconda3/envs/gsoc/lib/libiomp5.so',
  'prefix': 'libiomp',
  'user_api': 'openmp',
  'internal_api': 'openmp',
  'version': None,
  'num_threads': 12}]

version_2

  • numpy 1.20.0.dev0+4ccfbe6
  • python 3.8
threadpool_info()
[{'filepath': '/opt/OpenBLAS/lib/libopenblas_haswellp-r0.3.10.dev.so',
  'prefix': 'libopenblas',
  'user_api': 'blas',
  'internal_api': 'openblas',
  'version': '0.3.10.dev',
  'num_threads': 12,
  'threading_layer': 'pthreads'}]

version_3

  • numpy 1.19
  • python 3.8
threadpool_info()
[{'filepath': '/home/scottzhuang/anaconda3/lib/libmkl_rt.so',
  'prefix': 'libmkl_rt',
  'user_api': 'blas',
  'internal_api': 'mkl',
  'version': '2020.0.0',
  'num_threads': 6,
  'threading_layer': 'intel'},
 {'filepath': '/home/scottzhuang/anaconda3/lib/libiomp5.so',
  'prefix': 'libiomp',
  'user_api': 'openmp',
  'internal_api': 'openmp',
  'version': None,
  'num_threads': 12}]

Will check if it's version related.
EDIT: Update with other clean-installed version of numpy (OpenBlas, MKL).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment