如何加速python代码--关于Numpy(CuPy)--Numpy(pycuda)--Numpy(pyopencl)--numba--MinPy--阿里Mars等

参考书籍；

《Hands-On GPU Programming with Python and CUDA》by Dr. Brian Tuomanen

参考网站：

https://www.cnblogs.com/noluye/p/11517489.html

 

如何将Numpy加速700倍？用 CuPy 呀

https://github.com/cupy/cupy

例子：

https://github.com/cupy/cupy/tree/master/examples

教程：

https://docs-cupy.chainer.org/en/stable/tutorial/

社区：

https://groups.google.com/forum/#!forum/cupy

https://www.jiqizhixin.com/articles/2019-08-29-8

https://towardsdatascience.com/heres-how-to-use-cupy-to-make-numpy-700x-faster-4b920dda1f56

i7–8700k CPU
1080 Ti GPU
32 GB of DDR4 3000MHz RAM
CUDA 9.0

The latest version of cuDNN and NCCL libraries are included in binary packages (wheels). For the source package, you will need to install cuDNN/NCCL before installing CuPy, if you want to use it.

依赖CUDA cuDNN NCCL

pip install cupy
pip install cupy-cuda90
pip install cupy-cuda100
pip install cupy-cuda101


import numpy as np
import cupy as cp
import time
### Numpy and CPU
s = time.time()
*x_cpu = np.ones((1000,1000,1000))*
e = time.time()
print(e - s)### CuPy and GPU
s = time.time()
*x_gpu = cp.ones((1000,1000,1000))*
e = time.time()
print(e - s)

---

https://documen.tician.de/pycuda/

https://pypi.org/project/pycuda/

https://github.com/inducer/pycuda

https://mathema.tician.de/software/pycuda/

PyCUDA 文档中文翻译-Tutorial 中文版

https://zhuanlan.zhihu.com/p/32062796?utm_source=weibo

---

pyopencl 2019.1.2

https://github.com/inducer/pyopencl

https://mathema.tician.de/software/pyopencl/

https://wiki.tiker.net/PyOpenCL

https://documen.tician.de/pyopencl/

---

https://zhuanlan.zhihu.com/p/77307505

http://numba.pydata.org/

http://numba.pydata.org/numba-doc/latest/cuda/index.html

https://github.com/numba/numba

https://zhuanlan.zhihu.com/p/27152060

---

https://github.com/dmlc/minpy

https://minpy.readthedocs.io/en/latest/

https://minpy.readthedocs.io/en/latest/feature/limitation.html

专栏 | MinPy：剑气双修的武功秘籍

http://www.sohu.com/a/124626121_465975

---

Mars

https://www.jianshu.com/p/2c77b5b1f853

https://www.jiqizhixin.com/articles/2018-12-26-13