win10下安装theano和实现GPU加速

win10 

显卡 GT 720

*  安装Anaconda2 (python2.7)

http://www.continuum.io/downloads

确保系统环境变量PATH已添加

C:\Anaconda2

C:\Anaconda2\Scripts

C:\Anaconda2\Library\bin

* 安装 Visual studio 2013

这里用的是 VS2013_RTM_ULT_CHS.iso， 之前一直不成功不知道是不是没有用这个版本 。

系统环境变量PATH添加

C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin

C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\x86_amd64

VS的版本要和CUDA匹配，详见注意事项

*  安装 CUDA 7.5

https://developer.nvidia.com/cuda-downloads

测试cmd:

nvcc -V

*  安装minGW

cmd:

conda install mingw libpython

确保系统环境变量PATH已添加

C:\Anaconda2\MinGW\bin
C:\Anaconda2\MinGW\x86_64-w64-mingw32\lib

*  安装theano

cmd:

pip install theano

在当前用户的目录（如 C:\Users\valex, 一般为cmd的默认路径，但也不一定）下

创建文件 .theanorc.txt, 内容如下：

[blas]
ldflags=

[global]
openmp=False 
floatX = float32
device = gpu
allow_input_downcast=True 

[nvcc]
fastmath = True
flags = -LC:\Anaconda2\libs
compiler_bindir = C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin

[gcc]
cxxflags = -IC:\Anaconda2\MinGW
fastmath = True

[cuda]
root = C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v7.5\bin

*  测试是否可用GPU加速

from theano import function, config, shared, sandbox
import theano.tensor as T
import numpy
import time

vlen = 10 * 30 * 768  # 10 x #cores x # threads per core
iters = 1000

rng = numpy.random.RandomState(22)
x = shared(numpy.asarray(rng.rand(vlen), config.floatX))
f = function([], T.exp(x))
print (f.maker.fgraph.toposort())
t0 = time.time()
for i in range(iters):
    r = f()
t1 = time.time()
print ('Looping %d times took' % iters, t1 - t0, 'seconds')
print ('Result is', r)
if numpy.any([isinstance(x.op, T.Elemwise) for x in f.maker.fgraph.toposort()]):
    print ('Used the cpu')
else:
    print ('Used the gpu')

*  注意事项

http://blog.sina.com.cn/s/blog_87ecc6830102wnh1.html