在看這篇部落格之前，作者的執行環境首先要裝好cuda和cudnn。然後我的環境為：

ubuntu 16.04
anaconda
python 3.6
theano                    1.0.2            py36h6bb024c_0 

首先安裝pygpu

conda install -c conda-forge pygpu

然後進行theano-gpu測試，我的測試程式碼為theano-gpu-test.py

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') 
 

執行的命令gpu版本：

 THEANO_FLAGS=mode=FAST_RUN,device=cuda,floatX=float32 python theano-gpu-test.py

我的輸出是：

Using cuDNN version 6021 on context None
Mapped name None to device cuda: GeForce GTX 1080 Ti (0000:03:00.0)
[GpuElemwise{exp,no_inplace}(<GpuArrayType<None>(float32, vector)>), HostFromGpu(gpuarray)(GpuElemwise{exp,no_inplace}.0)]
Looping 1000 times took 0.23069453239440918 seconds
Result is [1.2317803 1.6187935 1.5227807 ... 2.2077181 2.2996776 1.623233 ]
Used the cpu　　＃這行沒有參考意義，看上面的用時
 

執行命令CPU版本：

python theano-gpu-test.py

我的電腦輸出為：

[Elemwise{exp,no_inplace}(<TensorType(float64, vector)>)]
Looping 1000 times took 26.147013664245605 seconds
Result is [1.23178032 1.61879341 1.52278065 ... 2.20771815 2.29967753 1.62323285]
Used the cpu　　＃這行沒有參考意義，看上面的用時

用GPU執行程式花了0.23秒，用CPU執行程式花了26秒

如果不想每次用gpu的時候加上前面的一長串，則需要配置.theanorc檔案，然後就直接輸入python theano-gpu-test.py命令就行了,我這裡把我配置的過程分享一下:

vim ~/.theanorc

然後加上

[global]
floatX=float32
device=cuda

按esc :wq退出就行了，親測可用。

參考文獻

[1].Theano2.1.12-基礎知識之使用GPU.https://www.cnblogs.com/shouhuxianjian/p/4590224.html

[2].Theano坑--無法呼叫GPU. https://www.jianshu.com/p/12af936e20ef

[3].gpuarray. http://deeplearning.net/software/libgpuarray/installation.html