如何使CUDA与python 3.5和numba一起使用

Question

By running the code:

import numba

numba.cuda.api.detect()

I get the error:

Traceback (most recent call last):
  File "<pyshell#1>", line 1, in <module>
    numba.cuda.api.detect()
  File "C:\Users\Wakasuki\AppData\Local\Programs\Python\Python35-32\lib\site-packages\numba\cuda\api.py", line 257, in detect
    print('Found %d CUDA devices' % len(devlist))
  File "C:\Users\Wakasuki\AppData\Local\Programs\Python\Python35-32\lib\site-packages\numba\cuda\cudadrv\devices.py", line 49, in __len__
    return len(self.lst)
  File "C:\Users\Wakasuki\AppData\Local\Programs\Python\Python35-32\lib\site-packages\numba\cuda\cudadrv\devices.py", line 26, in __getattr__
    numdev = driver.get_device_count()
  File "C:\Users\Wakasuki\AppData\Local\Programs\Python\Python35-32\lib\site-packages\numba\cuda\cudadrv\driver.py", line 269, in get_device_count
    self.cuDeviceGetCount(byref(count))
  File "C:\Users\Wakasuki\AppData\Local\Programs\Python\Python35-32\lib\site-packages\numba\cuda\cudadrv\driver.py", line 218, in __getattr__
    self.initialization_error)
numba.cuda.cudadrv.error.CudaSupportError: Error at driver init: 
CUDA disabled by user:

How do I enable CUDA within numba? I just recently installed CUDA toolkit 7.5 and the most recent numba and numpy in python 3.5.

Answer 1

Try as follows:

'...' is your installation path of CUDA toolkit 7.5 for "nvvm64_31_0.dll" and "libdevice"

import os
os.environ['NUMBAPRO_NVVM']=r'C:\NVIDIA GPU Computing Toolkit\CUDA\v8.0\nvvm\bin\nvvm64_31_0.dll'
os.environ['NUMBAPRO_LIBDEVICE']=r'C:\NVIDIA GPU Computing Toolkit\CUDA\v8.0\nvvm\libdevice'