[GitHub] [incubator-mxnet] DickJC123 commented on issue #15551: Bypass cuda/cudnn checks if no driver.

[GitBox <gi...@apache.org>]

DickJC123 commented on issue #15551: Bypass cuda/cudnn checks if no driver.
URL: https://github.com/apache/incubator-mxnet/pull/15551#issuecomment-512545023
 
 
   I went back and confirmed that the warnings are properly generated, although now they appear at the time of first use of a gpu context:
   ```
   $ python
   >>> import mxnet as mx
   >>> ctx = mx.gpu(0)
   >>> x = mx.nd.array([1,2,3], ctx=ctx)
   [19:46:59] src/base.cc:51: Upgrade advisory: this mxnet has been built against cuda library version 9000, which is older than the oldest version tested by CI (10000).  Set MXNET_CUDA_LIB_CHECKING=0 to quiet this warning.
   [19:46:59] src/base.cc:84: Upgrade advisory: this mxnet has been built against cuDNN lib version 7102, which is older than the oldest version tested by CI (7600).  Set MXNET_CUDNN_LIB_CHECKING=0 to quiet this warning.
   >>> y = mx.nd.array([4,5,6], ctx=mx.gpu(1))
   >>> exit()
   
   ```
   Moving the libmxnet.so to a system with a different cudnn version generates the additional warning at the same point as shown above:
   ```
   [19:50:10] src/base.cc:80: cuDNN lib mismatch: linked-against version 7201 != compiled-against version 7102.  Set MXNET_CUDNN_LIB_CHECKING=0 to quiet this warning.
   ```
   I've also verified that the env vars can be used to eliminate these warnings.
   I feel this PR should be merged when it passes CI.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services