[GitHub] reminisce commented on a change in pull request #10710: Add cuda version check to skip building quantization ops for versions less than 8.0

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

reminisce commented on a change in pull request #10710: Add cuda version check to skip building quantization ops for versions less than 8.0
URL: https://github.com/apache/incubator-mxnet/pull/10710#discussion_r184837659
 
 

 ##########
 File path: src/operator/quantization/quantized_conv.cu
 ##########
 @@ -270,7 +270,7 @@ void QuantizedConvForwardGPU(const nnvm::NodeAttrs& attrs,
   const ConvolutionParam& param = nnvm::get<ConvolutionParam>(attrs.parsed);
   CHECK_EQ(param.kernel.ndim(), 2U)
     << "QuantizedConvForward<gpu> only supports 2D convolution for now";
-#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
+#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6 && CUDA_VERSION >= 8000
 
 Review comment:
   There will be an error message printed saying that otherwise is not supported. I have made it clearer on specifying cudnn and cuda versions.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on 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