cuDNN Release Notes v7.1.4
cuDNN Release Notes v7.1.4 (PDF)
Key Features and Enhancements
The following enhancements have been added to this release:
- Improved performance for some cases of data-gradient convolutions and maxpooling. This is expected to improve performance of ResNet-50 like networks.
- The runtime of the RNN Find algorithm suite is improved in v7.1.4 resulting in slightly improved runtime of
Following are known issues in this release:
cudnnGetpicks a slow algorithm that does not use Tensor Cores on Volta when inputs are FP16 and it is possible to do so.
cudnnConvolutionBackwardFilter()function may output incorrect results for
CUDNN_CONVOLUTION_BWD_FILTER_ALGO_FFT_TILINGwhen the convolution mode is
CUDNN_CONVOLUTION. This function should not be used in this mode.
The following issues have been fixed in this release:
cudnnAddTensorNdmight cause a segmentation fault if called with bad arguments (e.g. null pointer), this issue is in 7.1.3 only and fixed in 7.1.4.
cudnnRNNBackwardDataLSTM cell with fp16 (half) inputs might generate wrong values (silently), this issue exists in cudnn 7.1.3 binaries compiled with cuda toolkit 9.0 and toolkit cuda 9.2, and does not exist in cudnn 7.1.3 binaries compiled with toolkit 9.1.
cudnnGetRNNLinLayerMatrixParamswrongly returns CUDNN_STATUS_BAD_PARAM when
cudnnSetRNNDescriptoris called with dataType == CUDNN_DATA_FLOAT. This is an issue in 7.1.3 only and will be fixed in 7.1.4. The dataType argument as of today supports only
CUDNN_DATA_FLOATand we plan to support additional compute types in the future.
- There is a small memory leak issue when calling
CUDNN_RNN_ALGO_STANDARD. This issue also affects previous cuDNN v7 releases. This is fixed in 7.1.4.
- RNN with half precision returns
CUDNN_EXECUTION_FAILEDon Kepler gpu in 7.1.3. This is fixed in 7.1.4 to use pseudo-fp16 computation
- The RNN Find algorithm suite mistakenly did not test
CUDNN_RNN_ALGO_PERSIST_DYNAMICkernels with tensor operations enabled when it was possible to do so. This is fixed in v7.1.4.