Module cudarc::cudnn::result

A thin wrapper around sys providing Results with CudnnError.

Structs

• CudnnError

Functions

• convolution_backward_data^⚠
  Launch the backward data kernel. See nvidia docs.
• convolution_backward_filter^⚠
  Launch the backward data kernel. See nvidia docs.
• convolution_forward^⚠
  Launch the conv forward kernel. See nvidia docs.
• create_convolution_descriptor
  Allocates a convolution descriptor. See nvidia docs.
• create_filter_descriptor
  Creates a filter descriptor. See nvidia docs
• create_handle
  Creates a handle to the cuDNN library. See nvidia docs
• create_reduce_tensor_descriptor
  See nvidia docs.
• create_tensor_descriptor
  Allocates a new tensor descriptor. See nvidia docs
• destroy_convolution_descriptor^⚠
  Destroys a descriptor. See nvidia docs.
• destroy_filter_descriptor^⚠
  Destroys a filter descriptor. See nvidia docs
• destroy_handle^⚠
  Destroys a handle previously created with create_handle(). See nvidia docs
• destroy_reduce_tensor_descriptor^⚠
  See nvidia docs.
• destroy_tensor_descriptor^⚠
  Destroys a tensor descriptor. See nvidia docs
• get_convolution_backward_data_algorithm^⚠
  See nvidia docs
• get_convolution_backward_data_workspace_size^⚠
  Returns size in bytes. See nvidia docs
• get_convolution_backward_filter_algorithm^⚠
  See nvidia docs
• get_convolution_backward_filter_workspace_size^⚠
  Returns size in bytes. See nvidia docs
• get_convolution_forward_algorithm^⚠
  See nvidia docs
• get_convolution_forward_workspace_size^⚠
  Returns size in bytes. See nvidia docs
• get_cudart_version
  The same version of a given cuDNN library can be compiled against different CUDA toolkit versions. This routine returns the CUDA toolkit version that the currently used cuDNN library has been compiled against.
• get_reduction_indices_size^⚠
  See nvidia docs
• get_reduction_workspace_size^⚠
  See nvidia docs
• get_version
  This function returns the version number of the cuDNN library. It returns the CUDNN_VERSION defined present in the cudnn.h header file.
• reduce_tensor^⚠
  See nvidia docs
• set_convolution2d_descriptor^⚠
  Sets data on a conv descriptor. See nvidia docs
• set_convolution_group_count^⚠
  See nvidia docs
• set_convolution_math_type^⚠
  See nvidia docs.
• set_convolutionnd_descriptor^⚠
  Sets data on a conv descriptor. See nvidia docs
• set_filter4d_descriptor^⚠
  Sets data on a pre allocated filter descriptor. See nvidia docs
• set_filternd_descriptor^⚠
  Sets data on a pre allocated filter descriptor. See nvidia docs
• set_reduce_tensor_descriptor^⚠
  See nvidia docs
• set_stream^⚠
  Sets the stream cuDNN will use. See nvidia docs
• set_tensor4d_descriptor^⚠
  Sets data on a tensor descriptor. See nvidia docs
• set_tensor4d_descriptor_ex^⚠
  Sets data on a tensor descriptor. See nvidia docs
• set_tensornd_descriptor^⚠
  Sets data on a tensor descriptor. See nvidia docs
• version_check
  Runs all *VersionCheck functions.

Type Aliases

• CudnnResult