Memory descriptor. More...
#include <mkldnn_types.h>
Collaboration diagram for mkldnn_memory_desc_t:
Public Attributes | |
| mkldnn_primitive_kind_t | primitive_kind |
| The kind of primitive. More... | |
| int | ndims |
| Number of dimensions. More... | |
| mkldnn_dims_t | dims |
| Dimensions in the following order: More... | |
| mkldnn_data_type_t | data_type |
| Data type of the tensor elements. More... | |
| mkldnn_memory_format_t | format |
| Memory format. More... | |
| union { | |
| mkldnn_blocking_desc_t blocking | |
| Description of the data layout for memory formats that use blocking. More... | |
| mkldnn_wino_desc_t wino_desc | |
| Tensor of weights for integer 8bit winograd convolution. More... | |
| mkldnn_rnn_packed_desc_t rnn_packed_desc | |
| Tensor of packed weights for RNN. More... | |
| } | layout_desc |
Detailed Description
Memory descriptor.
The description is based on a number of dimensions, dimensions themselves, plus information about elements type and memory format. Additionally, contains format-specific descriptions of the data layout.
Member Data Documentation
◆ primitive_kind
| mkldnn_primitive_kind_t mkldnn_memory_desc_t::primitive_kind |
The kind of primitive.
Used for self identifying the primitive descriptor. Must be mkldnn_memory.
◆ ndims
| int mkldnn_memory_desc_t::ndims |
Number of dimensions.
◆ dims
| mkldnn_dims_t mkldnn_memory_desc_t::dims |
Dimensions in the following order:
- CNN data tensors: mini-batch, channel, spatial (
{N, C, [[D,] H,] W}) - CNN weight tensors: group (optional), output channel, input channel, spatial (
{[G,] O, I, [[D,] H,] W}) - RNN data tensors: time, mini-batch, channels (
{T, N, C}) or layers, directions, states, mini-batch, channels ({L, D, S, N, C}) - RNN weight tensor: layers, directions, input channel, gates, output channels (
{L, D, I, G, O}).
- Note
- The order of dimensions does not depend on the memory format, so no matter whether the data is laid in mkldnn_nchw or mkldnn_nhwc the dims for 4D CN data tensor would be
{N, C, H, W}
◆ data_type
| mkldnn_data_type_t mkldnn_memory_desc_t::data_type |
Data type of the tensor elements.
◆ format
| mkldnn_memory_format_t mkldnn_memory_desc_t::format |
Memory format.
◆ blocking
| mkldnn_blocking_desc_t mkldnn_memory_desc_t::blocking |
Description of the data layout for memory formats that use blocking.
◆ wino_desc
| mkldnn_wino_desc_t mkldnn_memory_desc_t::wino_desc |
Tensor of weights for integer 8bit winograd convolution.
◆ rnn_packed_desc
| mkldnn_rnn_packed_desc_t mkldnn_memory_desc_t::rnn_packed_desc |
Tensor of packed weights for RNN.
◆ layout_desc
| union { ... } mkldnn_memory_desc_t::layout_desc |
The documentation for this struct was generated from the following file:
- include/mkldnn_types.h
