Skip to content
Branch: master
Find file History
ayzhuang and diyessi Create mkldnn primitives at first iteration for codegen - part2 (#2859)
* Create mkldnn primitives at first iteration for CODEGEN.

 OPs: add, lstm, and rnn.

*  OPs: batchnorm.

*  OPs: concat and lrn.

Remove dead code.

* Skip in place concat, relu, reshape, and slice when building node_primitive_string_deps_index map.

* Change NGRAPH_ASSERT to NGRAPH_CHECK.

* Address PR Feedback.

* Create mkldnn primitives at first iteration for CODEGEN.
 OPs: convertlayout, relu, leakyrelu, boundedrelu, sigmoid, softmax, slice.

* Fix bugs.

*  OPs: quantizedconcat.

Check if there are descriptors before emitting code to read desc_file.

*  OPs: convolution backward.

Use macro to write mkldnn memory dims to generated file.

*  OPs: MaxPoolWithIndices and MaxPoolWithIndicesBackprop.

Add unit tests for MaxPoolWithIndices, MaxPoolWithIndicesBackprop, and MaxPoolBackprop.

* Fix style error.

*  OPs: AvgPoolBackprop and MaxPoolBackprop.

Add unit test for AvgPoolBackprop.

*  OPs: DeconvolutionBias.

*  OPs: Quantize and Dequantize.

*  OPs: QuantizedDot and QuantizedDotBias.

* Use reference kernel for QuantizedConvolution for CODEGEN when mkldnn does not support the parameter types.
Get scales for quantization ops in cpu_emitter.

* Fix Windows build error: add CPU_BACKEND_API.

* Use template for quantization ops.

*  OPs: QuantizedMatmul.

Emit referece kernel for QuantizedDot in CODEGEN.

* Remove QuantizedDot from get_scale_index.

* Address PR feedback.
Latest commit 9335e41 May 21, 2019
Permalink
Type Name Latest commit message Commit time
..
Failed to load latest commit information.
cpu Create mkldnn primitives at first iteration for codegen - part2 (#2859) May 21, 2019
dynamic Address Klocwork issue. Fix bug? (#2915) May 14, 2019
generic_cpu Workaround for #pragma GCC diagnostic error. (Another version) (#2945) May 17, 2019
gpu Add support for more types in gather op. (#2926) May 17, 2019
gpuh Create backend as shared_ptr (#2793) May 8, 2019
hybrid Fix python stuff on Scott's PR (#2843) May 1, 2019
intelgpu Workaround for #pragma GCC diagnostic error. (Another version) (#2945) May 17, 2019
interpreter Workaround for #pragma GCC diagnostic error. (Another version) (#2945) May 17, 2019
nop Create backend as shared_ptr (#2793) May 8, 2019
plaidml Remove parent from PlaidML tensor initializer (#2923) May 21, 2019
reference Generic Dot & builders for QLinearMatmul (#2857) May 17, 2019
CMakeLists.txt Function call working (#2472) Feb 27, 2019
aligned_buffer.cpp Add support for move semantics to AlignedBuffer (#2956) May 21, 2019
aligned_buffer.hpp Add support for move semantics to AlignedBuffer (#2956) May 21, 2019
backend.cpp Remove runtime/cpu includes from things outside runtime/cpu (#2906) May 13, 2019
backend.hpp Dynamic-shape backend API change, with prototype implementation (#2844) May 10, 2019
backend_manager.cpp Add more info to error messages (#2894) May 10, 2019
backend_manager.hpp Create backend as shared_ptr (#2793) May 8, 2019
executable.cpp Dynamic-shape backend API change, with prototype implementation (#2844) May 10, 2019
executable.hpp Split runtime backend and executable source files (#2544) Mar 5, 2019
host_tensor.cpp Remove Backend parent from runtime::Tensor (#2810) Apr 27, 2019
host_tensor.hpp Remove Backend parent from runtime::Tensor (#2810) Apr 27, 2019
performance_counter.hpp PerformanceCounter rework (#2790) Apr 27, 2019
tensor.cpp Dynamic-shape backend API change, with prototype implementation (#2844) May 10, 2019
tensor.hpp Dynamic-shape backend API change, with prototype implementation (#2844) May 10, 2019
You can’t perform that action at this time.