[PATCH] D146483: Add infrastructure to use EmitC-generated models for inlining.

[Jacob Hegna via Phabricator via llvm-commits]

jacobhegna updated this revision to Diff 517635.
jacobhegna added a comment.

Address comments from reviewers. The main changes are:

1. Include the python files to generate the EmitC models in this patch, instead

of in the mlcompileropt github repo.

2. Remove the global declaration of the emitc::Tensor type and make it an

implementation detail. This requires changing the signatures of the generated
models to speak in terms of T* instead of Tensor<T, ...>.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D146483/new/

https://reviews.llvm.org/D146483

Files:
  llvm/include/llvm/Analysis/EmitCModelRegistry.h
  llvm/include/llvm/Analysis/EmitCTensor.h
  llvm/include/llvm/Analysis/InlineAdvisor.h
  llvm/include/llvm/Analysis/InlineModelFeatureMaps.h
  llvm/include/llvm/Analysis/MLInlineEmitCModel.h
  llvm/include/llvm/Analysis/MLModelRunner.h
  llvm/lib/Analysis/CMakeLists.txt
  llvm/lib/Analysis/InlineAdvisor.cpp
  llvm/lib/Analysis/MLInlineAdvisor.cpp
  llvm/lib/Analysis/MLInlinerEmitCRunner.h
  llvm/lib/Analysis/models/CMakeLists.txt
  llvm/lib/Analysis/models/emitc/CMakeLists.txt
  llvm/lib/Analysis/models/emitc/InlineOzTestModel.emitc.cpp
  llvm/lib/Analysis/models/emitc/InlineOzTestModel.emitc.h
  llvm/lib/Analysis/models/tflite_to_cpp.py
  llvm/lib/Analysis/models/tflite_to_cpp_lib.py
  llvm/lib/Passes/PassBuilderPipelines.cpp
  llvm/test/Transforms/Inline/ML/ml-test-emitc-mode.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D146483.517635.patch
Type: text/x-patch
Size: 111005 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20230427/628b8929/attachment.bin>