Stella Stamenova
Stella Stamenova
There are multiple impactful changes from mlir: - [LLVM] Update C++ standard to 17 (https://reviews.llvm.org/D130689) - [ADT] Deprecate Optional::{hasValue,getValue} (NFC) (https://reviews.llvm.org/D131349) - [mlir] Remove types from attributes (https://reviews.llvm.org/D130092) - [mlir]...
Right now there are several different types of tests in onnx-mlir: 1) The lit tests in test/mlir 2) The numerical tests in test/numerical 3) The backend tests in test/backend 4)...
llvm keeps track of a couple of triples in cmake for each compilation: `TARGET_TRIPLE` (which is set by using `LLVM_DEFAULT_TARGET_TRIPLE`, for example) & `LLVM_HOST_TRIPLE` (which is set by the llvm...
Please use https://github.com/llvm/torch-mlir/issues/1135 for process discussion. I'm adding the new issue, so that we can track just the green LLVM commits and corresponding commits/branches in torch-mlir, onnx-mlir and MHLO until...
Following up on a couple of conversations we've had about coordinating LLVM commits across projects (in [onnx-mlir](https://github.com/onnx/onnx-mlir/issues/1558) and in [LLVM](https://discourse.llvm.org/t/coordinate-llvm-commits-for-different-project/63990/64)), we had a follow up discussion in the Torch MLIR...
There's an ongoing experiment to keep mhlo, torch-mlir and onnx-mlir in sync with respect to their llvm commit. These projects historically have different update cadences and rarely converge on a...
Based on the discussion in the past couple of community meetings, I've added a schedule and instructions for the llvm updates in onnx-mlir: https://github.com/onnx/onnx-mlir/wiki/LLVM-Update-Schedule If you are able to contribute,...
There are several onnx_to_krnl tests that have started failing since #2441 (one was actually added later, but looks to be the same failure). They only fail in debug builds -...