xla
xla copied to clipboard
openxla
A machine learning compiler for GPUs, CPUs, and ML accelerators
The issue was introduced here -> https://github.com/openxla/xla/commit/b30b420510a89caa6ce9ae12c16add7e416e9034#diff-923ec2612240244a60ae28c17f798a7daeccbab84a8e9de9efd7ba96c311c763L115 Some of the ROCm files need `dso_loader.h`.
With the existing `enable_mock_nccl` setting it is impossible to warm up compilation cache when there are multiple processes per node. This is because the cache key includes topology and GPU...
Integrate LLVM at llvm/llvm-project@6292f117c39b Updates LLVM usage to match [6292f117c39b](https://github.com/llvm/llvm-project/commit/6292f117c39b)
![copybara-service[bot] avatar](https://avatars.githubusercontent.com/in/44061?v=4 "copybara-service[bot] avatar"){kind=avatar}
This PR https://github.com/openxla/xla/pull/11880 created a latent bug on ROCM side which was really hard to track. Due to [gemm_algorithm_picker](https://github.com/ROCm/xla/blob/58cd0e78dc19075e7c935d7cdb31676ce868e64c/xla/service/gpu/autotuning/gemm_algorithm_picker.cc#L299), the problem occurs only for non-zero beta when the output matrix...
[HLO Componentization] Create hlo/builder sub-component (Phase II). This CL takes care of 1. Migrating external projects dependencies from xla/client --> xla/hlo/builder Phase I takes care of 1. Migrating xla/translate -->...
[HLO Componentization] Create hlo/builder sub-component (Phase II). This CL takes care of 1. Migrating external projects dependencies from xla/client --> xla/hlo/builder Phase I takes care of 1. Migrating xla/translate -->...
[HLO Componentization] Create hlo/builder sub-component (Phase II). This CL takes care of 1. Migrating external projects dependencies from xla/client --> xla/hlo/builder Phase I takes care of 1. Migrating xla/translate -->...
Integrate LLVM at llvm/llvm-project@a5cd5d351ddb Updates LLVM usage to match [a5cd5d351ddb](https://github.com/llvm/llvm-project/commit/a5cd5d351ddb)
Allow simple host call that uses host tensor as parameter/result.