iree
iree copied to clipboard
iree-org
A retargetable MLIR-based machine learning compiler and runtime toolkit.
We depend on a few files hosted in a GCP project using various buckets. Most uses can be discovered in this repo with a regex search of `https://storage\.googleapis\.com.*iree`: ``` 21...
The current reshape propagation patterns upstream bail on multi-result operations, but the implementation seems to support such cases as far as I can tell: https://github.com/llvm/llvm-project/blob/3733528e521b7ee6af3950c65c3ff421c8fd0af6/mlir/lib/Dialect/Linalg/Transforms/ElementwiseOpFusion.cpp#L1253-L1258 This gist is an example...
**Initial V0 Goal** Main Requirements: - Needs to be able to be run locally by developer easily (don't require a bunch of unnecessary requirements such as RocBLAS, HipBLASLt , etc.)...
### What happened? During decode phase in sharktank paged LLM, there is a case where we expand shapes on one of attention's reduction dim/K2. When fused, this causes attention pipeline...
Splitting https://github.com/iree-org/iree/pull/18519 into four patches. Depends https://github.com/iree-org/iree/pull/18784 and #18800. This is the third patch, including changes along the VectorDistribute pipeline to support reduction operations. Additionally, a relevant test has been...
Progress on https://github.com/iree-org/iree-test-suites/issues/2 and https://github.com/iree-org/iree-test-suites/issues/6 .
Dispatch count is very sensitive to the placement of `tensor.expand_shape` and `tensor.collapse_shape` ops which often shows up as regressions in changes that shouldn't have a negative impact on dispatch count....
Similar to https://github.com/iree-org/iree/issues/18814. We used to have a mac mini running macOS builds/tests, including of the Metal HAL. Right now we use [standard GitHub-hosted runners](https://docs.github.com/en/actions/using-github-hosted-runners/using-github-hosted-runners/about-github-hosted-runners#standard-github-hosted-runners-for-public-repositories) for macOS, which includes arm...
We currently run the [`.github/workflows/ci_windows_x64_msvc.yml`](https://github.com/iree-org/iree/blob/main/.github/workflows/ci_windows_x64_msvc.yml) workflow on a nightly schedule using [standard GitHub-hosted runners](https://docs.github.com/en/actions/using-github-hosted-runners/using-github-hosted-runners/about-github-hosted-runners#standard-github-hosted-runners-for-public-repositories) (currently `windows-2022` with 4 CPU cores, 16 GB of RAM, and 14GB of SSD). Looking at...
We used to have some Linux runners with NVIDIA T4 and A100 GPUs on the GCP runner cluster. Our new Azure runner cluster currently only has Linux CPU runners. We...