| commit | 71c22dadbae930cd6585940b75d0aa44b0a8ec12 | [log] [tgz] |
|---|---|---|
| author | bjacob <benoitjacob@google.com> | Fri Oct 20 08:14:34 2023 -0400 |
| committer | GitHub <noreply@github.com> | Fri Oct 20 08:14:34 2023 -0400 |
| tree | 493d610bfbee70e0b1ceb91a0924797268429dbe | |
| parent | 5a20dce23516316eb57a9f67a1564768ce8fb70c [diff] |
Improvements to e2e matmul tests (#15243) * Test f16 on CPU, not just on GPU. * Test bf16 * Test many SIMD variants that we now have dedicated ukernel code paths for. These code paths were so far tested only in isolated unit tests, not in integration. * Rationalize/reorganize the collection of e2e matmul tests: - Drop no longer relevant cases. - Rename: `mmt4d` -> `dt` (for data-tiling), `direct` -> `nondt`, etc. The `nondt` tests explicitly disable data-tiling, so they will continue testing that through the merging of #15215. - Group by target backend.
IREE (Intermediate Representation Execution Environment, pronounced as “eerie”) is an MLIR-based end-to-end compiler and runtime that lowers Machine Learning (ML) models to a unified IR that scales up to meet the needs of the datacenter and down to satisfy the constraints and special considerations of mobile and edge deployments.
See our website for project details, user guides, and instructions on building from source.
IREE is still in its early phase. We have settled down on the overarching infrastructure and are actively improving various software components as well as project logistics. It is still quite far from ready for everyday use and is made available without any support at the moment. With that said, we welcome any kind of feedback on any communication channels!
See our website for more information.
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.