commit | 2401be20d63aa2abb9ebf2a5a763a8384d10c205 | [log] [tgz] |
---|---|---|
author | Han-Chung Wang <hanhan0912@gmail.com> | Wed Jun 26 14:21:30 2024 -0700 |
committer | GitHub <noreply@github.com> | Wed Jun 26 14:21:30 2024 -0700 |
tree | 26302a7137b8ad25535d877498cf05bdf0405a55 | |
parent | 9da0309b0491df57629a2177ab1dbec4aa73ae6e [diff] |
[CPU] Tile outer parallel dims with 1 before lowering to ukernels. (#17731) The revision drops the support of "cache level tiling" because 1. Nobody is actively developing the path. 2. The dummy config is set which is doing nothing. 3. It is causing maintenance burden when we're developing new features. The new pipeline is: 1. Distribute mmt4d ops 2. Tile and fuse ops along parallel dims 3. Convert the mmt4d ops to ukernel ops. 4. Tile reduction dims if there are no ukernels for the mmt4d ops. 5. The rest is still the same. Fixes https://github.com/iree-org/iree/issues/17717 --------- Signed-off-by: hanhanW <hanhan0912@gmail.com>
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.
Community meeting recordings: IREE YouTube channel
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.