commit | a3b74bcc7bd29958c8188e3b79d7000de3427786 | [log] [tgz] |
---|---|---|
author | Benjamin Maxwell <benjamin.maxwell@arm.com> | Fri May 17 18:15:38 2024 +0100 |
committer | GitHub <noreply@github.com> | Fri May 17 10:15:38 2024 -0700 |
tree | 3eda6da8dce6a674dbb7cf518b68858e4e92c625 | |
parent | 4132d2e29a416b444b4f2c801f22b83b052d1a06 [diff] |
[CPU][ArmSME] Update tiling to use all SME accumulators (#16389) Previously, we only tiled for a single SME accumulator. This patch updates the lowering_config to make use of all SME accumulators. This is done by increasing the tile size to [8]x[8] for f32 and to [4]x[8] for f64. This lowers to four [4]x[4] 32-bit accumulators and eight [2]x[2] 64-bit accumulators respectively. Signed-off-by: Benjamin Maxwell <benjamin.maxwell@arm.com> ci-extra: build_test_all_arm64 --------- Signed-off-by: Benjamin Maxwell <benjamin.maxwell@arm.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.