[CPU] Remove special distribution tile sizes from setting matmul config. (#15968) They were used for DoubleTilingPadExpert pipeline, and were specialized for MiniLMBert models. The pipeline was retired and we have been tracking other models today. We don't have to use the configs for all defaults. It does not regress MiniLMBert model. It saves 30 ms (25.35%↓) for MobileBert_int8 and 4 seconds (10.70%↓) for Falcon7bGptqPT. It regresses a little for DeepLabV3, but it improves other ~10 models. The regression only happens on single-threaded, and I think they should be fixed by cache-level tiling afterwards. We are likely sitting in a sweet spot because we mixed the context with distribution config.
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.