Add AMDGPU pattern for chained matmuls (#16398) This PR adds a pattern for chained matmuls operators like Flash Attention where by swapping the operands of the sequential matmuls, one can keep the data in registers and avoid a trip to shared memory and/or additional shuffle instructions. The pattern swaps the operands of the contract op, and inserts transposes for the accumulator and result, if the contracts are chained and satisfy the MMT indexing maps.
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.