`ukernel.generic_raw` op passing only buffer and offset for memref operands (#13053) This is less opinionated than the existing `ukernel.generic` in that `memref` operands are lowered to just base buffer and offset operands, the bare minimum for any callee to be able to do anything with the memref. By not appending stride values, this leaves it up to each use case to decide which strides to pass. This makes `ukernel.generic_raw` flexible enough to replace all other existing ukernel ops. Though we want to retain `ukernel.generic` for now, we will at least proceed with migrating `ukernel.mmt4d` and all the ukernels in VMVXOps.td. This PR also makes the `offset` be passed along with the buffer in all cases, including the rank-0 case where it was omitted (that was #13054, now folded into this). This PR also does some naming harmonization: `micro_kernel` -> `ukernel`, `foo_ukernel` -> `ukernel_foo`.
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.