Introduce Flow InferNumericNarrow, OptimizeNumerics and CleanupNumericNarrowing passes (#7975) Together, these passes: * Infer where it is safe to use a narrower, low precision type in place of floating point arithmetic. * Use this information to do targeted rewrites of linalg math ops. * Propagate casts. When combined with constant hoisting and eval, this will perform most of the work needed for low-precision optimization of inference workloads. Also adds a new Util::NumericCastOpInterface and applies it to the casting ops in the arith dialect. I'm going to put some more mileage on this and then will likely propose it for upstream.
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.