commit | f06c804f9ec79e16da7dcee1af81afd7b95a8bb1 | [log] [tgz] |
---|---|---|
author | Scott Todd <scotttodd@google.com> | Thu Nov 04 08:09:46 2021 -0700 |
committer | GitHub <noreply@github.com> | Thu Nov 04 08:09:46 2021 -0700 |
tree | 03832448cdcd7e0a269157b85ef40ea24de56408 | |
parent | 4395a123bfc709ff2c5203b5c9407a43f7ee5c26 [diff] |
Add statistics to Util/Transforms/FoldGlobals.cpp. (#7535) Following up on https://github.com/google/iree/pull/7312#discussion_r728261108, this adds more [statistics](https://mlir.llvm.org/docs/PassManagement/#pass-statistics) to IREE's compiler passes. I personally build with the CMake option `-DLLVM_FORCE_ENABLE_STATS=ON` so the `-pass-statistics` flag works in `RelWithDebInfo` builds, but we don't yet have anything like that configured on our CI or recommended in our developer documentation. See also https://github.com/google/iree/issues/6161#issuecomment-912635307 for how this could be useful. Sample output on [bert_encoder_unrolled_fake_weights.mlir](https://github.com/google/iree/blob/main/iree/test/e2e/models/bert_encoder_unrolled_fake_weights.mlir) (fake weights are folded aggressively 😛): ``` mlir::iree_compiler::IREE::Util::`anonymous-namespace'::FoldGlobalsPass (S) 1113 global ops before folding - Number of util.global ops before folding (S) 7 global ops after folding - Number of util.global ops after folding ```
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.