commit | e96e3c004011550c6b4ecc1520e0f022984c4b17 | [log] [tgz] |
---|---|---|
author | Stanley Winata <68087699+raikonenfnu@users.noreply.github.com> | Thu Oct 24 16:56:38 2024 -0700 |
committer | GitHub <noreply@github.com> | Thu Oct 24 16:56:38 2024 -0700 |
tree | 7789c2739d0ba2b030ccb617d06ad5d230567dee | |
parent | aef6e1fc8fe85f4a57bc06b0272ada18975494fb [diff] |
[VectorLayout] Fix insertion of new constOp for non dominate issue. (#18894) Main motivation of this patch is to resolve issue where we have the same constOp being used by multiple operations. But with a twist where first time the constOp needs a layout is on a op that topologically comes after other ops that use constOp. This will generate a copy of constOp in the location right before the latter op, which is problematic because this constOp will be used by other ops before it. Previously for the test added we get this error: ``` within split at contraction_layout.mlir:1 offset :24:10: note: see current operation: %9 = "arith.addf"(%8, %6) <{fastmath = #arith.fastmath<none>}> : (vector<96x64xf16>, vector<96x64xf16>) -> vector<96x64xf16> within split at contraction_layout.mlir:1 offset :22:19: error: operand #1 does not dominate this use %scaled_rhs = arith.mulf %read_1, %cst_1 : vector<96x64xf16> ``` While minor, this is also problematic because this error seem to stopped layout analysis (but not fatally) S.T it fails to vector distribute in some cases, making it hard to debug. Signed-off-by: Stanley Winata <stanley.winata@amd.com>
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
Package | Release status |
---|---|
GitHub release (stable) | |
GitHub release (nightly) | |
Python iree-compiler | |
Python iree-runtime |
Host platform | Build status |
---|---|
Linux | |
macOS | |
Windows |
For the full list of workflows see https://iree.dev/developers/general/github-actions/.
See our website for more information.
Community meeting recordings: IREE YouTube channel
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.