| commit | 04b7922787aba321ea04423f6080e895de1b7d31 | [log] [tgz] |
|---|---|---|
| author | Geoffrey Martin-Noble <gcmn@google.com> | Thu Sep 09 15:19:35 2021 -0700 |
| committer | GitHub <noreply@github.com> | Thu Sep 09 15:19:35 2021 -0700 |
| tree | 90553bf1430d4a42a9b2a10e0f2ae6a494efff8a | |
| parent | 76ebae70c534ff779d86910fae61b98642c0da67 [diff] |
Extend linalg_ext lowering to handle signedness and more ops (#6987) MHLO ops use signed (although written as signless for historical reasons) and unsigned ops. The upstream MHLO->Linalg already handles this and we need to handle it here as well. In addition, this changes the approach to dealing with detensorizing, relying more heavily on the dialect conversion framework type converter. This leaves things as unrealized_conversion_cast ops, to be cleaned up later after lowering the rest of MHLO. We could continue to use "real" ops for [de]tensoring, but we need unrealized_conversion_cast for signedness conversions anyway, so I think it's cleaner to use it everywhere. It's perhaps unfortunate because really the correct cast there *is* a bitcast, but like all std ops, std.bitcast does not support [un]signed integers. Finally, it adds a bunch of ops for lowering within linalg_ext regions. We should be handling everything that lowers to standard naturally, not picking the handful of ops we've run into so far. Fixes https://github.com/google/iree/issues/6154
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!
| CI System | Build System | Platform | Architecture | Configuration / Component | Status |
|---|---|---|---|---|---|
| Kokoro | Bazel | Linux | x86-64 |  | |
| Kokoro | CMake & Bazel | Linux | x86-64 (swiftshader) | Integrations |  |
| Kokoro | CMake & Bazel | Linux | x86-64 (turing) | Integrations |  |
| Kokoro | CMake | Linux | x86-64 (swiftshader) |  | |
| Kokoro | CMake | Linux | x86-64 (swiftshader) | asan |  |
| Kokoro | CMake | Linux | x86-64 (turing) |  | |
| Kokoro | CMake | Android | arm64-v8a | Runtime (build only) |  |
| Kokoro | CMake | Bare Metal | risc-v-32 | Runtime |  |
| Kokoro | CMake | Linux | risc-v-64 | Runtime |  |
| Buildkite | CMake | Android | arm64-v8a | Runtime |  |
| BuildKite | CMake | Android | arm64-v8a | Runtime Benchmarks |  |
| BuildKite | CMake | Linux | x86-64 | Tracing + Standalone Runtime |  |
IREE is licensed under the terms of the Apache 2.0 License with LLVM Exceptions. See LICENSE for more information.