commit | 9bf6fb1bcd6d5bc0d61c0a736e2555ad2745207b | [log] [tgz] |
---|---|---|
author | Okwan Kwon <okwan@google.com> | Mon May 15 19:32:07 2023 -0700 |
committer | GitHub <noreply@github.com> | Mon May 15 19:32:07 2023 -0700 |
tree | 5bb78ec1fc7daba154ae748aedf2c52cd7aae383 | |
parent | 747b309fd7b353bc8c6f14e8c773cf15d83c2f74 [diff] |
Support `num_replicas` and `num_partitions` (#13288) Support `num_replicas` and `num_partitions` through cross_replica and cross_partition for all_reduce, all_gather, and reduce_scatter, replica_id, and partition_id. Based on the channel ID and use_global_device_ids, the communication mode can be different in the 2D grid. See https://github.com/openxla/stablehlo/blob/main/docs/spec.md#collective-ops for more information. Note that stablehlo still uses `mhlo.num_replicas` and `mhlo.num_partitions` to embed the info in the module.
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.