commit | d8d140764c40c328537e374e9143ca6bab8f5a14 | [log] [tgz] |
---|---|---|
author | Han-Chung Wang <hanhan0912@gmail.com> | Wed Jul 31 10:43:49 2024 -0700 |
committer | GitHub <noreply@github.com> | Wed Jul 31 10:43:49 2024 -0700 |
tree | b15f645bd5c052389480f9c9f7c0d79aa0d5a896 | |
parent | 91433fc05de4500d05a88f4caf76e753ceddf0fc [diff] |
[GPU] Fix offsets calculation formula in MultiMmaOp distribution. (#18055) It was ``` vtid: virtual thread id tid: lane id vtid = (tid floordiv stride_i) mod size_i ``` However, it does not take `element` into account. Each thread grabs `element` contiguous data, so the vtid needs to be multiplied by `element` to get the next bunch of data. I.e., it becomes ``` vtid: virtual thread id tid: lane id vtid = ((tid floordiv stride_i) mod size_i) * element_i ``` Fixes https://github.com/iree-org/iree/issues/17973 --------- Signed-off-by: hanhanW <hanhan0912@gmail.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!
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.