Kunwar Grover

Added support for hosting using docker

1

Docker allows for a smoother and easier hosting experience. To run using docker: 1. Edit docker-compose.yml and update the environment variables. 2. Run using: `docker-compose up`

This issue tracks progress on vector distribution framework. ```[tasklist] ### Layouts - [ ] Add warp enum to LayoutV1 - [ ] https://github.com/openxla/iree/pull/16368 ``` ```[tasklist] ### Distribution Patterns - [...

Generalize overriding llvm func attr flags in translation info

2

Previously, we were specifically querying for waves_per_eu attr in translation info in ROCMTarget. This patch makes this general by attaching a llvm function attribute override dictionary in the translation info...

[VectorDistribution] Add iree_vector_ext.subgroup/thread_ids to make distribution opaque

3

This patch adds two new operations: `iree_vector_ext.subgroup_ids` and `iree_vector_ext.thread_ids`. These operations simply decouple vector distribution from the actual thread indexing and hide them behind another operation. This makes distribution patterns...

[NestedLayout] Make layout use strides instead of basis

1

This patch makes the layout store how threads and subgroups are distributed as "strides" instead of basis. Strides are just a mapping from `virtual tid --> tid` (where tid is...

iree-e2e-matmul-test duplicate flags

4

### What happened? On running: ``` /tools/testing/e2e/iree-e2e-matmul-test --help ``` there are duplicate flag options for test_utils.h: ``` # ===----------------------------------------------------------------------=== # Flags in iree/tools/testing/e2e/test_utils.h # ===----------------------------------------------------------------------=== # Requires floating point result...

This issue tracks tasks to support attention op in GPU codegen targeting tensor cores properly. Initial proof of concept PR that works e2e for gpus with MFMA: https://github.com/iree-org/iree/pull/17212 Short Term...