[Issue]: MIGRAPHX_MLIR_USE_SPECIFIC_OPS with where operator and boolean inputs + slowdown
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 333
- Forks
- 150
- Avg merge
- 4d 19h
- Merged PRs (30d)
- 54
Description
Problem Description
Running a test model as part of : https://github.com/ROCm/AMDMIGraphX/pull/3816
I seem to see a large slowdown when using the where operator. Printed out using --python
attention_no_where.log
attention_with_where.log
If I don't add the int8 convert before I see errors with the where operator when using the attention MLIR flag
Operating System
Ubuntu
CPU
Threadripper
GPU
AMD Instinct MI250
Other
No response
ROCm Version
ROCm 6.0.0
Steps to Reproduce
No response
(Optional for Linux users) Output of /opt/rocm/bin/rocminfo --support
No response
Additional Information
No response
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the attention_no_where.log, attention_with_where.log, and attention_no_int8_convert.log attachments, comparing the --python output for the where operator under MIGRAPHX_MLIR_USE_SPECIFIC_OPS. Reproduce the model from PR 3816 on Ubuntu with ROCm 6.0.0 and the MI250, then document or fix the boolean-input errors and slowdown so the affected runs behave as expected.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cpp
- Domain
- performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100