Performance gap on a 7-point stencil Laplacian kernel on Frontier MI250x GPUs
Nobody has claimed this yet.
- Dominant language
- Julia
- Stars
- 344
- Forks
- 79
- Avg merge
- 2d 19h
- Merged PRs (30d)
- 25
Description
In a recent study on Frontier, a 7-point stencil kernel under performs at half the bandwidth (~300 GB/s) of its HIP counterpart (~600 GB/s) on a single MI250x. The behavior is replicated at large scale up to 4K GPUs. While this was a first attempt using AMDGPU v0.4
Some to-do items:
- Test with AMDGPU.jl v0.5 and onwards
- Understand the performance difference wrt HIP 7-point stencil driven by a Laplacian operator available here
- Implement a performance test to avoid regression
Opening after discussion in the HPC call with @vchuravy and @gbaraldi
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 by testing the 7-point stencil with AMDGPU.jl v0.5 and later versions, then compare its performance with the HIP Laplacian reference in kernel1.hpp. Investigate the bandwidth difference on a single MI250x and at scale, and add a performance test whose results can detect regressions.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- hpc, performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100