MIOpen `Group-Xdlops` solvers segfaulting on gfx942
Open
Nobody has claimed this yet.
- Dominant language
- Julia
- Stars
- 344
- Forks
- 79
- Avg merge
- 2d 23h
- Merged PRs (30d)
- 27
Description
While MIOpen works on CDNA3 arch, there is a workaround needed for MIOpen CK Group-Xdlops solvers segfaulting on gfx942 (MI300). This could be removed (together with test skip in conv) once fixed upstream: https://github.com/ROCm/rocm-libraries/issues/9088
MIOPEN_DEBUG_CONV_IMPLICIT_GEMM_ASM_FWD_GTC_XDLOPS_NHWC: "0"
MIOPEN_DEBUG_GROUP_CONV_IMPLICIT_GEMM_HIP_FWD_XDLOPS: "0"
MIOPEN_DEBUG_GROUP_CONV_IMPLICIT_GEMM_HIP_WRW_XDLOPS: "0"
MIOPEN_DEBUG_3D_CONV_IMPLICIT_GEMM_HIP_FWD_XDLOPS: "0"
MIOPEN_DEBUG_3D_CONV_IMPLICIT_GEMM_HIP_WRW_XDLOPS: "0"
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
Review the workaround and convolution test skip introduced around JuliaGPU/AMDGPU.jl pull request #956, then check upstream rocm-libraries issue #9088. The work is done when the upstream Group-Xdlops gfx942 segfault is fixed and the listed MIOpen debug overrides and related test skip can be removed safely.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- machine-learning
- Issue type
- Bug
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100