JuliaGPU / JuliaGPU/KernelAbstractions.jl

StackOverflow for `get_device` on `ROCArray`

Open
#325 2 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

AMD bug
Dominant language
Julia
Stars
523
Forks
88
Avg merge
1d 11h
Merged PRs (30d)
25

Description

julia> a = ROCArray(rand(10,10));

julia> KernelAbstractions.get_device(a)
ERROR: StackOverflowError:
Stacktrace:
 [1] get_device(A::ROCMatrix{Float64}) (repeats 79984 times)
   @ KernelAbstractions ~/.julia/packages/KernelAbstractions/DqITC/src/KernelAbstractions.jl:355

Calling AMDGPU.device(a) works:

julia> AMDGPU.device(a)
GPU-XX [AMD Radeon RX 6700 XT (gfx1030)]

So maybe there's a way to get rid of one layer of abstraction here?

As a note:

julia> b = Array(rand(10,10));

julia> KernelAbstractions.get_device(b)
CPU()

So shouldn't we be returning ROCdevice() instead of GPU-XX [AMD Radeon RX 6700 XT (gfx1030)]? Or was there something about the recent change with how ROCDevice() that changes things?

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Reproduce the ROCArray example and inspect get_device at src/KernelAbstractions.jl:355, comparing its dispatch with AMDGPU.device(a) and the working Array case. Determine the intended device abstraction for ROCArray, then verify that get_device no longer recurses and returns the expected device type.

Written by the indexing model from the issue text.

Assessment

Tech stack
julia
Domain
hpc
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.