llnl / llnl/smith

`hipErrorInvalidDeviceFunction: invalid device function` on tioga/ rzvernal

Open
#1,586 2 comments 0 reactions 0 assignees View on GitHub
bug build gpu
Dominant language
C++
Stars
245
Forks
36
Avg merge
8d 23h
Merged PRs (30d)
4

Description

from a convo with @ebchin

>
> C++ exception with description "scan failed on 2nd step: hipErrorInvalidDeviceFunction: invalid device function" thrown in SetUp().
>
> so the rocm device code isn't compatible with rzvernal
>
> prob need to change a compiler setting
>

now that we don't test against tioga/ rzvernal, something about our HIP setup isn't compatible with those machines. maybe enzyme is the problem? needs more investigation.

Contributor guide

Open the contributing guide

Research direction

Start with the failing SetUp() path and reproduce the second-step scan on tioga or rzvernal. Investigate the HIP setup, compiler settings, and whether Enzyme contributes to the incompatible device code. Done means the scan runs successfully on those machines and the relevant configuration is validated.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
hpc
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.