-
Notifications
You must be signed in to change notification settings - Fork 63
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
CUDA.jl CI broken again because of Enzyme #1745
Comments
So we do have a CUDA CI.....which passes......https://buildkite.com/julialang/enzyme-dot-jl/builds/3855#01917b02-f5bc-453b-a7ef-68591a5f3c89 |
@maleadt did GPUCompiler/etc change gpu_report_exception change names to report_exception? That's what it looks like? |
No, nothing should have changed in that regard. |
Well in any case, let's see if that fixes it: #1746 |
What about the other errors? See the PR JuliaGPU/CUDA.jl#2472 -- is Enzyme so sensitive to a minor change in |
I think that's genuinely due to the PR. The kernel call is changed, which seems to be the issue there. Specifically, we hook into the kernel call for a custom rule. The previous defn was
Now with that PR it is
Not sure if that specifically is the issue, but I do think the issue is from that PR itself |
Closing as needs resolution in cuda.jl extension and specific to PR at hand. |
The error I reported at the top of my issue remains:
|
Was indeed caused by a GPUCompiler bump, resolved in JuliaGPU/GPUCompiler.jl#619 |
As seen on CUDA.jl#master:
Also, as seen on JuliaGPU/CUDA.jl#2472, but probably related to Enzyme and not the PR:
TBH I'm getting kinda annoyed by the amount of breakage of the Enzyme.jl reverse CI job. Please add CUDA.jl CI here so that the Enzyme extension doesn't break that often...
The text was updated successfully, but these errors were encountered: