On Wed, Apr 19, 2023 at 1:12 PM Christian König <ckoenig.leichtzumer...@gmail.com> wrote: > > I'm already looking into this, but can't figure out why we run into > problems here. > > What happens is that a CS is aborted without sending the job to the > scheduler and in this case the cleanup function doesn't seem to work. > > Christian.
I can easily reproduce it on any AMD GPU hardware. You can add more logs to debug and I return with new logs which explains this. Thanks. -- Best Regards, Mike Gavrilov.
