** Description changed: - When running Comfy-UI with Strix Halo an illegal memory access occurs - intermittently like below: + [Impact] + When running Comfy-UI with Strix Halo an illegal memory access occurs intermittently like below: 0%| | 0/20 [00:00<?, ?it/s] 5%|▌ | 1/20 [00:01<00:23, 1.21s/it] 10%|█ | 2/20 [00:06<01:01, 3.43s/it] 15%|█▌ | 3/20 [00:11<01:10, 4.14s/it] 20%|██ | 4/20 [00:16<01:11, 4.49s/it] 25%|██▌ | 5/20 [00:21<01:13, 4.90s/it] 30%|███ | 6/20 [00:26<01:09, 4.93s/it] 35%|███▌ | 7/20 [00:31<01:04, 4.96s/it] 40%|████ | 8/20 [00:36<00:59, 4.98s/it] 45%|████▌ | 9/20 [00:41<00:54, 4.98s/it] 50%|█████ | 10/20 [00:46<00:49, 4.99s/it] 55%|█████▌ | 11/20 [00:52<00:46, 5.20s/it] 60%|██████ | 12/20 [00:57<00:41, 5.14s/it] 65%|██████▌ | 13/20 [01:02<00:35, 5.09s/it] 70%|███████ | 14/20 [01:07<00:30, 5.07s/it] 75%|███████▌ | 15/20 [01:12<00:25, 5.06s/it] 80%|████████ | 16/20 [01:18<00:21, 5.26s/it] 85%|████████▌ | 17/20 [01:23<00:15, 5.24s/it] 90%|█████████ | 18/20 [01:28<00:10, 5.24s/it] 95%|█████████▌| 19/20 [01:33<00:05, 5.24s/it] 100%|██████████| 20/20 [01:39<00:00, 5.24s/it] 100%|██████████| 20/20 [01:39<00:00, 4.96s/it] /home/taccuser/workspace/7.1-18/dart/scripts/rocm-on-radeon/semi-automated-scripts/pyt_rocm_flux_inf_comfy-ui/ComfyUI/comfy/sd.py:694: UserWarning: HIP warning: an illegal memory access was encountered (Triggered internally at /pytorch/aten/src/ATen/hip/impl/HIPGuardImplMasqueradingAsCUDA.h:83.) out = self.process_output(self.first_stage_model.decode(samples, **vae_options).to(self.output_device).float()) !!! Exception during processing !!! HIP error: an illegal memory access was encountered - This issue is caused by a mismatch of expectations between the KFD and - thunk CWSR save area size. It is fixed by this change from 6.18: + [Test Plan] + Run Comfy-UI using ROCm stack with matching ROCr change on Strix halo and ensure it's stable. + + [ Where problems could occur ] + ROCm applications (which use KFD interface) + + [Other Info] + This issue is caused by a mismatch of expectations between the KFD and thunk CWSR save area size. It is fixed by this change from 6.18: https://git.kernel.org/torvalds/c/d15deafab5d72
-- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2131160 Title: Illegal memory access with Strix Halo To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2131160/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
