Segmentation fault but with more than 40GB swap available.

[Giger22]

Read Troubleshoot

[x] I admit that I have read the Troubleshoot before making this issue.

Describe the problem
I have 48GB swap but still got this error.
I have tried using --always-cpu flag and I haven't got this error. I only get the error when I use GPU (AMD 5700XT 8GB) without chaning swap.

Full Console Log

1. Using command
(fooocus_env) artix:[artix]:~/Applications/fooocus$ python entry_with_update.py

Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py', '--always-gpu']
Python 3.11.6 (main, Nov 14 2023, 18:04:26) [GCC 13.2.1 20230801]
Fooocus version: 2.1.859
Running on local URL: http://127.0.0.1:7865

To create a public link, set share=True in launch().
Total VRAM 8176 MB, total RAM 15913 MB
Set vram state to: HIGH_VRAM
Always offload VRAM
/home/artix/Applications/fooocus/fooocus_env/lib/python3.11/site-packages/torch/cuda/init.py:611: UserWarning: Can't initialize NVML
warnings.warn("Can't initialize NVML")
Device: cuda:0 AMD Radeon RX 5700 XT : native
VAE dtype: torch.float32
Using sub quadratic optimization for cross attention, if you have memory or speed issues try using: --attention-split
Refiner unloaded.
Segmentation fault

Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py']
Python 3.11.6 (main, Nov 14 2023, 18:04:26) [GCC 13.2.1 20230801]
Fooocus version: 2.1.859
Running on local URL: http://127.0.0.1:7865

To create a public link, set share=True in launch().
Total VRAM 8176 MB, total RAM 15913 MB
Set vram state to: NORMAL_VRAM
Always offload VRAM
/home/artix/Applications/fooocus/fooocus_env/lib/python3.11/site-packages/torch/cuda/init.py:611: UserWarning: Can't initialize NVML
warnings.warn("Can't initialize NVML")
Device: cuda:0 AMD Radeon RX 5700 XT : native
VAE dtype: torch.float32
Using sub quadratic optimization for cross attention, if you have memory or speed issues try using: --attention-split
Refiner unloaded.
model_type EPS
UNet ADM Dimension 2816
Using split attention in VAE
Working with z of shape (1, 4, 32, 32) = 4096 dimensions.
Using split attention in VAE
extra {'cond_stage_model.clip_g.transformer.text_model.embeddings.position_ids', 'cond_stage_model.clip_l.text_projection', 'cond_stage_model.clip_l.logit_scale'}
Base model loaded: /home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors
Request to load LoRAs [['sd_xl_offset_example-lora_1.0.safetensors', 0.1], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0]] for model [/home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors].
Loaded LoRA [/home/artix/Applications/fooocus/models/loras/sd_xl_offset_example-lora_1.0.safetensors] for UNet [/home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 788 keys at weight 0.1.
Fooocus V2 Expansion: Vocab with 642 words.
Fooocus Expansion engine loaded for cuda:0, use_fp16 = True.
Requested to load SDXLClipModel
Requested to load GPT2LMHeadModel
Loading 2 new models
[Fooocus Model Management] Moving model(s) has taken 0.85 seconds
App started successful. Use the app with http://127.0.0.1:7865/ or 127.0.0.1:7865
Enter LCM mode.
[Fooocus] Downloading LCM components ...
[Parameters] Adaptive CFG = 1.0
[Parameters] Sharpness = 0.0
[Parameters] ADM Scale = 1.0 : 1.0 : 0.0
[Parameters] CFG = 1.0
[Parameters] Seed = 1708197939703660366
[Parameters] Sampler = lcm - lcm
[Parameters] Steps = 8 - 8
[Fooocus] Initializing ...
[Fooocus] Loading models ...
Refiner unloaded.
Request to load LoRAs [['sd_xl_offset_example-lora_1.0.safetensors', 0.1], ['None', 1.0], ['None', 1.0], ['None', 1.0], ['None', 1.0], ('sdxl_lcm_lora.safetensors', 1.0)] for model [/home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors].
Loaded LoRA [/home/artix/Applications/fooocus/models/loras/sd_xl_offset_example-lora_1.0.safetensors] for UNet [/home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 788 keys at weight 0.1.
Loaded LoRA [/home/artix/Applications/fooocus/models/loras/sdxl_lcm_lora.safetensors] for UNet [/home/artix/Applications/fooocus/models/checkpoints/juggernautXL_version6Rundiffusion.safetensors] with 788 keys at weight 1.0.
Requested to load SDXLClipModel
Loading 1 new model
unload clone 1
[Fooocus Model Management] Moving model(s) has taken 0.76 seconds
[Fooocus] Processing prompts ...
[Fooocus] Preparing Fooocus text #1 ...
Segmentation fault

So I got a one warning and a segmentation fault.
/home/artix/Applications/fooocus/fooocus_env/lib/python3.11/site-packages/torch/cuda/init.py:611: UserWarning: Can't initialize NVML
warnings.warn("Can't initialize NVML")
Segmentation fault

2. Using command:
(fooocus_env) artix:[artix]:~/Applications/fooocus$ python entry_with_update.py --always-gpu
Already up-to-date
Update succeeded.
[System ARGV] ['entry_with_update.py', '--always-gpu']
Python 3.11.6 (main, Nov 14 2023, 18:04:26) [GCC 13.2.1 20230801]
Fooocus version: 2.1.859
Running on local URL: http://127.0.0.1:7865

To create a public link, set share=True in launch().
Total VRAM 8176 MB, total RAM 15913 MB
Set vram state to: HIGH_VRAM
Always offload VRAM
/home/artix/Applications/fooocus/fooocus_env/lib/python3.11/site-packages/torch/cuda/init.py:611: UserWarning: Can't initialize NVML
warnings.warn("Can't initialize NVML")
Device: cuda:0 AMD Radeon RX 5700 XT : native
VAE dtype: torch.float32
Using sub quadratic optimization for cross attention, if you have memory or speed issues try using: --attention-split
Refiner unloaded.
Segmentation fault

So using --always-gpu I also get a warning:
-warnings.warn("Can't initialize NVML")
and "Segmentation fault".

[mashb1t]

Thank you for the well structured report.
Can you please try running Fooocus without any flags on your machine, so without --always-gpu and check if this still occurs? This would be appreciated.

[Giger22]

I have already provided this log. It is a first log and only a second log was a result with running the command with a --always-gpu flag.

[mashb1t]

ah ok, i was confused as it also mentions

[System ARGV] ['entry_with_update.py', '--always-gpu']

but the other log is below

[OEvgeny]

I saw similar error on 7800XT, I had to set HSA_OVERRIDE_GFX_VERSION variable to run with ROCm 5.7 and also update linux kernel to 6.6.x.

Based on the thread, you should be using HSA_OVERRIDE_GFX_VERSION=10.3.0 python entry_with_update.py as pointed here:
comfyanonymous/ComfyUI#1119