Unterschiede

Hier werden die Unterschiede zwischen zwei Versionen angezeigt.

--- amd-ryzen-ai-max-395-ai-lab-notizen [2026/01/30 18:32] – [Upgrade auf 6.18.x Kernel] benny
+++ amd-ryzen-ai-max-395-ai-lab-notizen [2026/01/31 14:45] (aktuell) – benny
@@ Zeile 1181: / Zeile 1181: @@
     --disable-api-nodes \
     --enable-cors
+</code>
+====== Verschiedene runs mit unterschiedlichen Flags ======
+<code>
+(rocm-nightly2) benny@ailab:~$ uname -a
+Linux ailab 6.18.5+deb14-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.18.5-1 (2026-01-16) x86_64 GNU/Linux
+</code>
+===== Standard =====
+<code>
+(rocm-nightly2) benny@ailab:~/ComfyUI$ python main.py \
+    --listen 192.168.11.190 \
+    --port 8189 \
+    --gpu-only \
+    --disable-smart-memory \
+    --disable-auto-launch \
+    --enable-manager \
+    --disable-api-nodes \
+    --enable-cors
+</code>
+<code>
+(rocm-nightly2) benny@ailab:~/rocm-nightly2/ComfyUI$ python main.py     --listen 192.168.11.190     --port 8189     --gpu-only     --disable-smart-memory     --disable-auto-launch     --enable-manager     --disable-api-nodes     --enable-cors
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+[START] Security scan
+[ComfyUI-Manager] Using uv as Python module for pip operations.
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[DONE] Security scan
+** ComfyUI startup time: 2026-01-30 19:24:37.737
+** Platform: Linux
+** Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+** Python executable: /home/benny/rocm-nightly2/bin/python
+** ComfyUI Path: /home/benny/rocm-nightly2/ComfyUI
+** ComfyUI Base Folder Path: /home/benny/rocm-nightly2/ComfyUI
+** User directory: /home/benny/rocm-nightly2/ComfyUI/user
+** ComfyUI-Manager config path: /home/benny/rocm-nightly2/ComfyUI/user/__manager/config.ini
+** Log path: /home/benny/rocm-nightly2/ComfyUI/user/comfyui.log
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[PRE] ComfyUI-Manager
+Checkpoint files will always be loaded safely.
+Total VRAM 63219 MB, total RAM 126438 MB
+pytorch version: 2.9.1+rocm7.11.0a20260118
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: HIGH_VRAM
+Disabling smart memory management
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 120115.0
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Using pytorch attention
+Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+ComfyUI version: 0.11.1
+ComfyUI frontend version: 1.37.11
+[Prompt Server] web root: /home/benny/rocm-nightly2/lib/python3.13/site-packages/comfyui_frontend_package/static
+[START] ComfyUI-Manager
+[ComfyUI-Manager] network_mode: public
+[ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
+	To use this feature, please run the following command:
+	/home/benny/rocm-nightly2/bin/python -m pip install matrix-nio
+Import times for custom nodes:
+.0 seconds: /home/benny/rocm-nightly2/ComfyUI/custom_nodes/websocket_image_save.py
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.011s (created=0, skipped_existing=28, total_seen=28)
+Starting server
+To see the GUI go to: http://192.168.11.190:8189
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+FETCH ComfyRegistry Data [DONE]
+[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
+FETCH DATA from: /home/benny/rocm-nightly2/ComfyUI/user/__manager/cache/1514988643_custom-node-list.json [DONE]
+[ComfyUI-Manager] All startup tasks have been completed.
+Requested to load QwenImage
+loaded completely;  38967.90 MB loaded, full load: True
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [12:44<00:00, 15.29s/it]
+Requested to load WanVAE
+loaded completely;  242.03 MB loaded, full load: True
+Prompt executed in 00:13:38
+</code>
+===== ohne --disable-smart-memory =====
+<code>
+(rocm-nightly2) benny@ailab:~/ComfyUI$ python main.py \
+    --listen 192.168.11.190 \
+    --port 8189 \
+    --gpu-only \
+    --disable-auto-launch \
+    --enable-manager \
+    --disable-api-nodes \
+    --enable-cors
+</code>
+<code>
+(rocm-nightly2) benny@ailab:~/rocm-nightly2/ComfyUI$ python main.py     --listen 192.168.11.190     --port 8189     --gpu-only     --disable-auto-launch     --enable-manager     --disable-api-nodes     --enable-cors
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+[START] Security scan
+[ComfyUI-Manager] Using uv as Python module for pip operations.
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[DONE] Security scan
+** ComfyUI startup time: 2026-01-30 19:55:18.770
+** Platform: Linux
+** Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+** Python executable: /home/benny/rocm-nightly2/bin/python
+** ComfyUI Path: /home/benny/rocm-nightly2/ComfyUI
+** ComfyUI Base Folder Path: /home/benny/rocm-nightly2/ComfyUI
+** User directory: /home/benny/rocm-nightly2/ComfyUI/user
+** ComfyUI-Manager config path: /home/benny/rocm-nightly2/ComfyUI/user/__manager/config.ini
+** Log path: /home/benny/rocm-nightly2/ComfyUI/user/comfyui.log
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[PRE] ComfyUI-Manager
+Checkpoint files will always be loaded safely.
+Total VRAM 63219 MB, total RAM 126438 MB
+pytorch version: 2.9.1+rocm7.11.0a20260118
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: HIGH_VRAM
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 120115.0
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Using pytorch attention
+Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+ComfyUI version: 0.11.1
+ComfyUI frontend version: 1.37.11
+[Prompt Server] web root: /home/benny/rocm-nightly2/lib/python3.13/site-packages/comfyui_frontend_package/static
+[START] ComfyUI-Manager
+[ComfyUI-Manager] network_mode: public
+[ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
+	To use this feature, please run the following command:
+	/home/benny/rocm-nightly2/bin/python -m pip install matrix-nio
+Import times for custom nodes:
+.0 seconds: /home/benny/rocm-nightly2/ComfyUI/custom_nodes/websocket_image_save.py
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.015s (created=0, skipped_existing=28, total_seen=28)
+Starting server
+To see the GUI go to: http://192.168.11.190:8189
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+Requested to load QwenImage
+loaded completely;  38967.90 MB loaded, full load: True
+%|████████▋                                                                                                                                       | 3/50 [00:27<08:30, 10.87s/it]FETCH ComfyRegistry Data [DONE]
+[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
+FETCH DATA from: /home/benny/rocm-nightly2/ComfyUI/user/__manager/cache/1514988643_custom-node-list.json [DONE]
+[ComfyUI-Manager] All startup tasks have been completed.
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [12:39<00:00, 15.20s/it]
+Requested to load WanVAE
+loaded completely;  242.03 MB loaded, full load: True
+Prompt executed in 00:13:24
+</code>
+===== Mit corectrl auf "Advanced" =====
+<code>
+(rocm-nightly2) benny@ailab:~/ComfyUI$ python main.py \
+    --listen 192.168.11.190 \
+    --port 8189 \
+    --gpu-only \
+    --disable-auto-launch \
+    --enable-manager \
+    --disable-api-nodes \
+    --enable-cors
+</code>
+<code>
+(rocm-nightly2) benny@ailab:~/rocm-nightly2/ComfyUI$ python main.py     --listen 192.168.11.190     --port 8189     --gpu-only     --disable-smart-memory     --disable-auto-launch     --enable-manager     --disable-api-nodes     --enable-cors
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+[START] Security scan
+[ComfyUI-Manager] Using uv as Python module for pip operations.
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[DONE] Security scan
+** ComfyUI startup time: 2026-01-30 20:45:12.260
+** Platform: Linux
+** Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+** Python executable: /home/benny/rocm-nightly2/bin/python
+** ComfyUI Path: /home/benny/rocm-nightly2/ComfyUI
+** ComfyUI Base Folder Path: /home/benny/rocm-nightly2/ComfyUI
+** User directory: /home/benny/rocm-nightly2/ComfyUI/user
+** ComfyUI-Manager config path: /home/benny/rocm-nightly2/ComfyUI/user/__manager/config.ini
+** Log path: /home/benny/rocm-nightly2/ComfyUI/user/comfyui.log
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[PRE] ComfyUI-Manager
+Checkpoint files will always be loaded safely.
+Total VRAM 63219 MB, total RAM 126438 MB
+pytorch version: 2.9.1+rocm7.11.0a20260118
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: HIGH_VRAM
+Disabling smart memory management
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 120115.0
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Using pytorch attention
+Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+ComfyUI version: 0.11.1
+ComfyUI frontend version: 1.37.11
+[Prompt Server] web root: /home/benny/rocm-nightly2/lib/python3.13/site-packages/comfyui_frontend_package/static
+[START] ComfyUI-Manager
+[ComfyUI-Manager] network_mode: public
+[ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
+	To use this feature, please run the following command:
+	/home/benny/rocm-nightly2/bin/python -m pip install matrix-nio
+Import times for custom nodes:
+.0 seconds: /home/benny/rocm-nightly2/ComfyUI/custom_nodes/websocket_image_save.py
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.010s (created=0, skipped_existing=28, total_seen=28)
+Starting server
+To see the GUI go to: http://192.168.11.190:8189
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+Requested to load QwenImage
+loaded completely;  38967.90 MB loaded, full load: True
+%|███████████▌                                                                                                                                    | 4/50 [00:42<09:38, 12.57s/it]FETCH ComfyRegistry Data [DONE]
+[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
+FETCH DATA from: /home/benny/rocm-nightly2/ComfyUI/user/__manager/cache/1514988643_custom-node-list.json [DONE]
+[ComfyUI-Manager] All startup tasks have been completed.
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [12:39<00:00, 15.20s/it]
+Requested to load WanVAE
+loaded completely;  242.03 MB loaded, full load: True
+Prompt executed in 00:13:24
+</code>
+===== UEFI/BIOS: Performance Mode =====
+<code>
+benny@ailab:~$ source rocm-nightly2/bin/activate
+(rocm-nightly2) benny@ailab:~$ cd rocm-nightly2/ComfyUI/
+(rocm-nightly2) benny@ailab:~/rocm-nightly2/ComfyUI$ python main.py     --listen 192.168.11.190     --port 8189     --gpu-only     --disable-smart-memory     --disable-auto-launch     --enable-manager     --disable-api-nodes     --enable-cors
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+[START] Security scan
+[ComfyUI-Manager] Using uv as Python module for pip operations.
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[DONE] Security scan
+** ComfyUI startup time: 2026-01-31 14:49:55.074
+** Platform: Linux
+** Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+** Python executable: /home/benny/rocm-nightly2/bin/python
+** ComfyUI Path: /home/benny/rocm-nightly2/ComfyUI
+** ComfyUI Base Folder Path: /home/benny/rocm-nightly2/ComfyUI
+** User directory: /home/benny/rocm-nightly2/ComfyUI/user
+** ComfyUI-Manager config path: /home/benny/rocm-nightly2/ComfyUI/user/__manager/config.ini
+** Log path: /home/benny/rocm-nightly2/ComfyUI/user/comfyui.log
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[PRE] ComfyUI-Manager
+Checkpoint files will always be loaded safely.
+Total VRAM 63219 MB, total RAM 126438 MB
+pytorch version: 2.9.1+rocm7.11.0a20260118
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: HIGH_VRAM
+Disabling smart memory management
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 120115.0
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Using pytorch attention
+Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+ComfyUI version: 0.11.1
+ComfyUI frontend version: 1.37.11
+[Prompt Server] web root: /home/benny/rocm-nightly2/lib/python3.13/site-packages/comfyui_frontend_package/static
+[START] ComfyUI-Manager
+[ComfyUI-Manager] network_mode: public
+[ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
+	To use this feature, please run the following command:
+	/home/benny/rocm-nightly2/bin/python -m pip install matrix-nio
+Import times for custom nodes:
+.0 seconds: /home/benny/rocm-nightly2/ComfyUI/custom_nodes/websocket_image_save.py
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.016s (created=0, skipped_existing=32, total_seen=32)
+Starting server
+To see the GUI go to: http://192.168.11.190:8189
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+FETCH ComfyRegistry Data [DONE]
+[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
+FETCH DATA from: /home/benny/rocm-nightly2/ComfyUI/user/__manager/cache/1514988643_custom-node-list.json [DONE]
+[ComfyUI-Manager] All startup tasks have been completed.
+Requested to load QwenImage
+loaded completely;  38967.90 MB loaded, full load: True
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [12:25<00:00, 14.92s/it]
+Requested to load WanVAE
+loaded completely;  242.03 MB loaded, full load: True
+Prompt executed in 00:13:20
+</code>
+===== BIOS/UEFI Performance Mode + GPU 96 GB VRAM fest =====
+<code>
+benny@ailab:~$ source rocm-nightly2/bin/activate
+(rocm-nightly2) benny@ailab:~$ cd rocm-nightly2/ComfyUI/
+(rocm-nightly2) benny@ailab:~/rocm-nightly2/ComfyUI$ python main.py     --listen 192.168.11.190     --port 8189     --gpu-only     --disable-smart-memory     --disable-auto-launch     --enable-manager     --disable-api-nodes     --enable-cors
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+[START] Security scan
+[ComfyUI-Manager] Using uv as Python module for pip operations.
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[DONE] Security scan
+** ComfyUI startup time: 2026-01-31 15:11:30.580
+** Platform: Linux
+** Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+** Python executable: /home/benny/rocm-nightly2/bin/python
+** ComfyUI Path: /home/benny/rocm-nightly2/ComfyUI
+** ComfyUI Base Folder Path: /home/benny/rocm-nightly2/ComfyUI
+** User directory: /home/benny/rocm-nightly2/ComfyUI/user
+** ComfyUI-Manager config path: /home/benny/rocm-nightly2/ComfyUI/user/__manager/config.ini
+** Log path: /home/benny/rocm-nightly2/ComfyUI/user/comfyui.log
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+Using Python 3.13.5 environment at: /home/benny/rocm-nightly2
+[PRE] ComfyUI-Manager
+Checkpoint files will always be loaded safely.
+Total VRAM 98304 MB, total RAM 31734 MB
+pytorch version: 2.9.1+rocm7.11.0a20260118
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: HIGH_VRAM
+Disabling smart memory management
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 30147.0
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Using pytorch attention
+Python version: 3.13.5 (main, Jun 25 2025, 18:55:22) [GCC 14.2.0]
+ComfyUI version: 0.11.1
+ComfyUI frontend version: 1.37.11
+[Prompt Server] web root: /home/benny/rocm-nightly2/lib/python3.13/site-packages/comfyui_frontend_package/static
+[START] ComfyUI-Manager
+[ComfyUI-Manager] network_mode: public
+[ComfyUI-Manager] The matrix sharing feature has been disabled because the `matrix-nio` dependency is not installed.
+	To use this feature, please run the following command:
+	/home/benny/rocm-nightly2/bin/python -m pip install matrix-nio
+Import times for custom nodes:
+.0 seconds: /home/benny/rocm-nightly2/ComfyUI/custom_nodes/websocket_image_save.py
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.011s (created=0, skipped_existing=32, total_seen=32)
+Starting server
+To see the GUI go to: http://192.168.11.190:8189
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cuda:0, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cuda:0, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+Requested to load QwenImage
+loaded completely;  38967.90 MB loaded, full load: True
+%|██▉                                                                                                                                             | 1/50 [00:00<00:06,  7.29it/s]FETCH ComfyRegistry Data [DONE]
+[ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
+FETCH DATA from: /home/benny/rocm-nightly2/ComfyUI/user/__manager/cache/1514988643_custom-node-list.json [DONE]
+[ComfyUI-Manager] All startup tasks have been completed.
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [10:58<00:00, 13.17s/it]
+Requested to load WanVAE
+loaded completely;  242.03 MB loaded, full load: True
+Prompt executed in 00:11:45
+</code>
+===== Distrobox mit Performance (UEFI) und 96 GB VRAM BIOS =====
+<code>
+benny@ailab:~/rocm-nightly2/ComfyUI$ distrobox enter strix-halo-comfyui
+Starting container...                   	 [ OK ]
+Installing basic packages...            	 [ OK ]
+Setting up devpts mounts...             	 [ OK ]
+Setting up read-only mounts...          	 [ OK ]
+Setting up read-write mounts...         	 [ OK ]
+Setting up host's sockets integration...	 [ OK ]
+Integrating host's themes, icons, fonts...	 [ OK ]
+Setting up distrobox profile...         	 [ OK ]
+Setting up sudo...                      	 [ OK ]
+Setting up user's group list...         	 [ OK ]
+Container Setup Complete!
+███████╗████████╗██████╗ ██╗██╗  ██╗      ██╗  ██╗ █████╗ ██╗      ██████╗
+██╔════╝╚══██╔══╝██╔══██╗██║╚██╗██╔╝      ██║  ██║██╔══██╗██║     ██╔═══██╗
+███████╗   ██║   ██████╔╝██║ ╚███╔╝       ███████║███████║██║     ██║   ██║
+╚════██║   ██║   ██╔══██╗██║ ██╔██╗       ██╔══██║██╔══██║██║     ██║   ██║
+███████║   ██║   ██║  ██║██║██╔╝ ██╗      ██║  ██║██║  ██║███████╗╚██████╔╝
+╚══════╝   ╚═╝   ╚═╝  ╚═╝╚═╝╚═╝  ╚═╝      ╚═╝  ╚═╝╚═╝  ╚═╝╚══════╝ ╚═════╝
+                          C O M F Y   U I
+AMD Ryzen AI Max “Strix Halo” — Image & Video Toolbox (gfx1151, ROCm via TheRock)
+ROCm nightly: 7.11.0a20260115
+Machine: Micro Computer (HK) Tech Limited MS-S1 MAX
+GPU    : AMD RYZEN AI MAX+ 395 w/ Radeon 8060S
+Repo   : https://github.com/kyuz0/amd-strix-halo-comfyui-toolboxes
+Image  : docker.io/kyuz0/amd-strix-halo-comfyui:latest
+Included:
+  - ComfyUI          → start_comfy_ui (http://localhost:8000)
+SSH tip: ssh -L 8000:localhost:8000 user@host
+benny@ailab:~/rocm-nightly2/ComfyUI$ start_comfy_ui
+Adding extra search path text_encoders /home/benny/comfy-models/text_encoders
+Adding extra search path vae /home/benny/comfy-models/vae
+Adding extra search path diffusion_models /home/benny/comfy-models/diffusion_models
+Adding extra search path unet /home/benny/comfy-models/unet
+Adding extra search path loras /home/benny/comfy-models/loras
+Adding extra search path latent_upscale_models /home/benny/comfy-models/latent_upscale_models
+Adding extra search path clip_vision /home/benny/comfy-models/clip_vision
+Setting output directory to: /home/benny/comfy-outputs
+Checkpoint files will always be loaded safely.
+Total VRAM 98304 MB, total RAM 31734 MB
+pytorch version: 2.11.0a0+rocm7.11.0a20260115
+Set: torch.backends.cudnn.enabled = False for better AMD performance.
+AMD arch: gfx1151
+ROCm version: (7, 2)
+Set vram state to: NORMAL_VRAM
+Device: cuda:0 Radeon 8060S Graphics : native
+Using async weight offloading with 2 streams
+Enabled pinned memory 30147.0
+Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
+Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
+Using pytorch attention
+Python version: 3.13.11 (main, Dec  5 2025, 00:00:00) [GCC 15.2.1 20251111 (Red Hat 15.2.1-4)]
+ComfyUI version: 0.9.1
+ComfyUI frontend version: 1.36.14
+[Prompt Server] web root: /opt/venv/lib64/python3.13/site-packages/comfyui_frontend_package/static
+Traceback (most recent call last):
+  File "/opt/ComfyUI/nodes.py", line 2157, in load_custom_node
+    module_spec.loader.exec_module(module)
+    ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^
+  File "<frozen importlib._bootstrap_external>", line 1023, in exec_module
+  File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
+  File "/opt/ComfyUI/custom_nodes/ComfyUI-GGUF/__init__.py", line 7, in <module>
+    from .nodes import NODE_CLASS_MAPPINGS
+  File "/opt/ComfyUI/custom_nodes/ComfyUI-GGUF/nodes.py", line 16, in <module>
+    from .ops import GGMLOps, move_patch_to_device
+  File "/opt/ComfyUI/custom_nodes/ComfyUI-GGUF/ops.py", line 2, in <module>
+    import gguf
+ModuleNotFoundError: No module named 'gguf'
+Cannot import /opt/ComfyUI/custom_nodes/ComfyUI-GGUF module for custom nodes: No module named 'gguf'
+AMD GPU Monitor thread started
+AMD GPU Monitor: Web directory set to /opt/ComfyUI/custom_nodes/ComfyUI-AMDGPUMonitor/web
+Import times for custom nodes:
+.0 seconds: /opt/ComfyUI/custom_nodes/websocket_image_save.py
+.0 seconds: /opt/ComfyUI/custom_nodes/ComfyUI-AMDGPUMonitor
+.0 seconds (IMPORT FAILED): /opt/ComfyUI/custom_nodes/ComfyUI-GGUF
+.0 seconds: /opt/ComfyUI/custom_nodes/ComfyUI_essentials
+Using AMD SMI tool: /opt/venv/bin/rocm-smi
+Context impl SQLiteImpl.
+Will assume non-transactional DDL.
+Assets scan(roots=['models']) completed in 0.021s (created=4, skipped_existing=28, total_seen=33)
+Disabling intermediate node cache.
+Starting server
+To see the GUI go to: http://127.0.0.1:8000
+got prompt
+Using split attention in VAE
+Using split attention in VAE
+VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
+Found quantization metadata version 1
+Using MixedPrecisionOps for text encoder
+Requested to load QwenImageTEModel_
+loaded completely;  7910.28 MB loaded, full load: True
+CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
+model weight dtype torch.bfloat16, manual cast: None
+model_type FLUX
+Requested to load QwenImage
+loaded completely; 96341.98 MB usable, 38967.90 MB loaded, full load: True
+%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 50/50 [10:53<00:00, 13.06s/it]
+Requested to load WanVAE
+loaded completely; 89893.71 MB usable, 242.03 MB loaded, full load: True
+Prompt executed in 00:12:49
+^C
 </code>