Description: I encountered a performance issue with the preview Gaussian node in the GeometryPack. The node works fine when increasing the scale, but triggers a freeze when the scale is adjusted from a larger value to a smaller one.
Steps to Reproduce:
Set the scale parameter to a small value (e.g., 1.0) and run the workflow. It works normally.
Increase the scale to a larger value (e.g., 5.0) and run again. It works normally.
Decrease the scale back to 1.0.
The node freezes/hangs during execution. After next workflow run finishes, the preview window turns completely black.
ComfyUI 0.5.1
ComfyUI_frontend v1.34.9
EasyUse v1.3.4
rgthree-comfy v1.0.2509092031
ComfyUI-Manager V3.38.1
OS
win32
Python Version
3.12.12 | packaged by Anaconda, Inc. | (main, Oct 21 2025, 20:05:38) [MSC v.1929 64 bit (AMD64)]
Embedded Python
false
Pytorch Version
2.8.0+cu128
Arguments
.\main.py --use-sage-attention
RAM Total
95.12 GB
Name
cuda:0 NVIDIA GeForce RTX 5060 Ti : cudaMallocAsync
Type
cuda
VRAM Total
15.93 GB
Description: I encountered a performance issue with the preview Gaussian node in the GeometryPack. The node works fine when increasing the scale, but triggers a freeze when the scale is adjusted from a larger value to a smaller one.
Steps to Reproduce:
ComfyUI 0.5.1
ComfyUI_frontend v1.34.9
EasyUse v1.3.4
rgthree-comfy v1.0.2509092031
ComfyUI-Manager V3.38.1
OS
win32
Python Version
3.12.12 | packaged by Anaconda, Inc. | (main, Oct 21 2025, 20:05:38) [MSC v.1929 64 bit (AMD64)]
Embedded Python
false
Pytorch Version
2.8.0+cu128
Arguments
.\main.py --use-sage-attention
RAM Total
95.12 GB
Name
cuda:0 NVIDIA GeForce RTX 5060 Ti : cudaMallocAsync
Type
cuda
VRAM Total
15.93 GB