Cuda out of memory when calling the distCUDA2

[dhruvilB01]

I am trying to reproduce the results in RTX 5080 GPU. I faced building issues with submodules, and hence updated their setup.py in order to correctly install it. Not sure if that created this problem. I need help fixing this issue.

Logs:

python slam.py --config configs/mono/tum/fr3_office.yaml
MonoGS: saving results in results/datasets_tum/2025-11-03-13-51-59
WARNING: Using soft CircularBuffer (6144 KiB)
FEngine (64 bits) created at 0x78e61e040010 (threading is enabled)
FEngine resolved backend: OpenGL
MonoGS: Resetting the system
Process Process-3:
Traceback (most recent call last):
File "/home/dhruvil/miniforge3/envs/MonoGS/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap
self.run()
File "/home/dhruvil/miniforge3/envs/MonoGS/lib/python3.10/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/home/dhruvil/EC720/GS_SLAM/MonoGS/utils/slam_backend.py", line 403, in run
self.add_next_kf(
File "/home/dhruvil/EC720/GS_SLAM/MonoGS/utils/slam_backend.py", line 68, in add_next_kf
self.gaussians.extend_from_pcd_seq(
File "/home/dhruvil/EC720/GS_SLAM/MonoGS/gaussian_splatting/scene/gaussian_model.py", line 239, in extend_from_pcd_seq
self.create_pcd_from_image(cam_info, init, scale=scale, depthmap=depthmap)
File "/home/dhruvil/EC720/GS_SLAM/MonoGS/gaussian_splatting/scene/gaussian_model.py", line 131, in create_pcd_from_image
return self.create_pcd_from_image_and_depth(cam, rgb, depth, init)
File "/home/dhruvil/EC720/GS_SLAM/MonoGS/gaussian_splatting/scene/gaussian_model.py", line 185, in create_pcd_from_image_and_depth
distCUDA2(torch.from_numpy(np.asarray(pcd.points)).float().cuda()),
MemoryError: std::bad_alloc: cudaErrorMemoryAllocation: out of memory

[2659958644]

Hello, have you resolved this issue? My graphics card is similar to yours. I would like to know if your 5080 can successfully replicate this code. I look forward to your reply.