About the problem of orinagx running failure on arm architecture

[z157181773]

Hello! Thank you for your work!
I encountered a problem in my recent operation. I would like to ask if this code can be run on orinagx on the arm architecture? Because I encountered problems during the implementation, I installed cuda11.4 and pytorch adapted to orinagx normally, and installed third-party libraries normally, but there was a problem when running:

nvidia@nvidia-desktop:~/ssd/zzq/MonoGS$ python slam.py --config ./configs/rgbd/replica/office0.yaml
MonoGS: saving results in results/Replica_office0/2024-12-01-20-13-29
MonoGS: training_setup Done.
Traceback (most recent call last):
File "slam.py", line 263, in
slam = SLAM(config, save_dir=save_dir)
File "slam.py", line 112, in init
backend_process.start()
File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/process.py", line 121, in start self._popen = self._Popen(self) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/context.py", line 224, in _Popen return _default_context.get_context().Process._Popen(process_obj) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/context.py", line 284, in _Popen return Popen(process_obj) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 32, in init super().init(process_obj) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/popen_fork.py", line 19, in init self._launch(process_obj) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/popen_spawn_posix.py", line 47, in _launch reduction.dump(process_obj, fp) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/multiprocessing/reduction.py", line 60, in dump ForkingPickler(file, protocol).dump(obj) File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/site-packages/torch/multiprocessing/reductions.py", line 270, in reduce_tensor event_sync_required) = storage.share_cuda() File "/home/nvidia/miniconda3/envs/SLAM/lib/python3.8/site-packages/torch/storage.py", line 1034, in share_cuda return self._untyped_storage.share_cuda(*args, **kwargs) RuntimeError: CUDA error: operation not supported CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1. Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.