Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Segfault on from triro.ray.ray_optix import RayMeshIntersector #4

Open
XDynames opened this issue Jul 30, 2024 · 9 comments
Open

Segfault on from triro.ray.ray_optix import RayMeshIntersector #4

XDynames opened this issue Jul 30, 2024 · 9 comments

Comments

@XDynames
Copy link

XDynames commented Jul 30, 2024

Just trying to set this up in a project that uses trimesh to do some raycasting and I am getting a segmentation fault when importing your package.

I am using CUDA 12.3, Pytorch 2.3 with Optix 8.0.0 on ubunutu in the nvcr.io/nvidia/pytorch:24.02-py3 docker container

Can you think of anything off the bat that could be wrong or how to chase down that problem the further?
I figure it must be some sort of version conflict

Segmentation fault: address not mapped to object at address (nil)

Did some poking around in gdb it get thrown by triro.so in optixDeviceContextCreate ()

@XDynames XDynames changed the title Segfault on import Segfault on from triro.ray.ray_optix import RayMeshIntersector Jul 30, 2024
@HS2290
Copy link

HS2290 commented Dec 15, 2024

Same problem.

@lcp29
Copy link
Owner

lcp29 commented Dec 15, 2024

Sorry for the late reply. Is there more outputs form the program? And perhaps a minimal code set to reproduce the problem:)
I will try getting the docker run first so maybe let me know what GPUs are you using?

@HS2290
Copy link

HS2290 commented Dec 15, 2024

Sorry for the late reply. Is there more outputs form the program? And perhaps a minimal code set to reproduce the problem:) I will try getting the docker run first so maybe let me know what GPUs are you using?

Thanks for your reply. I am using H100 Server, CUDA 12.1, Ubuntu 22.04, python 3.10.
My environment only includes python==3.10 and triro==1.3.0 (successfully installed).
When I finish the installation successfully and run 'import triro.ray', the code will only output 'Segmentation fault (core dumped)'. I run 'import triro' is OK, but run 'import triro.ray' fails :(

@lcp29
Copy link
Owner

lcp29 commented Dec 16, 2024

Sorry for the late reply. Is there more outputs form the program? And perhaps a minimal code set to reproduce the problem:) I will try getting the docker run first so maybe let me know what GPUs are you using?

Thanks for your reply. I am using H100 Server, CUDA 12.1, Ubuntu 22.04, python 3.10. My environment only includes python==3.10 and triro==1.3.0 (successfully installed). When I finish the installation successfully and run 'import triro.ray', the code will only output 'Segmentation fault (core dumped)'. I run 'import triro' is OK, but run 'import triro.ray' fails :(

Than you for the feedback! I will investigate into it.

@kangyiyang
Copy link
Contributor

I found that I get a segmentation fault error on the A40, but not on the 4090D. Is this related to the GPU as well? I am using CUDA 11.8, PyTorch 2.1.2, and OptiX 8.0.0

@XDynames
Copy link
Author

XDynames commented Jan 4, 2025

I had the seg fault issue on a 4090
Maybe it’s more to do with the exact combination of drivers/dependancies at compile time?

@kangyiyang
Copy link
Contributor

kangyiyang commented Jan 5, 2025 via email

@lcp29
Copy link
Owner

lcp29 commented Jan 5, 2025

I got the same problem with OptiX 8.1.0, which requires the driver version >= 555 but mine is 550. After switching to OptiX 8.0.0 (>=535) it worked fine.

@kangyiyang
Copy link
Contributor

kangyiyang commented Jan 5, 2025 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants