Skip to content

keops works well on cpu, but not works on gpu #398

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

Open
LekangJia opened this issue Nov 15, 2024 · 0 comments
Open

keops works well on cpu, but not works on gpu #398

LekangJia opened this issue Nov 15, 2024 · 0 comments

Comments

@LekangJia
Copy link

Hello, I use pykeops on centsOS 7 or AlmaLinux 9
I have cuda 11.8, gcc v11.4.1, python 3.9.18
my code work well on cpu, but when turns to gpu, I will get the following error:

0: gpu024.ihep.ac.cn
1: gpu024.ihep.ac.cn
GPU 0: Tesla V100-SXM2-32GB (UUID: GPU-323af0ee-b324-c688-6e72-4b0b4fee9374)
Allocate GPU cards : 0
36 Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz
[KeOps] Generating code for Sum_Reduction reduction (with parameters 0) of formula dExp(-(aSum((b-c)**2))) with a=Var(0,1,2), b=Var(1,8,0), c=Var(2,8,1), d=Var(3,1,1) ... OK
/var/spool/slurm/spool/job809169/slurm_script: line 27: 86666 Segmentation fault (core dumped) /hpcfs/others/ucas/jialekang/miniconda/envs/NN/bin/python /hpcfs/others/ucas/jialekang/hardon/code/train.py

At first I use the cuda 11.8 which is installed together with pytorch via conda, it shows that keops can not detect cuda. So then I install a cuda 11.8 again via Nvidia CUDA Toolkit. Now keops can detect cuda and compile some file, but then is the above Segmentation fault. It is really comfused.

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

1 participant