I managed to get SimNet working on my local machine through a conda env. However setting up the same conda env on a cluster, I get the following error:
[OptiX][ 4][ DISK CACHE]: Opened database: "/var/tmp/OptixCache_usr/cache7.db" [OptiX][ 4][ DISK CACHE]: Cache data size: "0 Bytes" terminate called after throwing an instance of 'std::runtime_error' what(): the provided PTX was compiled with an unsupported toolchain. Aborted
stl library error.txt (7.9 KB)
The first command shows how libsdf.so was added to the library path as instructed in the installation guide.
The conda environment and HPC are listed here:
Conda env and GPU config.txt (9.6 KB)
The main significant difference between the local installation and the HPC installation was the need to add a Symbolic link to the correct cuda location on the HPC. I used @martiningram 's answer to fix this cuda library path issue.
$ ln -s /u/usr/Physics_informed_nueral_network/SimNet_sandboxv21/usr/local/cuda ./cuda_sdk_lib
Where SimNet_sandboxv21 is a singularity sandbox of the SimNet container. As mentioned this is the only significant difference between the local installation where anuerysm.py and its dependent stl libraries worked, and the HPC installation where I’m encountering the mentioned error.
My question is how do I resolve the PTX error? Any advice would be appreciated.