undefined symbol: _ZN9LAMMPS_NS6LAMMPSC1EiPPci

[phys-chem]

The compiled LAMMPS using1080Ti can run but I have to use lower torch version. I cannot run my tasks on nodes with 2080Ti. I just get this error:
Kokkos::Cuda::initialize ERROR: running kernels compiled for compute capability 6.1 on device with compute capability 7.5 is not supported by CUDA!

However, when I try compile LAMMPS using 2080Ti. I run LAMMPS and get this error even if I do not use pair mace.
undefined symbol: _ZN9LAMMPS_NS6LAMMPSC1EiPPci

I am not sure what is the compiled problem. Could you give me some suggestions. Thank you very much!

[wcwitt]

Sorry I don't have much experience on that hardware. Can you post your exact install commands?

[phys-chem]

Thank you for the reply! I use this script for making and then run "make -j".

cmake \
    -D CMAKE_BUILD_TYPE=Release \
    -D CMAKE_INSTALL_PREFIX=$(pwd) \
    -D BUILD_MPI=yes \
    -D BUILD_OMP=yes \
    -D BUILD_SHARED_LIBS=yes \
    -D LAMMPS_EXCEPTIONS=yes \
    -D PKG_KOKKOS=yes \
    -D Kokkos_ARCH_AMDAVX=yes \
    -D Kokkos_ARCH_AMPERE100=yes \
    -D Kokkos_ENABLE_CUDA=yes \
    -D Kokkos_ENABLE_OPENMP=yes \
    -D Kokkos_ENABLE_DEBUG=no \
    -D Kokkos_ENABLE_DEBUG_BOUNDS_CHECK=no \
    -D Kokkos_ENABLE_CUDA_UVM=no \
    -D CMAKE_CXX_COMPILER=$(pwd)/../lib/kokkos/bin/nvcc_wrapper \
    -D PKG_ML-MACE=yes \
    -D CMAKE_PREFIX_PATH=$(pwd)/../../libtorch-gpu \
    ../cmake

[wcwitt]

What about modules (if applicable), libtorch version, etc?

[phys-chem]

torch 2.0.1+cu118
torch-ema 0.3
torchaudio 2.0.2+cu118
torchvision 0.15.2+cu118
mace-torch 0.3.4

[wcwitt]

Sorry, I'm not sure. Do you have a system administrator who can help with these kinds of things?