I am testing GPU … but got no gpu detected
I googled and found NVIDIA Tesla K80 not compatible with cuda above 11.3
idk if that mean I should reinstall GROMACS … this version is the built in for the HPC
see at the last of your query
CUDA driver: 0.0
CUDA runtime: N/A
i thing you have not installed CUDA driver and it can be installed by nvidia app in driver i think so and you also have to download same version of cuda toolkit and install it your system then you have install gromacs as you can see of mine at the time of my gromacs installation i have same CUDA driver: 12.60
CUDA runtime: 12.60 but now when i updated my nvidia app it is currently
GROMACS version: 2024.4
Precision: mixed
Memory model: 64 bit
MPI library: thread_mpi
OpenMP support: enabled (GMX_OPENMP_MAX_THREADS = 128)
GPU support: CUDA
NBNxM GPU setup: super-cluster 2x2x2 / cluster 8
SIMD instructions: AVX2_256
CPU FFT library: fftw-3.3.8-sse2-avx-avx2-avx2_128
GPU FFT library: cuFFT
Multi-GPU FFT: none
RDTSCP usage: enabled
TNG support: enabled
Hwloc support: disabled
Tracing support: disabled
C compiler: /usr/bin/cc GNU 13.2.0
C compiler flags: -fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -O3 -DNDEBUG
C++ compiler: /usr/bin/c++ GNU 13.2.0
C++ compiler flags: -fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -Wno-cast-function-type-strict SHELL:-fopenmp -O3 -DNDEBUG
BLAS library: Internal
LAPACK library: Internal
CUDA compiler: /usr/local/cuda/bin/nvcc nvcc: NVIDIA (R) Cuda compiler driver;Copyright (c) 2005-2024 NVIDIA Corporation;Built on Thu_Sep_12_02:18:05_PDT_2024;Cuda compilation tools, release 12.6, V12.6.77;Build cuda_12.6.r12.6/compiler.34841621_0
CUDA compiler flags:-std=c++17;--generate-code=arch=compute_50,code=sm_50;--generate-code=arch=compute_52,code=sm_52;--generate-code=arch=compute_60,code=sm_60;--generate-code=arch=compute_61,code=sm_61;--generate-code=arch=compute_70,code=sm_70;--generate-code=arch=compute_75,code=sm_75;--generate-code=arch=compute_80,code=sm_80;--generate-code=arch=compute_86,code=sm_86;--generate-code=arch=compute_89,code=sm_89;--generate-code=arch=compute_90,code=sm_90;-Wno-deprecated-gpu-targets;--generate-code=arch=compute_53,code=sm_53;--generate-code=arch=compute_80,code=sm_80;-use_fast_math;-Xptxas;-warn-double-usage;-Xptxas;-Werror;-D_FORCE_INLINES;-Xcompiler;-fopenmp;-fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -Wno-cast-function-type-strict SHELL:-fopenmp -O3 -DNDEBUG
CUDA driver: 12.80
CUDA runtime: 12.60
Tesla K80 has compute capability 3.7, which is compatible with CUDA up to, and including, 11.8 (it’s deprecated, but still works). For example, CUDA 11.7 even includes tuning guide for Kepler architecture. You use CUDA 11.7 and GROMACS 2023.3, which are compatible (although we suggest you switch to GROMACS 2024.5, or at least 2023.5, there are quite a few bugs fixed there since 2023.3). So far so good.
You did not specify how you build GROMACS, but the “CUDA compiler flags” in the version report says that it is only built for sm_70 architecture, which is Tesla V100 and such. Did you set GMX_CUDA_TARGET_SM=70 when configuring GROMACS? If so, why 70, and why did you change it at all? For Tesla K80, you should either set GMX_CUDA_TARGET_SM=37, or not set this option at all (don’t forget to clear the build directory in this case!), and rebuild GROMACS.
Another helpful thing to do in such cases is to run nvidia-smi prior to running GROMACS to see that NVIDIA driver is installed, which version it is, and whether it can detect GPUs.
Thx for your reply, first this was installed already on a HPC,
I will try to install the versions you have recommended,
Second … the cuda didn’t show in my report cuz I ran “GMX --version” out of slurm … The local node does not have NVIDIA gpus.
Before reading your comment I already tried to install gromacs 2023.3 with sm 37 but it didn’t work … so yes I will try the versions you recommend… Thx for your reply again ☺️☺️