# My gpu not detected **URL:** \ **Category:** User discussions\ **Created:** [February 27, 2025, 11:55pm UTC](https://gromacs.bioexcel.eu/t/my-gpu-not-detected/11649 "2025-02-27T23:55:31Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![abdulla](https://avatars.discourse-cdn.com/v4/letter/a/c6cbf5/32.png) [@abdulla](https://gromacs.bioexcel.eu/u/abdulla)\ **Post date:** [February 27, 2025, 11:55pm UTC](https://gromacs.bioexcel.eu/t/my-gpu-not-detected/11649/1 "2025-02-27T23:55:31Z")
GROMACS version: GROMACS modification: Yes/No 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 #!/bin/bash #SBATCH --time=00:10:00 #SBATCH --nodes=1 #SBATCH --ntasks=1 #SBATCH --gres=gpu:1 #SBATCH --cpus-per-task=8 #SBATCH --partition=gpu module load GROMACS/2023.3-foss-2022a-CUDA-11.7.0 gmx mdrun -ntomp 8 -nb gpu -pme gpu -deffnm nvt GROMACS version: 2023.3 Precision: mixed Memory model: 64 bit MPI library: thread\_mpi OpenMP support: enabled (GMX\_OPENMP\_MAX\_THREADS = 128) GPU support: CUDA NB cluster size: 8 SIMD instructions: AVX2\_256 CPU FFT library: fftw-3.3.10-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: \*\*\*\*\*OpenMPI/4.1.4-GCC-11.3.0/bin/mpicc GNU 11.3.0 C compiler flags: -fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -O3 -DNDEBUG C++ compiler: \*\*\*\*\*OpenMPI/4.1.4-GCC-11.3.0/bin/mpicxx GNU 11.3.0 C++ compiler flags: -fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -Wno-cast-function-type-strict -fopenmp -O3 -DNDEBUG BLAS library: External - user-supplied LAPACK library: External - user-supplied CUDA compiler: \*\*\*\*\*\*/CUDA/11.7.0/bin/nvcc nvcc: NVIDIA (R) Cuda compiler driver;Copyright (c) 2005-2022 NVIDIA Corporation;Built on Tue\_May\_\_3\_18:49:52\_PDT\_2022;Cuda compilation tools, release 11.7, V11.7.64;Build cuda\_11.7.r11.7/compiler.31294372\_0 CUDA compiler flags:-std=c++17;–generate-code=arch=compute\_70,code=sm\_70;-use\_fast\_math;-Xptxas;-warn-double-usage;-Xptxas;-Werror;-D\_FORCE\_INLINES;-fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -Wno-cast-function-type-strict -fopenmp -O3 -DNDEBUG CUDA driver: 0.0 CUDA runtime: N/A ---
**Author:** ![Ashutosh](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/ashutosh/32/2587_2.png) [@Ashutosh](https://gromacs.bioexcel.eu/u/Ashutosh)\ **Post date:** [March 1, 2025, 10:53am UTC](https://gromacs.bioexcel.eu/t/my-gpu-not-detected/11649/2 "2025-03-01T10:53:37Z")
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 ```auto 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 ``` but still for me it is running fine with gpu ---
**Author:** ![al42and](https://dub1.discourse-cdn.com/flex017/user_avatar/gromacs.bioexcel.eu/al42and/32/1393_2.png) [@al42and](https://gromacs.bioexcel.eu/u/al42and)\ **Post date:** [March 1, 2025, 6:33pm UTC](https://gromacs.bioexcel.eu/t/my-gpu-not-detected/11649/3 "2025-03-01T18:33:02Z")
> [@abdulla](#): > > 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 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](https://docs.nvidia.com/cuda/archive/11.7.0/kepler-tuning-guide/index.html). 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](https://manual.gromacs.org/2023.5/release-notes/2023/2023.5.html#)). 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. > [@abdulla](#): > > CUDA driver: 0.0 > CUDA runtime: N/A 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. ---
**Author:** ![abdulla](https://avatars.discourse-cdn.com/v4/letter/a/c6cbf5/32.png) [@abdulla](https://gromacs.bioexcel.eu/u/abdulla)\ **Post date:** [March 9, 2025, 9:17am UTC](https://gromacs.bioexcel.eu/t/my-gpu-not-detected/11649/4 "2025-03-09T09:17:19Z")
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 ☺️☺️