GROMACS 2023.4 on RTX5080

GROMACS version: 2023.4
GROMACS modification: No

Dear GROMACS community,

I am encountering an issue when trying to run GROMACS 2023.4 on an RTX 5080. The installation, including make check, completed successfully. However, when I attempt to run a simulation using the GPU, GROMACS reports that it does not detect any capable GPU.

Interestingly, I was able to compile and run GROMACS 2021.7, 2024.6, and 2025.3 successfully on the same machine, so the problem seems to be specific to version 2023.4. I am aware that some older GROMACS versions have issues with newer GPU models, but I might have missed a workaround for this case.

Unfortunately, I need this specific version because I want to prolong some of my simulations that were previously running on a cluster that is currently down for the next six days.

All the best and many thanks already in advance,
Marius

Some additional information:

How I compiled GROMACS

#!/bin/bash

export CUDA_HOME=/usr/local/cuda-12.9
export PATH=$CUDA_HOME/bin:$PATH
export LD_LIBRARY_PATH=$CUDA_HOME/lib64:$LD_LIBRARY_PATH

cd gromacs-2023.4
mkdir build

cd build
cmake .. -DGMX_BUILD_OWN_FFTW=ON \
	 -DREGRESSIONTEST_DOWNLOAD=ON \
	 -DGMX_GPU=CUDA \
	 -DCMAKE_INSTALL_PREFIX=/opt/gromacs_2023.4 \
	 -DCMAKE_CXX_COMPILER=/usr/bin/g++-11 \
	 -DCMAKE_C_COMPILER=/usr/bin/gcc-11 \
	 -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-12.9 \
	 -DCMAKE_CUDA_COMPILER=/usr/local/cuda-12.9/bin/nvcc \
	 -DCMAKE_DISABLE_FIND_PACKAGE_MPI=on

make -j24

make check -j24

Output of gmx mdrun

WARNING: An error occurred while sanity checking device #0. An unhandled error from a previous CUDA operation was detected. CUDA error #209 (cudaErrorNoKernelImageForDevice): no kernel image is available for execution on the device.

[...]

Inconsistency in user input:
Update task on the GPU was required,
but the following condition(s) were not satisfied:
Compatible GPUs must have been found.

For more information and tips for troubleshooting, please check the GROMACS
website at https://manual.gromacs.org/current/user-guide/run-time-errors.html

Output of gmx --version

GROMACS version:    2023.4
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-3.3.8-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/gcc-11 GNU 11.5.0
C compiler flags:   -fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -O3 -DNDEBUG
C++ compiler:       /usr/bin/g++-11 GNU 11.5.0
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:       External - detected on the system
LAPACK library:     External - detected on the system
CUDA compiler:      /usr/local/cuda-12.9/bin/nvcc nvcc: NVIDIA (R) Cuda compiler driver;Copyright (c) 2005-2025 NVIDIA Corporation;Built on Tue_May_27_02:21:03_PDT_2025;Cuda compilation tools, release 12.9, V12.9.86;Build cuda_12.9.r12.9/compiler.36037853_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;-fexcess-precision=fast -funroll-all-loops -mavx2 -mfma -Wno-missing-field-initializers -Wno-cast-function-type-strict SHELL:-fopenmp -O3 -DNDEBUG
CUDA driver:        13.0
CUDA runtime:       12.90

Hi,

Forward compatibility with new architectures has indeed been broken in GROMACS 2022-2024.6, see the release note.

Using -DGMX_CUDA_TARGET_SM=120 should help: you will compile the code specifically for your GPU.

If that fails (there could be issues, haven’t tested), -DGMX_CUDA_TARGET_COMPUTE=80 is a more gentle fix: it will compile the code for the old architecture but in a forward-compatible way (how it should have worked if not for that bug).