# Fatal error: Cannot run short-ranged nonbonded interactions on a GPU because no GPU is detected **URL:** \ **Category:** User discussions\ **Tags:** mdrun, gpu, installation-error\ **Created:** [March 13, 2025, 10:44am UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737 "2025-03-13T10:44:37Z")\ **Posts on this page:** 12\ **Page:** 1
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [March 13, 2025, 10:44am UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/1 "2025-03-13T10:44:37Z")
GROMACS version: 2024.5 GROMACS modification: No I have compiled GROMACS with the following cmake options: cmake … -DGMX\_MPI=ON -DGMX\_BUILD\_OWN\_FFTW=ON -DREGRESSIONTEST\_DOWNLOAD=ON -DGMX\_GPU=opencl -DOpenCL\_LIBRARY=/usr/lib/x86\_64-linux-gnu/libOpenCL.so -DGMX\_GPU\_NB\_CLUSTER\_SIZE=4 My Hardware: NVS Quadro 5400M Driver: NVIDIA-390.157 (installed through apt repository - legacy version) Client driver `nvidia-legacy-390xx-opencl-icd` has also been installed and my installation aborted with the following error on `make check`. ```auto Essential dynamics tests FAILED with 7 errors! 92% tests passed, 7 tests failed out of 91 Label Time Summary: GTest = 134.20 sec*proc (87 tests) IntegrationTest = 78.79 sec*proc (30 tests) MpiTest = 59.39 sec*proc (23 tests) QuickGpuTest = 35.72 sec*proc (20 tests) SlowGpuTest = 131.97 sec*proc (16 tests) SlowTest = 23.31 sec*proc (13 tests) UnitTest = 32.10 sec*proc (44 tests) Total Test time (real) = 179.75 sec The following tests FAILED: 28 - DomDecMpiTests (Failed) 35 - MdrunUtilityMpiUnitTests (Failed) 74 - MdrunMultiSimTests (Failed) 75 - MdrunMultiSimReplexTests (Failed) 76 - MdrunMultiSimReplexEquivalenceTests (Failed) 90 - regressiontests/complex (Failed) 91 - regressiontests/essentialdynamics (Failed) Errors while running CTest make[3]: *** [CMakeFiles/run-ctest-nophys.dir/build.make:71: CMakeFiles/run-ctest-nophys] Error 8 make[2]: *** [CMakeFiles/Makefile2:3310: CMakeFiles/run-ctest-nophys.dir/all] Error 2 make[1]: *** [CMakeFiles/Makefile2:3343: CMakeFiles/check.dir/rule] Error 2 make: *** [Makefile:628: check] Error 2 ``` Essentially I am trying to use my discrete GPU to use opencl for GROMACS run. my `mdrun` runs fine. However my installation went fine, no complaints whatsoever. When I try to use the GPU for computation it throws the following error: ```auto ~~> gmx_mpi mdrun -v -deffnm nvt_1 -nb gpu -pin on -pinoffset 0 :-) GROMACS - gmx mdrun, 2024.5 (-: Executable: /usr/local/gromacs/bin/gmx_mpi Data prefix: /usr/local/gromacs Working dir: /home/russellb/Documents/work/staph_nuclease/snase_charm/6eql Command line: gmx_mpi mdrun -v -deffnm nvt_1 -nb gpu -pin on -pinoffset 0 Back Off! I just backed up nvt_1.log to ./#nvt_1.log.2# Reading file nvt_1.tpr, VERSION 2024.5 (single precision) Changing nstlist from 10 to 80, rlist from 1.007 to 1.171 ------------------------------------------------------- Program: gmx mdrun, version 2024.5 Source file: src/gromacs/taskassignment/findallgputasks.cpp (line 85) Fatal error: Cannot run short-ranged nonbonded interactions on a GPU because no GPU is detected. For more information and tips for troubleshooting, please check the GROMACS website at https://manual.gromacs.org/current/user-guide/run-time-errors.html ------------------------------------------------------- -------------------------------------------------------------------------- MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 1. NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. You may or may not see output from other processes, depending on exactly when Open MPI kills them ``` Could someone shed some light on what is that I am missing (or) how to make use of my GPU with OpenCL. Many thanks in advance. :-) **Update:** My driver module is loaded: ```auto ~~> nvidia-smi Thu Mar 13 21:20:01 2025 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 390.157 Driver Version: 390.157 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 NVS 5400M Off | 00000000:01:00.0 N/A | N/A | | N/A 73C P8 N/A / N/A | 150MiB / 963MiB | N/A Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 Not Supported | +-----------------------------------------------------------------------------+ ``` and ```auto ~~> lsmod | grep nvidia nvidia_uvm 921600 4 nvidia_drm 53248 1 drm_kms_helper 204800 1 nvidia_drm nvidia_modeset 1056768 7 nvidia_drm nvidia 15888384 276 nvidia_uvm,nvidia_modeset drm 643072 5 drm_kms_helper,thinkpad_acpi,nvidia_drm ipmi_msghandler 69632 2 ipmi_devintf,nvidia video 61440 2 nvidia,thinkpad_acpi ``` GPU is detected by other programs such as `pymol`. Here, I am trying to compile `GROMACS` with `OpenCL`. Any help would be much appreciated. :-) ---
**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 14, 2025, 4:51pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/2 "2025-03-14T16:51:59Z")
Hi! Regarding the device detection: `GMX_GPU_NB_CLUSTER_SIZE=4` is the setting for consumer Intel GPUs, not compatible with NVIDIA GPUs at all. You should set `GMX_GPU_NB_CLUSTER_SIZE=8` and rebuild GROMACS. Have you checked the output of `clinfo` output to see whether the OpenCL runtime itself is functioning? And, well, you probably know that GROMACS 2024 does not officially support this GPU, so getting it through OpenCL is the best option but still not guaranteed to work. It _might_ be worthwhile to use older version of GROMACS and build it with CUDA 9.x instead of OpenCL. Regarding the failed tests: sharing the full output of `make check` is needed. Just knowing that certain tests failed is not enough to understand _why_ they failed. ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [March 15, 2025, 3:36pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/3 "2025-03-15T15:36:43Z")
Thank you for this insight. I am building it now with `GMX_GPU_NB_CLUSTER_SIZE=8`. Following is the output of my `clinfo`: ```auto ~~> clinfo Number of platforms 1 Platform Name NVIDIA CUDA Platform Vendor NVIDIA Corporation Platform Version OpenCL 1.2 CUDA 9.1.84 Platform Profile FULL_PROFILE Platform Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer Platform Extensions function suffix NV Platform Name NVIDIA CUDA Number of devices 1 Device Name NVS 5400M Device Vendor NVIDIA Corporation Device Vendor ID 0x10de Device Version OpenCL 1.1 CUDA Driver Version 390.157 Device OpenCL C Version OpenCL C 1.1 Device Type GPU Device Topology (NV) PCI-E, 0000:01:00.0 Device Profile FULL_PROFILE Device Available Yes Compiler Available Yes Max compute units 2 Max clock frequency 950MHz Compute Capability (NV) 2.1 Max work item dimensions 3 Max work item sizes 1024x1024x64 Max work group size 1024 Preferred work group size multiple (kernel) 32 Warp size (NV) 32 Preferred / native vector sizes char 1 / 1 short 1 / 1 int 1 / 1 long 1 / 1 half 0 / 0 (n/a) float 1 / 1 double 1 / 1 (cl_khr_fp64) Half-precision Floating-point support (n/a) Single-precision Floating-point support (core) Denormals Yes Infinity and NANs Yes Round to nearest Yes Round to zero Yes Round to infinity Yes IEEE754-2008 fused multiply-add Yes Support is emulated in software No Correctly-rounded divide and sqrt operations No Double-precision Floating-point support (cl_khr_fp64) Denormals Yes Infinity and NANs Yes Round to nearest Yes Round to zero Yes Round to infinity Yes IEEE754-2008 fused multiply-add Yes Support is emulated in software No Address bits 64, Little-Endian Global memory size 1010106368 (963.3MiB) Error Correction support No Max memory allocation 252526592 (240.8MiB) Unified memory for Host and Device No Integrated memory (NV) No Minimum alignment for any data type 128 bytes Alignment of base address 4096 bits (512 bytes) Global Memory cache type Read/Write Global Memory cache size 32768 (32KiB) Global Memory cache line size 128 bytes Image support Yes Max number of samplers per kernel 16 Max 2D image size 16384x16384 pixels Max 3D image size 2048x2048x2048 pixels Max number of read image args 128 Max number of write image args 8 Local memory type Local Local memory size 49152 (48KiB) Registers per block (NV) 32768 Max number of constant args 9 Max constant buffer size 65536 (64KiB) Max size of kernel argument 4352 (4.25KiB) Queue properties Out-of-order execution Yes Profiling Yes Profiling timer resolution 1000ns Execution capabilities Run OpenCL kernels Yes Run native kernels No Kernel execution timeout (NV) Yes Concurrent copy and kernel execution (NV) Yes Number of async copy engines 1 Device Extensions cl_khr_global_int32_base_atomics cl_khr_global_int32_extended_atomics cl_khr_local_int32_base_atomics cl_khr_local_int32_extended_atomics cl_khr_fp64 cl_khr_byte_addressable_store cl_khr_icd cl_khr_gl_sharing cl_nv_compiler_options cl_nv_device_attribute_query cl_nv_pragma_unroll cl_nv_copy_opts cl_nv_create_buffer NULL platform behavior clGetPlatformInfo(NULL, CL_PLATFORM_NAME, ...) NVIDIA CUDA clGetDeviceIDs(NULL, CL_DEVICE_TYPE_ALL, ...) Success [NV] clCreateContext(NULL, ...) [default] Success [NV] clCreateContextFromType(NULL, CL_DEVICE_TYPE_DEFAULT) No platform clCreateContextFromType(NULL, CL_DEVICE_TYPE_CPU) No devices found in platform clCreateContextFromType(NULL, CL_DEVICE_TYPE_GPU) No platform clCreateContextFromType(NULL, CL_DEVICE_TYPE_ACCELERATOR) No devices found in platform clCreateContextFromType(NULL, CL_DEVICE_TYPE_CUSTOM) Invalid device type for platform clCreateContextFromType(NULL, CL_DEVICE_TYPE_ALL) No platform ICD loader properties ICD loader Name OpenCL ICD Loader ICD loader Vendor OCL Icd free software ICD loader Version 2.3.1 ICD loader Profile OpenCL 3.0 ``` I am in the cross-roads and I don’t think I can afford to go back to `CUDA9.x` without resorting to a complete re-install of my OS ^\_^. I would probably stick to this and try to get it work with `OpenCL` if not impossible. Could you direct me which version of GROMACS would support my GPU officially? I can try that. Will update this post with the outcome of `GMX_GPU_NB_CLUSTER=8` shortly. Many thank you for the support. **UPDATE** Here is my test failure log: ```auto . . . 87/91 Test #87: MdrunRotationTests ........................ Passed 0.82 sec Start 88: MdrunSimulatorComparison 88/91 Test #88: MdrunSimulatorComparison .................. Passed 0.64 sec Start 89: MdrunVirtualSiteTests 89/91 Test #89: MdrunVirtualSiteTests ..................... Passed 0.91 sec Start 90: regressiontests/complex 90/91 Test #90: regressiontests/complex ...................***Failed 86.03 sec Will test on 8 MPI ranks (if possible) Will test using executable suffix _mpi Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/acetonitrilRF gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in acetonitrilRF for acetonitrilRF Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/aminoacids gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in aminoacids for aminoacids Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/argon gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in argon for argon Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/awh_multibias gmx_mpi mdrun -notunepme -cpi /home/russellb/extpacks/gromacs-2024.5/build/tests/regressiontests-2024.5/complex/awh_multibias/continue -noappend >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in awh_multibias for awh_multibias Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/awh_multidim gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in awh_multidim for awh_multidim Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/butane gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in butane for butane Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/cbt gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in cbt for cbt Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/dd121 gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in dd121 for dd121 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/dec+water gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in dec+water for dec+water Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/ethyleenglycol gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in ethyleenglycol for ethyleenglycol Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/field gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in field for field Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nacl gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nacl for nacl Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-energy-groups gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-energy-groups for nbnxn-energy-groups Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-free-energy gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-free-energy for nbnxn-free-energy Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-free-energy-vv gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-free-energy-vv for nbnxn-free-energy-vv Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-ljpme-geometric gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-ljpme-geometric for nbnxn-ljpme-geometric Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-ljpme-LB gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-ljpme-LB for nbnxn-ljpme-LB Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-ljpme-LB-geometric gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-ljpme-LB-geometric for nbnxn-ljpme-LB-geometric Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-vdw-force-switch gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-vdw-force-switch for nbnxn-vdw-force-switch Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-vdw-potential-switch gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-vdw-potential-switch for nbnxn-vdw-potential-switch Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn-vdw-potential-switch-argon gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn-vdw-potential-switch-argon for nbnxn-vdw-potential-switch-argon Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_pme gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_pme for nbnxn_pme Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_pme_order5 gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_pme_order5 for nbnxn_pme_order5 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_pme_order6 gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_pme_order6 for nbnxn_pme_order6 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_rf gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_rf for nbnxn_rf Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_rzero gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_rzero for nbnxn_rzero Abnormal return value for '/usr/bin/mpiexec -np 6 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nbnxn_vsite gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nbnxn_vsite for nbnxn_vsite Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/nst_mismatch gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in nst_mismatch for nst_mismatch Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/octahedron gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in octahedron for octahedron Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/position-restraints gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in position-restraints for position-restraints Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pr-vrescale gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pr-vrescale for pr-vrescale Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pull_constraint gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pull_constraint for pull_constraint Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pull_cylinder gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pull_cylinder for pull_cylinder Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pull_geometry_angle gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pull_geometry_angle for pull_geometry_angle Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pull_geometry_angle-axis gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pull_geometry_angle-axis for pull_geometry_angle-axis Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/pull_geometry_dihedral gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in pull_geometry_dihedral for pull_geometry_dihedral Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/reb gmx_mpi mdrun -nb cpu -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in reb for reb Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/swap_x gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in swap_x for swap_x Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/swap_y gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in swap_y for swap_y Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/swap_z gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in swap_z for swap_z Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/tip4p gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in tip4p for tip4p Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/tip4p_continue gmx_mpi mdrun -notunepme -cpi /home/russellb/extpacks/gromacs-2024.5/build/tests/regressiontests-2024.5/complex/tip4p_continue/continue -noappend >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in tip4p_continue for tip4p_continue Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/tip4pflex gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in tip4pflex for tip4pflex Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/urea gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in urea for urea Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/complex/walls gmx_mpi mdrun -notunepme >mdrun.out 2>&1' was -1 FAILED. Check mdrun.out, md.log file(s) in walls for walls 45 out of 48 complex tests FAILED Start 91: regressiontests/essentialdynamics 91/91 Test #91: regressiontests/essentialdynamics .........***Failed 9.70 sec Will test on 8 MPI ranks (if possible) Will test using executable suffix _mpi Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/linfix gmx_mpi mdrun -ei sam.edi -eo linfix.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/linacc gmx_mpi mdrun -ei sam.edi -eo linacc.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/radfix gmx_mpi mdrun -ei sam.edi -eo radfix.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/radacc gmx_mpi mdrun -ei sam.edi -eo radacc.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/radcon gmx_mpi mdrun -ei sam.edi -eo radcon.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/flooding1 gmx_mpi mdrun -ei /home/russellb/extpacks/gromacs-2024.5/build/tests/regressiontests-2024.5/essentialdynamics/flooding1/sam.edi -eo flooding1.xvg >mdrun.out 2>&1' was -1 Abnormal return value for '/usr/bin/mpiexec -np 8 -wdir /home/russellb/extpacks/gromacs-2024.5/build/tests/essentialdynamics/flooding2 gmx_mpi mdrun -ei /home/russellb/extpacks/gromacs-2024.5/build/tests/regressiontests-2024.5/essentialdynamics/flooding2/sam.edi -eo flooding2.xvg >mdrun.out 2>&1' was -1 Essential dynamics tests FAILED with 7 errors! 92% tests passed, 7 tests failed out of 91 Label Time Summary: GTest = 131.80 sec*proc (87 tests) IntegrationTest = 76.69 sec*proc (30 tests) MpiTest = 58.38 sec*proc (23 tests) QuickGpuTest = 34.51 sec*proc (20 tests) SlowGpuTest = 131.18 sec*proc (16 tests) SlowTest = 23.25 sec*proc (13 tests) UnitTest = 31.86 sec*proc (44 tests) Total Test time (real) = 178.09 sec The following tests FAILED: 28 - DomDecMpiTests (Failed) 35 - MdrunUtilityMpiUnitTests (Failed) 74 - MdrunMultiSimTests (Failed) 75 - MdrunMultiSimReplexTests (Failed) 76 - MdrunMultiSimReplexEquivalenceTests (Failed) 90 - regressiontests/complex (Failed) 91 - regressiontests/essentialdynamics (Failed) Errors while running CTest make[3]: *** [CMakeFiles/run-ctest-nophys.dir/build.make:71: CMakeFiles/run-ctest-nophys] Error 8 make[2]: *** [CMakeFiles/Makefile2:3310: CMakeFiles/run-ctest-nophys.dir/all] Error 2 make[1]: *** [CMakeFiles/Makefile2:3343: CMakeFiles/check.dir/rule] Error 2 make: *** [Makefile:628: check] Error 2 ``` While `-np 8` is impossible since I only have 4 CPU cores.`Abnormal return values..` is something I am not able to understand. Correct me if I am wrong in my assertion. Many thanks in advance. **Update-2** I am still getting the same issue: ```auto ~~> gmx_mpi mdrun -v -deffnm md_0_31 -nb gpu :-) GROMACS - gmx mdrun, 2024.5 (-: Executable: /usr/local/gromacs/bin/gmx_mpi Data prefix: /usr/local/gromacs Working dir: /home/russellb/Documents/work/Amyloid/MDSim/abeta42/7prd Command line: gmx_mpi mdrun -v -deffnm md_0_31 -nb gpu Reading file md_0_31.tpr, VERSION 2024.5 (single precision) Changing nstlist from 10 to 100, rlist from 1 to 1.164 ------------------------------------------------------- Program: gmx mdrun, version 2024.5 Source file: src/gromacs/taskassignment/findallgputasks.cpp (line 85) Fatal error: Cannot run short-ranged nonbonded interactions on a GPU because no GPU is detected. For more information and tips for troubleshooting, please check the GROMACS website at https://manual.gromacs.org/current/user-guide/run-time-errors.html ------------------------------------------------------- -------------------------------------------------------------------------- MPI_ABORT was invoked on rank 0 in communicator MPI_COMM_WORLD with errorcode 1. NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes. You may or may not see output from other processes, depending on exactly when Open MPI kills them. -------------------------------------------------------------------------- ``` May be I will try with an earlier version of GROMACS. Which one would support my GPU? ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [March 17, 2025, 1:09pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/4 "2025-03-17T13:09:42Z")
Hi @al42and. Would it be reasonable to downgrade to an earlier version of GROMACS? if so, could you kindly advise which version would be better? Many thanks in advance. ---
**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 30, 2025, 11:42pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/5 "2025-03-30T23:42:19Z")
> [@rube23](#): > > Following is the output of my `clinfo`: Looks fine. > [@rube23](#): > > While `-np 8` is impossible since I only have 4 CPU cores.`Abnormal return values..` is something I am not able to understand. Correct me if I am wrong in my assertion. Many thanks in advance. In regression tests, GROMACS tries different combinations, and `-np 8` is just the last one tried. As mentioned, it would be good to see _full_ output of `make check` (like, all thousands of lines; the parts you shared only show that some tests failed, without any details of why). However, the types of tests that fail hint (but it’s hard to tell for sure without actual error messages) that your MPI installation is broken / misconfigured. Have you checked that MPI works? Also, is there any reason you are building with libMPI and not threadMPI? It looks like you’re on a laptop, so multi-node runs are not a requirement for you, right? > [@rube23](#): > > Could you direct me which version of GROMACS would support my GPU officially? I can try that. GROMACS 2018.8 should support it, but still, your best bet is to try using CUDA instead of OpenCL (it will also have better performance). Also, correction to what I said earlier: for your GPU, you need CUDA 8, not CUDA 9. > [@rube23](#): > > I am still getting the same issue: Running `gmx_mpi mdrun -v -deffnm md_0_31 -nb gpu` will also produce `md_0_31.log` file, which, among other things, contains detailed hardware information, including the reasons why GPU was considered incompatible. ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [April 1, 2025, 2:51pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/6 "2025-04-01T14:51:17Z")
Thank you for the insights. I indeed checked the mpi and it works fine (for my other programs). I am relying on the `mpiexec` with `-ntomp` option and works for me. Going back to 2018.8 sounds not an option as you hinted. Having CUDA would require re-install, if not, recompilation of the kernel which I am not sure at this moment, I can afford it. However, I am not wiping that option. The log says the following: ```auto When checking whether update groups are usable: Domain decomposition is not active, so there is no need for update groups ------------------------------------------------------- Program: gmx mdrun, version 2024.5-plumed_2.11.0_dev Source file: src/gromacs/taskassignment/findallgputasks.cpp (line 85) Fatal error: Cannot run short-ranged nonbonded interactions on a GPU because no GPU is detected. For more information and tips for troubleshooting, please check the GROMACS website at https://manual.gromacs.org/current/user-guide/run-time-errors.html ------------------------------------------------------- ``` My GPU is operational and my `nvidia-smi` says: ```auto ~~> nvidia-smi . . . +-----------------------------------------------------------------------------+ | NVIDIA-SMI 390.157 Driver Version: 390.157 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 NVS 5400M Off | 00000000:01:00.0 N/A | N/A | | N/A 50C P8 N/A / N/A | 116MiB / 963MiB | N/A Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 Not Supported | +-----------------------------------------------------------------------------+ ``` I think it is because CUDA is not installed and OpenGL operates just for graphics rendering. Thank you for the insights. ---
**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:** [April 1, 2025, 11:21pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/7 "2025-04-01T23:21:55Z")
> [@rube23](#): > > The log says the following: The mdrun log file should also contain (above) the hardware report like: ```plaintext GROMACS version: 2025-dev-20241007-a45813a8bc GIT SHA1 hash: a45813a8bc5a6add3e61f2beab15716cf864bea9 Precision: mixed [skipped ...] Hardware detected on host: CPU info: Vendor: Intel Brand: 12th Gen Intel(R) Core(TM) i9-12900K Family: 6 Model: 151 Stepping: 2 Features: aes apic avx avx2 clfsh cmov cx8 cx16 f16c fma htt intel lahf mmx msr nonstop_tsc pcid pclmuldq pdcm pdpe1gb popcnt pse rdrnd rdtscp sha sse2 sse3 sse4.1 sse4.2 ssse3 tdt x2apic Hardware topology: Basic Packages, cores, and logical processors: [indices refer to OS logical processors] Package 0: [0 1] [2 3] [4 5] [6 7] [8 9] [10 11] [12 13] [14 15] [16] [17] [18] [19] [20] [21] [22] [23] CPU limit set by OS: -1 Recommended max number of threads: 24 GPU info: Number of GPUs detected: 1 #0: name: NVIDIA GeForce RTX 3060, architecture 8.6, vendor: NVIDIA Corporation, device version: 8.6, driver version: CUDA 12.6, status: compatible ``` In your case, presumably, there would either be no GPUs at all (which would indicate that GROMACS fails to initialize the OpenCL backend) or some reason for the GPU incompatibility will be printed. Without knowing that, it’s pretty much impossible to debug the issue. > [@rube23](#): > > Going back to 2018.8 sounds not an option as you hinted. Having CUDA would require re-install, if not, recompilation of the kernel which I am not sure at this moment, I can afford it. That sounds unusual. Driver is what interacts with the kernel; CUDA itself is purely user-space (you can, theoretically, install it without `root`), and both CUDA 8 and CUDA 9 are compatible with driver 390.x, so there should be no reason to do anything with the kernel or even kernel modules. But you know your setup better, so as you say. ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [April 3, 2025, 12:58am UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/8 "2025-04-03T00:58:34Z")
It’s very kind of your gesture. Yes. indeed the log says there is incompatibility with `CUDA`. Find below the log trace. ```auto . . . CPU limit set by OS: -1 Recommended max number of threads: 4 GPU info: Number of GPUs detected: 1 #0: name: NVS 5400M, vendor: NVIDIA Corporation, device version: OpenCL 1.1 CUDA, status: incompatible . . . ``` As to the Driver compatibility: I tried installing CUDA8, however, it pulls a previous version of NVIDIA driver, if my memory works good, it pulled `NVIDIA-357.xxx` whereas my system repository has `NVIDIA-390.xxx`. Current OS is Debian-12 whereas CUDA9 supports only up to `Ubuntu-17.06` which is Debian 9. I am bit hesitant as this might invite unwanted issues at this point. However, I will try my efforts with `OpenCL`. Many thanks. ---
**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:** [April 4, 2025, 1:49pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/9 "2025-04-04T13:49:35Z")
> [@rube23](#): > > `device version: OpenCL 1.1 CUDA` Thank you. That explains it, GROMACS (even the 2018 version) needs OpenCL at least 1.2. So, we now know why it’s not compatible. From what I can tell, this is simply how NVIDIA decided to do things for you GPU, and there’s not much you can do :( > [@rube23](#): > > As to the Driver compatibility: I tried installing CUDA8, however, it pulls a previous version of NVIDIA driver, if my memory works good, it pulled `NVIDIA-357.xxx` whereas my system repository has `NVIDIA-390.xxx`. Current OS is Debian-12 whereas CUDA9 supports only up to `Ubuntu-17.06` which is Debian 9. I am bit hesitant as this might invite unwanted issues at this point. However, I will try my efforts with `OpenCL`. Many thanks. I see. That’s painful. I’d say CUDA8 + GROMACS 2018 is still your best bet here. I would try to install CUDA 8 via a runfile (not package manager to avoid it pulling any dependencies), but afterward you are likely to encounter issues with the compiler (GCC) and have to install older versions. Using `spack` or `nix` or `docker` to install the software in an isolated environment without messing up your system could be an easier way. I don’t see fundamental reasons why it cannot work, but it’s not going to be easy to use hardware from ~2012, supported up to ~2018, with OS from ~2023. ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [April 8, 2025, 1:11am UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/10 "2025-04-08T01:11:41Z")
Hi Thanks for the insights. Yes, indeed, I thought about having `GROMACS` inside a `docker` environment. However, having `GROMACS` inside a container would significantly hamper the performance. I did forget to mention that I tried installing `CUDA8` from a run-file, explicitly here. As you mentioned, `NVIDIA` is known for these kind of cheap practices `:(` Thanks for your time and efforts though! But, going back to `CUDA8 + GROMACS 2018` would require an OS re-install which is not an option for me, as I mentioned earlier. Thanks. ---
**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:** [April 9, 2025, 9:54am UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/11 "2025-04-09T09:54:40Z")
> [@rube23](#): > > Yes, indeed, I thought about having `GROMACS` inside a `docker` environment. However, having `GROMACS` inside a container would significantly hamper the performance Would it? We haven’t run dedicated testing, but in my experience using Docker for local testing, the performance is close to native. But it could be problematic using GPU passthrough with 390-series driver. And `spack` and `nix` are not using containers, they are just user-space package managers, so they just help you install CUDA and all the dependencies into a separate directory, without impacting the rest of the OS (both can also build GROMACS; not sure about `nix`, but for `spack` it’s well tested). There can still be compatibility issues, of course, so depends on how much more effort you are willing to put ito that all. ---
**Author:** ![rube23](https://avatars.discourse-cdn.com/v4/letter/r/a8b319/32.png) [@rube23](https://gromacs.bioexcel.eu/u/rube23)\ **Post date:** [April 9, 2025, 12:38pm UTC](https://gromacs.bioexcel.eu/t/fatal-error-cannot-run-short-ranged-nonbonded-interactions-on-a-gpu-because-no-gpu-is-detected/11737/12 "2025-04-09T12:38:52Z")
Thanks.