CUDA starting from the version 8.0 is not separated from the NVML library any more, and cannot be downloaded independently. NVML is part of the toolkit. However, the documentation requires the GPU_DEPLOYMENT_KIT_ROOT_DIR to find NVML.
I have tried setting the variable to multiple combinations including the CUDA directories, nvidia driver directory and their subdirectories to no avail. At the same time the CUDA toolkit directory contains a single NVML code example which I can compile and run successfully. Can you advice me on how to configure gromacs to use NVML? Right now I am ending up with the message: -- Could NOT find NVML (missing: NVML_LIBRARY) CMake Error at cmake/gmxManageGPU.cmake:176 (message): NVML support was required, but was not detected. Please consult the install guide. Call Stack (most recent call first): CMakeLists.txt:214 (include) Thanks, Mateusz Bieniek -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.