GROMACS version: 2020.3
GROMACS modification: No
I am trying to simulate a system on a Tesla K80 GPU using a single GPU core, but I would like to specify the core that mdrun uses, say the core with ID=1. Tesla K80 comprises four GPU cores (IDs=0, 1, 2, 3). Is there a specific tag that is built into mdrun for this purpose or should I have the SGE scheduling system on the cluster assign the gpu core of interest to mdrun?
Please let me know if you need any further information.