Hello gmx users, I have a two-socket (8 physical cores each, and with hyperthreading 32 cores system) with 2 gpus. I want to run two independent simulations using 16 threads and a single GPU.
I tried the -pin on and -pinoffset 0 for the first instance of mdrun and -pin on and -pinoffset 8 for the second instance. However, that did not show all 32 threads active in htop. Then I tried -pin on -pinoffset 0 and -pin on -pinoffset 16, but still the same result. I noticed in the log file that mdrun selected a logical core stride of 2. I have attached a snapshot of htop. http://i1381.photobucket.com/albums/ah223/devapriyachem/Screenshot-mdrun_zpsi12o1hmq.png I hope I can get some advice on how to fix this, so that both simulations can share the hardware efficiently. Thanks, -- 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.
