From: Sharp, Kim (sharpk_at_pennmedicine.upenn.edu)
Date: Wed Jan 09 2019 - 09:30:32 CST
Steve,
please provide your complete slurm script.
charmrun needs to be told by slurm at run time the node names it should run on,
otherwise it will just run on the one node slurm picks to execute your namd command line
On 1/9/19 10:08 AM, Seibold, Steve Allan wrote:
I using a Slurm script for running NAMD across several machines (or attempting to that is). When I run 2 nodes with each having 15 cores, the Slurm output says, I am running "2 nodes" each with "15 cores". A total of 30 cores; however, my namd output (log file) says that I am running on 15 cores...
Below is my command line information...I know I don't use "+p ", but I have tried "mpirun" , "charmrun", and "mpiexe" placement in the below command line and I get "no such file or directory"...
I am using the NAMD_2.13 binary obtained from NAMD site...Do I need to build/compile it instead with mpi..?
~/NAMD2/NAMD_2.13_Linux-x86_64/namd2 md6_3BP.namd
Thanks, Steve
-- Kim Sharp, Ph.D Dept. of Biochemistry and Biophysics Univ. of Pennsylvania website: crystal.med.upenn.edu
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2019 - 23:20:25 CST