Dear AMBER usesrs,
I'm using AMBER12, but i'm having trouble in performance and scaling, the compilation with performed with the mvapich2 library and the connection is of infiniband type, Tesla type cluster. I tried to monitor the performance with different libraries and number of processors, the results are in the follwoing table:
Number of processor Performance(ns/day) MPI Library
2 ns/day = 0.15 mvapich2
4 ns/day = 0.21 mvapich2
8 ns/day = 0.28 mvapich2
16 ns/day = 0.20 mvapich2
32 ns/day = 0.03 mvapich2
2 ns/day = 0.15 mpich2
4 ns/day = 0.21 mpich2
8 ns/day = 0.28 mpich2
16 ns/day = 0.18 mpich2
32 ns/day = 0.05 mpich2
The system is of moderate size (32000 atoms), and i'm running a simple NVT at Zero temp...
As you can see, performance increases (although not too much) going from 2-----> 8 but drops rapidly after that...
Regards
Marawan
Here's my script:
#!/bin/bash
#PBS -l mem=1gb
#PBS -l walltime=01:10:00
#PBS -N m8_npt
source /usr/local/modules/init/bash
module load amber
module load mvapich2/1.8 (change this according to the used library)
cd $PBS_O_WORKDIR
mpirun -np 16 pmemd.MPI -O -i eq_1_heat.in -p com_solvated_m8.top -c min_solventonly_5.rst -r eq_1_heat.rst -x eq_1_heat.mdcrd -o eq_1_heat.out -ref min_solventonly_5.rst
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Sep 03 2012 - 18:30:03 PDT
