Recent site activity

  • TclMPI
    attachment removed by Axel Kohlmeyer
    attachment from Axel Kohlmeyer
    edited by Axel Kohlmeyer
  • VRPN-ICMS
    edited by Axel Kohlmeyer
  • TclMPI
    edited by Axel Kohlmeyer
  • LAMMPS-ICMS
    edited by Axel Kohlmeyer
  • TclMPI
    attachment from Axel Kohlmeyer
  • View All
Software‎ > ‎

LAMMPS Benchmarks

LAMMPS-ICMS version 19 Nov 2010

Here as some simple comparisons (note the different number of CPU cores) with results from lammps-icms binaries compiled for various machines on the NSF TeraGrid. This is mainly to give an orientation to see what kind of performance can be expected from using hybrid OpenMP/MPI or GPU/MPI parallel execution of LAMMPS. The input files and raw output can be retrieved from
the LAMMPS-ICMS git repository.

Machine
- NICS Kraken: Cray XT5, 2x 2.6 GHz 6-core AMD Opteron per node, SeaStar2
- TACC Ranger: Sun Constellation, 4x 2.3 GHz 4-core AMD Opteron per node, SDR Infiniband
- NCSA Abe: Dell Blade Cluster, 2x 2.33 GHz 4-core Intel Xeon (Clovertown), DDR Infiniband
- TACC Longhorn: Dell r610 Cluster, 2x 2.53 GHz 4-core Intel Xeon (Nehalem), 2x Nvidia Quadro FX 5800, QDR Infiniband


Test system 1: Lennard-Jones melt: 1,149,984 particles for 10,000 MD steps


 Machine Nodes
MPI tasks
OpenMP/MPI
 GPU/MPI  Time/s
 NICS Kraken 32  384 -
-
72.8 
 NICS Kraken
32 192 2
-
72.6 
 NICS Kraken
32
128
3
-
79.4 
 NICS Kraken
32
 64 6
-
 102.7 
 TACC Ranger  
 8  128  -  -  230.0 
 TACC Ranger
 8  64  2  -  215.3 
 TACC Ranger
 8  32  4  -
 305.7 
 TACC Longhorn
16
128
-
-
103.0 
 TACC Longhorn
16
64
2
-
110.5 
 TACC Longhorn
16
32
-
1
51.9 
 NCSA Abe
 16 112
-
-
193.9 
 NCSA Abe
 16  64  2  -   201.1
 NCSA Lincoln
 16  32  -  1   74.7 

Test system 2: Gay-Berne droplet: 64,000 particles for 10,000 MD steps


 Machine Nodes
MPI tasks
OpenMP/MPI
 GPU/MPI  Time/s
 NICS Kraken
16
 192 -
-
738.9 
 NICS Kraken
16
96
2
-
746.2 
 NICS Kraken
16
64
3
-
731.9 
 NICS Kraken
16
 32 6
-
 752.1 
 TACC Ranger
 8  128  -  - 905.8 
 TACC Ranger
 8  64  2  - 917.1 
 TACC Ranger
 8  32  4  - 909.1 
 TACC Longhorn
16
128
-
-
510.1 
 TACC Longhorn
16
32
4
-
508.4 
 TACC Longhorn
16
32
-
1
93.4 
 NCSA Abe
 16 112
-
-
1042.3 
 NCSA Abe
 16  64  2  -   921.3 

Test system 3: SDS coarse grain monolayer: 250,240 particles for 10,000 MD steps


 Machine Nodes
MPI tasks
OpenMP/MPI
 GPU/MPI  Time/s
 NICS Kraken
8
 96 -
-
254.0 
 NICS Kraken
8
48
2
-
230.9 
 NICS Kraken
8
32
3
-
248.9 
 NICS Kraken
8
 16 6
-
 344.9 
  TACC Ranger  8  128  -  - 294.7 
  TACC Ranger  8  64  2  - 193.5 
  TACC Ranger  8  32  4  - 262.0 
 TACC Longhorn
16
128
-
-
125.2 
 TACC Longhorn
16
64
2
-
108.3 
 TACC Longhorn
16
32
-
1
43.7 
 NCSA Abe
 16 112
-
-
212.3 
 NCSA Abe
 16  64  2  -   171.3 

Test system 4: Bulk CHARMM water : 84,000 atoms for 10,000 MD steps


 Machine Nodes
MPI tasks
OpenMP/MPI
 GPU/MPI  Time/s
 NICS Kraken
16
 192 -
-
430.7 
 NICS Kraken
16
96
2
-
347.8 
 NICS Kraken
16
64
3
-
348.8 
 NICS Kraken
16
 32 6
-
 508.1 
 TACC Ranger  8  128  -  - 681.7 
 TACC Ranger  8  64  2  - 547.3 
 TACC Ranger  8  32  4  - 668.0 
 TACC Longhorn
16
128
-
-
323.9 
 TACC Longhorn
16
64
2
-
298.6 
 TACC Longhorn
16
32
-
1
211.1 
 NCSA Abe
 16 112
-
-
527.1 
 NCSA Abe
 16  64  2  -   430.3