In this page
- MARCONI
- Version 2.12 (BDW A1 partition)
Version 2.12 (A1 partition) - work in progress
Input: (http://www.ks.uiuc.edu/Research/namd/utilities/)
STMV (virus) benchmark (1,066,628 atoms, periodic, PME)
Graphic 1: the NAMD performance (simulation time in ns/day) is reported vs. the increasing number of cores
Input: (http://www.ks.uiuc.edu/Research/namd/utilities/)
APOA1 - Acquaporin, (water channel) benchmark (92 224 atoms, periodic, PME)
Graphic 1: the NAMD performance (simulation time in ns/day) is reported vs. the increasing number of cores
MARCONI (KNL- A2 partition)
Problem
Performance data for NAMD on KNL is present on the Intel website,
but until recently it has been impossible to replicate the data reported by Intel. In particular, single node performance was very poor:
APOA1 benchmark,
| cores | days/ns | ns/day |
| 8 | 1.53863 | 0.649929 |
| 16 | 0.798383 | 1.25253 |
| 32 | 0.41594 | 2.40419 |
| 68 | 0.33502 | 2.9849 |
| 136 | 6.17328 | 0.16198 |
NB: SMP version of NAMD-12 downlowded directly from NAMD website
Notice the drop in performance >32 cores and particularly with hyperthreading (136 cores).
For this reason Intel were given access to Marconi: they obtained the same results.
© Copyright 2