AW: peculiar scaling on the GPU

From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Wed Aug 01 2012 - 01:14:26 CDT

Hi,

 

this looks in fact a little strange. What kind of machine is that?

How many CPU sockets and what cpus are inside?

Is this reproducible?

 

Norman Geist.

 

Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im Auftrag
von Ajasja Ljubetic
Gesendet: Dienstag, 31. Juli 2012 16:07
An: namd-l
Betreff: namd-l: peculiar scaling on the GPU

 

Dear all,

 

I'm noticing peculiar scaling when using a GPU (GTX 560). The graph is shown
here <http://img38.imageshack.us/img38/2933/scaling.png> . If I run the
calculation only on the CPUs I get excellent linear scaling, but when using
the GPU there are huge jumps at 3 and 5 cores.

 

What could be the reason for these jumps?

 

I'm running NAMD 2.9 Linux-x86_64-multicore, the systems is ~1.5k atoms
large. The important performance parts of the conf file are below. I'm also
using the colvars module to keep track of two dihedral angles. An example
logfile is attached.

 

Thanks for your help,

Ajasja

 

nonbondedFreq 1

fullElectFrequency 1

stepspercycle 20

 

switching on

longSplitting C2

switchdist 10

cutoff 12

pairlistdist 14

 

PME yes

PMEGridSpacing 1

 

 

 

  

 

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:52 CST