AW: ERROR in CUDA run

From: Norman Geist (norman.geist_at_uni-greifswald.de)
Date: Thu Dec 17 2015 - 09:24:20 CST

Hey,

 

NAMD uses a domain-decomposition to depart the simulation cell into pieces (patches) to

distribute the work to several processors and nodes. What NAMD tells you here is that you have more

processors than patches.

 

Solutions are:

 

1. Reduce the number of processors.

2. Artificially increase the number of patches by adding “twoawayx yes” to the NAMD jobscript.

 

Good luck

 

Norman Geist

 

Von: owner-namd-l_at_ks.uiuc.edu [mailto:owner-namd-l_at_ks.uiuc.edu] Im Auftrag von Lara rajam
Gesendet: Donnerstag, 17. Dezember 2015 15:51
An: namd-l_at_ks.uiuc.edu
Betreff: namd-l: ERROR in CUDA run

 

Dear NAMD !

 

I am trying to do the CUDA run for a system which has a total

of 1400 atoms.

when I tried to launch a job in GPU the job got stopped by saying

FATAL ERROR: CUDA-enabled NAMD requires at least one patch per process

 

could any one let me know it is because of the system is to small , if so how can i fix this thank you !

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:22:20 CST