From: Xinghua Shi (xinghua.shi_at_mf.mpg.de)
Date: Sun Nov 06 2005 - 05:41:18 CST
Thanks to Gengbin. It's working in IBM-SP computers now.
Yet it's still far away from my expectation. If I run a job without the
feature of ABF, NAMD 2.6 is fine. If I add the ABF into the input file, the
error will occur:
FATAL ERROR: Got the wrong number of atoms.
I have thought perhaps my input file caused such error. Yet it worked fine
in my linux computers. Another strange thing is: if I change the simulation
system to test another solute with the same ABF parameters, the error
information is different:
FATAL ERROR: Received too few groups.
Does this problem cause by TCL?
Xinghua Shi
Max Planck Institute for Metals Research
Heisenbergstr. 3, 70569 Stuttgart, Germany
Phone: (+49-711) 689-3567
Fax: (+49-711) 689-3512
E-mail: xinghua.shi_at_mf.mpg.de
http://shasta.mpi-stuttgart.mpg.de
-----Original Message-----
From: Gengbin Zheng [mailto:gzheng_at_ks.uiuc.edu]
Sent: Saturday, November 05, 2005 4:55 PM
To: Xinghua Shi
Cc: namd-l_at_ks.uiuc.edu
Subject: Re: namd-l: NAMD 2.6 in IBM SP
Hi,
You can try disable the context-based threads at:
charm-5.9/src/arch/mpi-sp/conv-mach.h
#define CMK_THREADS_USE_CONTEXT 0
recompile charm and NAMD.
Gengbin
Xinghua Shi wrote:
> Hi everyone,
>
>
>
> I wonder if somebody has successfully used NAMD 2.6 in IBM SP
> computers. I have successfully compiled it in IBM SP computers, yet in
> executing the error will occur with information:
>
>
>
> ATTENTION: 0031-408 8 tasks allocated by LoadLeveler,
> continuing...
>
> ------------- Processor 0 Exiting: Called CmiAbort ------------
>
> Reason: CthResume:
> swapcontext failed.
>
>
>
> I also tried the precompiled binary for IBM-SP to test it and the
> error information was the same as mentioned above.
>
> It seems the problem comes from the charm++ (version 5.9). If NAMD 2.6
> is compiled with the previous charm++ version (use charm.tar in the
> source code of NAMD 2.5), such problem will be solved. Yet the new
> feature of ABF (adaptive biasing force) in 2.6 will not work anymore,
> which is my original intention to compile version 2.6.
>
>
>
> Thanks for any solutions and suggestions!
>
>
>
> Xinghua Shi
>
> Max Planck Institute for Metals Research
> Heisenbergstr. 3, 70569 Stuttgart, Germany
> Phone: (+49-711) 689-3567
> Fax: (+49-711) 689-3512
>
> E-mail: xinghua.shi_at_mf.mpg.de
> http://shasta.mpi-stuttgart.mpg.de <http://shasta.mpi-stuttgart.mpg.de/>
>
>
>
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:41:21 CST