From: Gengbin Zheng (gzheng_at_ks.uiuc.edu)
Date: Thu Dec 30 2004 - 18:52:47 CST
Hi,
Charm++ and NAMD have been tested on clustermatic 4. I have never got a
chance to test it on 5. Also, x86_64 (opteron) make it different with
the previous builds on 32 bit based machines. So I have no idea if it
can compile right out of the source, but you can give it a shot and I
believe there is no fundamental reason why it would not work.
I believe you need to download the latest charm++ from CVS. Since the
cvs server of Charm++ now is closed from outside of campus for some
security reasons, I beileve you can only download the nightly build
source from: http://charm.cs.uiuc.edu/download/downloads.phtml. Please
*wait* until tomorrow to download tonight's version because I have made
couple changes to allow clustermatic build for amd64 machines.
Once you get Charm++ source, compile Charm++ with:
./build charm++ net-linux-amd64 clustermatic
Make sure charm/tests/charm++/megatest works before you proceed to NAMD.
Currently there is no config file for compiling x86-64 version of
clustermatic in NAMD, you can still use Linux-i686-Clustermatic.arch but
change the CHARMARCH to point to net-linux-amd64-clustermatic as you
just compiled in charm directory.
For information on compilation and how to run NAMD under clustermatic,
please see NAMD wiki page:
http://www.ks.uiuc.edu/Research/namd/wiki/index.cgi?NamdOnClustermatic
you are welcome to modify it to add your tips.
Gengbin
Rene Salmon wrote:
>Hi list,
>
>
>We have an x86_64 clustermatic 5 bproc cluster and we need to get
>Namd/charm running on it.
>
>I found posts on the list of people getting clustermatic 4 and namd/charm
>running. What I am not clear about is what i need to get this working.
>I am hopping some one can point me in the right direction:
>
>1) Can I use Namd from source or do I need to get it from CVS?
>2) Can I use the charm that came with the namd source? If not can i use
>the charm-5.8.tar.gz source that is at the charm www site or do I need
>charm from cvs?
>
>3)better yet any one have binaries??
>
>
>Thank you for any help.
>
>Rene
>
>-
>--
> Rene Salmon
> Tulane University
> Center for Computational Sciences
> Richardson Building 310
> New Orleans, LA 70118
> Tel 504-862-8393
> Fax 504-862-8392
>
>
>
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:38:06 CST