hello...
thanks for replying andrew.. i have got the binaries built..
tested a few simulations and they are running perfect now..
in case any oscar user wants a precompiled binary for namd simulations
with a mpi implementation built on a x86 linux box please write to
me..
you can be testing my build on your cluster.
thank you
genie
On 4/11/06, Andrew Friedley <afriedle_at_[hidden]> wrote:
> genie lfs wrote:
> > hello all...
> > i am trying to build Linux-i686-MPI on my OSCAR cluster.. the cluster
> > configuration is a 6 node one, with master and clients on Fedora Core
> > 3,parallel implementation being Mpich-1.2.7..
> >
> > the charm++ build was proper with the sample tests also running perfectly..
> > when i try to compile the namd using the command..
> > ./config Linux-i686-MPI and make in the Linux-i686-MPI directory..i
> > run into problems.. sample output is provided below..
> >
> > i have compiled tcl-8.4 with option --disable-shared ..
> > and fftw-2.1.5 with the option --enable-float --enable-type-prefix
> > --enable-static
> >
> >
> > the small fragment of the error pasted below...
> > can somebody please point out the error and the necessary correction...
>
> I see several different things going on here - why do you have tcl
> static? Try buiding it shared and using that. Also looks like namd is
> using one of its own memory libraries, which is stepping on existing
> libc functions.
>
> I suggest you post to a NAMD support list, as the LAM group does not
> support NAMD. I found relevant information here:
>
> http://www.ks.uiuc.edu/Research/namd/bugreport.html
>
> You might also consider talking to the charm++ group.
>
> Andrew
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
|
