VMD-L Mailing List
From: John Stone (johns_at_ks.uiuc.edu)
Date: Mon Jan 23 2006 - 13:21:14 CST
- Next message: John Stone: "Re: large DCD file "appears to contain no timesteps""
- Previous message: Linghao Zhong: "large DCD file "appears to contain no timesteps""
- In reply to: Joshua D. Moore: "64 bit NAMD on AIX and CatDCD"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]
Hi Josh,
Just last week I added a new #define of "_LARGE_FILES" which seems to be
what the new AIX 5.x headers need. The IBM docs from AIX 4.x suggested that
LFS support on AIX keyed off of "_LARGE_FILE", but this has either changed
or that IBM documentation was erroneous. I'm testing the new plugin builds
with this new flag presently and I expect the 32-bit builds of VMD and CatDCD
to use LFS 64-bit file support in the next builds I post later this week.
I believe the change I made has already fixed the problem, but I'll need to
do more testing to be sure of this.
In the short term, using the 64-bit VMD and CatDCD builds as you suggest
is is a perfectly reasonable workaround for those that have 64-bit
processors in their AIX machine.
John Stone
vmd_at_ks.uiuc.edu
On Sat, Jan 21, 2006 at 02:22:11PM -0500, Joshua D. Moore wrote:
> Hi Jorge and Others,
>
> This goes along with previous messages from Jorge concerning writing large
> (>2GB) dcd files on AIX operating systems.
>
> We determined that compiling a 64 bit version of NAMD fixes this problem.
> SDSC has a 64 bit version now compiled in their software directory for
> those that are interested as of 1/20/2006 I believe. I have since run
> simulations and indeed you can now write > 2 GB files.
>
> I discovered another problem. When I used the 32 bit version of catdcd,
> it did not work in reading dcd files written from the 64 bit NAMD. My
> other fortran programs compiled in 32 bit and 64 bit do work however in
> reading the dcd files. This confused me. However, I downloaded the
> latest CVS plugin tree from vmd and was able to compile a 64 bit version
> of catdcd. This works perfectly!
>
> On AIX 5.2, I did this by using the default AIX5_64 flags along with
>
> setenv OBJECT_MODE=64
> setenv TCLINC=-I(tcl path to tcl include files for version 8.4.5)
> setenv TCLLIB=-L(tcl path to tcl library files for version 8.4.5)
>
> Then
>
> gmake AIX5_64
>
> I did need to modify the Makefile-arch as given in this mailing list (but
> I am using libtcl8.4.so:
>
> http://www.ks.uiuc.edu/Research/vmd/mailing_list/vmd-l/5574.html
>
> Here are the compiling flags from the Makefile-arch file from the plugin
> tree from CVS:
>
> AIX5_64:
> $(MAKE) dynlibs staticlibs bins \
> "ARCH = AIX5_64" \
> "COPTO = -o " \
> "LOPTO = -o " \
> "CC = xlc" \
> "CXX = xlC" \
> "CCFLAGS = -w -qinlglue -q64 -qarch=com -qtune=pwr4" \
> "CXXFLAGS = -w -qstrict -Q -q64 -qarch=com -qtune=pwr4" \
> "LDFLAGS = -q64 -qarch=com -qtune=pwr4" \
> "TCLLDFLAGS = /(path to tcl 8.4.5 library libtcl8.4.so file)" \
> "AR = ar -X64" \
> "NM = nm -B" \
> "RANLIB = touch" \
> "SHLD = xlC -bM:SRE -bnoentry -bexpall -lm -q64 -qarch=com
> -qtune=pwr4"
>
>
> Hope this helps someone if they also have this problem.
>
> Josh
>
>
>
> On Wed, January 11, 2006 10:06 am, Jorge Pikunic wrote:
> > Hi Michel,
> >
> > Gracias.
> >
> >> Regarding the first question, you shouldn't have any problem at loading
> >> dcd files in either 32 or 64-bit machines. For example, I usually run my
> > It seems to be a problem with the NAMD implementation on AIX (see Joshua's
> > e-mail in this thread), which is used in the machine in which I ran the
> > simulations.
> >
> >> Second, what do you mean with "recover" a part of the trajectory?
> > The resulting dcd files are unreadable by most programs that can read dcd
> > files. I was wondering if I can somehow recover the trajectories stored in
> > these files. I think that the answer is no, after a couple of replies that
> > I
> > have received.
> >
> > Thank you,
> >
> > Jorge
> >
> >
> >
> >
> >
> >
>
>
> --------------------------------------------------
> Joshua D. Moore
> Graduate Student
> North Carolina State University
> Department of Chemical and Biomolecular Engineering
> Box 7905 Centennial Campus
> Engineering Building I
> 911 Partners Way
> Raleigh, NC 27695-7905
> Phone: (919) 513-2051
> Fax: (919) 513-2470
> Email: jdmoore_at_unity.ncsu.edu
> --------------------------------------------------
-- NIH Resource for Macromolecular Modeling and Bioinformatics Beckman Institute for Advanced Science and Technology University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801 Email: johns_at_ks.uiuc.edu Phone: 217-244-3349 WWW: http://www.ks.uiuc.edu/~johns/ Fax: 217-244-6078
- Next message: John Stone: "Re: large DCD file "appears to contain no timesteps""
- Previous message: Linghao Zhong: "large DCD file "appears to contain no timesteps""
- In reply to: Joshua D. Moore: "64 bit NAMD on AIX and CatDCD"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ] [ attachment ]