Re: AMBER: Running on MPI

From: Carlos Simmerling <carlos.ilion.bio.sunysb.edu>
Date: Wed, 06 Oct 2004 10:33:19 -0400

you still haven't told us what the error message was when you used
the Machine.g77_mpich file supplied with AMBER. It might also help to
tell us what
kind of processors you have, since some compiling options may depend on
that.

===================================================================
Carlos L. Simmerling, Ph.D.
Associate Professor Phone: (631) 632-1336
Center for Structural Biology Fax: (631) 632-1555
Stony Brook University Web: http://comp.chem.sunysb.edu/carlos
Stony Brook, NY 11794-5115 E-mail: carlos.simmerling.stonybrook.edu
===================================================================




wish_71 wrote:

>Dear Prof. Carlos L. Simmerling,
>
>Thanks a lot for your reply!
>
>I used Amber7 for molecular dynamics simulation of proteins in Single PC one year ago. Yes, sander does work under non-mpi. Now my system is more than 1000 residues, and the ruuning speed is very slowly to monitor results by using GB model. Thus, I want to move my present work into PC-cluster.
>
>The basic information of our PC-cluster is as following:
>nodes=16; ppn=2 (uses 32 CPUs total); OS: Red hat 9.0 (Fedora version).
>
>We first installed Amber7 according to the Amber7 manual, and created a link to the Machine.g77 _mpich. However, source files of Amber7 could not be successfully made. Then we just changed all "L1, L2 and L3" from all different "Makefile" files into L0 without modifying contents following "setenv". It seems that we then made install successfully. And we can run "sander" by the single CPU.
>
>This afternoon, we created a pbs-script by using more than 1 CPUs, "flag -p4pg" wrong message displayed, and the sander could not be run.
>
>After above failure, we tested another PG compiler. Unfortunately, we could not make install of Amber7 together with following wrong messages:
>
>=========================================
>......
>new2oldparm.o(.text+Ox3ec2): In function 'MAIN_':
>:undefined reference to 's_copy'
>new2oldparm.o(.text+Ox3f20): In function 'MAIN_':
>:undefined reference to 's_rsfe'
>new2oldparm.o(.text+Ox3f4e): In function 'MAIN_':
>:undefined reference to 'do_fio'
>......
>
>make[1]: ***[new2oldparm] Error 1
>make[1]: leaving directory '/usr/local/amber7/src/lib'
>make:***[install] Error 2
>
>==========================================
>
>That's all, looking forward to your warm-hearted help!
>
>Thank you very much invance!
>
>
>Best Wishes,
>
>Yingliang Wu
>
>=======================================
>College of Life Science
>Wuhan Univeristy
>Wuhan, China
>Email: wish_71_at_sina.com
>=======================================
>
>
>
>
>
>>it would be much easier to help if you tell us exactly what >sander does
>>and if you get any error message or output file.
>>
>>
>
>
>
>
>>does the single processor (non-mpi) sander work?
>>
>>
>
>
>
>
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Wed Oct 06 2004 - 15:53:00 PDT
Custom Search