AMBER Archive (2003)

Subject: Compilation error of Amber 7 using pgi and mpich on Linux Cluster

From: Fabrice Yerly (fabrice.yerly_at_epfl.ch)
Date: Tue Feb 25 2003 - 07:02:00 CST

Hi everybody,

I compiled Amber 7 on a Linux cluster using pgf77 4.0-1 and mpich 1.2.1-16.
The compilation was performed without major problems. I used the
Machine.pgf77_mpich file for the compilation of sander and gibbs.

leap works fine, but sander causes some problems...

$ sander -O -i ./md/dy300_vel.in
PGFIO-F-239/namelist read/unit=5/entity name is not member of group.
 File name = ./md/dy300_vel.in formatted, sequential access record = 11
 In source file _mdread_.f, at line number 759

The same command file works fine with amber6, also compiled with pgi. Does
anybody have an idea?

Also related with this upgrade: on amber 6 I used sander_classic to perform
ibelly equilibration simulations. Because it didn't work under sander, with
forbidds the use of cutoff raddii, forcing the use of PME. It is also useful
for some system that are globally electrically charged. sander_classic
dosen't appear anymore in amber7. It there any way to use standard cutoff
radius for energy integration?

Thanks in advance for any help!

Fabrice 

-- 
Fabrice.Yerly_at_epfl.ch
ICMB-BCH ; U-3207 ;  EPFL 
CH 1015 Lausanne / Phone +41 21 693 98 81
http://icmbcu001.epfl.ch/~fyerly