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

From: Tormod Skauge <>
Date: 25 Feb 2003 16:00:19 +0100

> leap works fine, but sander causes some problems...
> $ sander -O -i ./md/
> PGFIO-F-239/namelist read/unit=5/entity name is not member of group.
> File name = ./md/ 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?

I had similar problems. Checking mdread.f didn't give me more info, nor
did recompiling make any difference. In my case it seemed that the error
was a format problem. Amber7 don't accept all the parameters from
Amber6. I couldn't find a "changed since Amber6" chapter in the Amber7
manual (definitely something nice to have), so I made a *quick* (i.e.
don't believe everything listed !) check myself comparing the manuals.
Here's most of the parameters that changed:

Removed in Amber7: "-"
Changed in Amber7: "."
New in Amber 7: "+"











-A, B, C




Section Two:

Section Four:

Section Seven:


Hope this helps.

Tormod Skauge, PhD-student Biophysical Chemistry
Dep. of Chemistry, Univ. of Bergen
Tel: +47-55588220, Fax: +47-55589490
Received on Tue Feb 25 2003 - 15:53:00 PST
Custom Search