AMBER Archive (2009)

Subject: Re: [AMBER] amber 10 test failures...

From: Mark Williamson (mjw_at_sdsc.edu)
Date: Mon Nov 02 2009 - 18:11:40 CST

Ross Walker wrote:
> Hi Peker,
>
> ** NCSU-Error ** : expected list value for key 'i', got '<EMPTY>' instead
>
> I suspect this is because they use some new namelist option in the input
> files that your compiler is not supporting. That said I am sure others have
> used gfortran 4.1.2 and have not seen this problem. The nice thing at least

This is specific to the gfortran 4.1.2 compiler; I see it on my RHEL4
desktop machine. I think it is some form of compiler bug; I do not see
this issue with ifort or later versions of gfortran.

The error message is being emitted by the type_mismatch() routine, which
I think has been called by node_lookup_list() which has been called by
colvar_mdread() whilst I think parsing the section:

       # list of participating atoms
       i = (1, 5, 8, 10, 12, 29, 31, 33, 41,
            43, 53, 55, 57, 59, 70, 72, 74,
            84, 86, 88, 91, 93, 95, 105, 107,
            109, 129, 131, 133, 136)

of the test/ncsu/abmd_ANALYSIS/mdin file. I had a look at this a while
ago with the aim to generating some form of workaround, but I could not
ascertain what specifically was causing it. I've started to look into it
again, but I'm not sure how much joy I will have: compiler bugs are a
real time sink. 

-- 
Mark Williamson, Post Doc
Walker Molecular Dynamics Group
Room 395E
San Diego Supercomputer Center
9500 Gilman Drive
La Jolla, CA 92093-0505
Email:  mjw at sdsc.edu
Office: 858-246-0827

_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber