Ross -
The bloody pgf compilers have a long history of being "less than solid" in
my opinion. I think that even early versions of 5 will not compiler
pmemd;
the details are buried in doc somewhere. So I would not worry about this.
Anyone using pgf90 had better be pretty much up-to-date. Several years
ago,
I had tried using pgf90 with "amd" the precursor to pmemd, and back then
they apparently didn't even reliably keep the state of the floating point
stack (so in long runs the math would go funny). I only came back to
supporting them at all under pressure, and was dismayed to see Cray choose
them as their system compiler.
Regards - Bob
----- Original Message -----
From: "Ross Walker" <ross.rosswalker.co.uk>
To: <amber-developers.scripps.edu>
Sent: Monday, April 03, 2006 12:07 PM
Subject: amber-developers: Ancient PGF90 compiler - v3.2-4
> Hi All,
>
> Well it seems that one needs at least a relatively new version of the
> Portland Group compilers to compile Amber 9. E.g. trying PGF90 v3.2-4
(Ca.
> 1999) it fails miserably since it assumes every initialisation statement
> of
> the form:
>
> logical :: ready = .false.
>
> Is actually specifying that the variable ready is a constant and then
> barfs
> when something tries to change it.
>
> Oh well... Hopefully nobody will try such an ancient compiler...
<not>...
>
> All the best
> Ross
>
> /\
> \/
> |\oss Walker
>
> | HPC Consultant and Staff Scientist |
> | San Diego Supercomputer Center |
> | Tel: +1 858 822 0854 | EMail:- ross.rosswalker.co.uk |
> | http://www.rosswalker.co.uk | PGP Key available on request |
>
> Note: Electronic Mail is not secure, has no guarantee of delivery, may
not
> be read every day, and should not be used for urgent or sensitive
issues.
>
>
>
http://rosswalker.co.uk/adwords_top10/
Received on Wed Apr 05 2006 - 23:49:30 PDT
