Trouble in reinstalling molpro2000.1

[Peter Knowles]

The full deal on this issue is ...

1. You don't in principle need a Fortran 90 compiler for Molpro 2000.1

2. Future versions will definitely require a Fortran 90 compiler

3. The patch dft_grad_sym2, which fixes a bug in the Kohn-Sham
gradient, was prepared for a future version of Molpro, but
retro-fitted to 2000.1, and has some f90 features. Because 2000.1 now
has a limited remaining lifespan, it was not thought worthwhile to get
rid of the f90 constructions.

4. If you don't have a f90 compiler, you can either get rid of the
patch (patch/dft_grad_sym2/revert), or else, if using Intel/linux,
simply download the compiled code from http://www.tc.bham.ac.uk/molpro

5. If you want to plan to compile future versions of Molpro on Intel,
you should be aware that the Portland f90 compiler is fully supported,
but no others are. Initial attempts to use VAST f90 revealed that
there is some work to do to get Molpro to be able to use this, and in
view of the quality and availability of pgf90, we have so far done
nothing.

Peter
At Wed, 8 Aug 2001 07:49:03 -0400,
Dave Moore wrote:
> 
>  
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi,  
> 	I had the same problem .. basically it is because the free GNU
> compiler (g77) has only limited support for FORTRAN 90, and from what
> I can tell by their mission statement (www.gnu.org), there are no
> plans to expand f90 support in g77, or to release g90.  
> 	In the short term, you can fix the specific problem you mentioned by
> simply removing the appropriate patch (something like dft_grad_sym2).
>  The way to do this is to type "patch/<patchname>/revert", where
> <patchname> is dft_grad_sym2 (or whatever the correct name is).
> 	In the long term, we are making plans to switch to the Portland
> Group (www.pgroup.com) FORTRAN compiler suite (~ 300 dollars), which
> has full support for both F77 & F90.  You can download a trial
> version (good for 15 days) from their website ... I have done this
> and it compiles MOLPRO just fine, however the Athlon optimizations
> seem to run a bit slower than for g77, as far as I can tell from the
> limited benchmarking I have done.
> 	There are also a couple of free F90-F77 translators out there that
> can run as front ends to g77.  I have messed around with one of these
> (VASTF90), with limited success ... it looks like a fair amount of
> hacking around in the MOLPRO configuration and compilation routines
> is required to make this work, and I wouldn't advise it ... things
> are really easy to break and really hard to fix in there in my
> experience *grin*.  Anyway, these g77 front ends are not officially
> supported by MOLPRO either.
> 	Anyway, that is my two cents, hope it proves helpful for you.
> 
> Dave Moore
> Chemistry Dept.
> Univ. of North Carolina at Chapel Hill
> 
> 
> - -----Original Message-----
> From: owner-molprouser at tcpc3.bham.ac.uk
> [mailto:owner-molprouser at tcpc3.bham.ac.uk]On Behalf Of
> kinugawa at ils.uec.ac.jp
> Sent: Wednesday, August 08, 2001 2:35 AM
> To: molpro-user at tcpc3.bham.ac.uk
> Subject: Trouble in reinstalling molpro2000.1
> 
> 
> Dear Experienced users of molpro2000
> 
> We got the licence of molpro2000.1 during the last January and
> succeeded 
> in the first instalation in a Linux machine. Recently we had to 
> reinstall molpro2000.1 in the same machine as it was cracked in June.
> This time, however, we failed in compliling dftigrad.f. As we checked
> the error log, we found a line saying 'Fortran 90 feature at (^) 
> unsupported'. The manual says that we can do without Fortran 90. Do
> we 
> need to use Fortran 90 at now ? 
> 
> Best Regard
> 
> Tohru Kinugawa
> The University of Electro-Communication
> 
> -----BEGIN PGP SIGNATURE-----
> Version: PGPfreeware 7.0.3 for non-commercial use <http://www.pgp.com>
> 
> iQA/AwUBO3EnL15cg+YcPBRAEQK0KQCg7SO6OD4UOmSYRX2gmHckEV8FmxgAoKvP
> nPY2S91CIzLGTRg9y+pN7+EC
> =CUeW
> -----END PGP SIGNATURE-----

-- 
Prof. Peter J. Knowles              
Email P.J.Knowles at bham.ac.uk  Phone +44-121-414-7472  Fax +44-121-414-7471
School of Chemical Sciences, University of Birmingham, Birmingham, B15 2TT, UK
WWW http://www.tc.bham.ac.uk/~peterk/