[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: libsc8

On Tuesday 22 August 2006 16:43, Lennart Sorensen wrote:
> On Tue, Aug 22, 2006 at 08:34:56AM +0200, Francesco Pietra wrote:
> > Retried today to reinstall libsc8: same negation as below.
> >
> > Cheers
> > francesco
> > ________
> > Aimed at reinstalling mpqc 2.3.1 from scratch, I also wanted to reinstall
> > required tk8.4 and libsc8 (although "apt-cache show" shows both installed
> > and existing from The Scientific Computing Toolkit). No problem on
> > reinstalling tk8.4, while:
> >
> > #apt-get --reinstall install libsc8
> >
> > reported "not reinstalled because it cannot be downloaded"
> >
> > Is that libsc8 has first to be removed? Not attempted.
> >
> > All that because mpqc is unable to complete complex computations very
> > similar to those that it was able to complete successfully before. I want
> > to be sure that there is no corruption (which is not so easy - or speedy
> > - to verify otherwise: with simple, speedy computations no problem
> > arises).
> Usually that message means you have it installed, but it is no longer
> available for download from one of the sites in your sources.list
> (assuming it ever was).

As you explain below, it never was.
> Since it actually appears to be one of the packages that comes from
> mpqc, you should download it from where you downloaded the mpqc packages
> from.

I downloaded the deb packages from your site (you graciously had compiled 
it!).  I still have the packages , libsc8 is included, I removed it and 
installed freshly everything. We will see if problems came from mpqc/lib 
corruption, or this molecule deceives mpqc, which, during the computation, 
messes the internal/cartesian coordinate relationships, dying by the effort. 

If there is no corruption, the alternative is to carry out the computation 
without transformation cartesian to internal. But that would probably be too 
slow. That messing is supposed to be known issue (that I never met before 
with very similar sets of cartesian coordinates) of the program, therefore I 
am very interested to see what happens now.

What is agreed by all experts, is that big molecules are strange, sometimes 
unpredictable stuff for computations. But I am doomed to pursue such stuff.

Sorry for not having checked your packages.

Thanks a lot

francesco pietra
> --
> Len Sorensen

Reply to: