Regarding<br>"...Maybe PETSc is an overkill for you, but if the problem is
that simple, why not use Fortran and regular LAPACK/PLAPACK?,,,"<br>I haven't written fortran since IV and 77 in the early 80's. The matrix multiplication part is simple but the ambient application is squirrelly and I"m more facile in C or C++. So I just want basic linear algebra library I can link my C to, but maybe all y'all use Matlab for simple things.<br>
Probably I should just do it with, say, Armadillo, and learn something without spilling too much blood from fingertips.<br>Peter<br>
<br><br><div class="gmail_quote">On Mon, Nov 22, 2010 at 12:38 PM, Gus Correa <span dir="ltr"><<a href="mailto:gus@ldeo.columbia.edu">gus@ldeo.columbia.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="im">Peter St. John wrote:<br>
> Gus,<br>
> In my case, I"m not concerned with PDEs but the "linear system solvers"<br>
> themselves,which PETSci gets from other packages?<br>
</div>Hi Peter<br>
<br>
Yes, you can compile PETSC with support of a variety of linear algebra<br>
packages. See their installation instructions and documentation.<br>
I think there is support for PLAPACK, please check their<br>
install and docs pages:<br>
<br>
<a href="http://www.mcs.anl.gov/petsc/petsc-2/documentation/installation.html" target="_blank">http://www.mcs.anl.gov/petsc/petsc-2/documentation/installation.html</a><br>
<a href="http://www.mcs.anl.gov/petsc/petsc-2/documentation/index.html" target="_blank">http://www.mcs.anl.gov/petsc/petsc-2/documentation/index.html</a><br>
<div class="im"><br>
> So for example,<br>
> looking over their table at:<br>
> <a href="http://www.mcs.anl.gov/petsc/petsc-as/documentation/linearsolvertable.html" target="_blank">http://www.mcs.anl.gov/petsc/petsc-as/documentation/linearsolvertable.html</a><br>
> I'd want the "dense" case and therefore PLAPACK, so now I know to<br>
> google PLAPACK.<br>
> I just want to solve a system of simultaneous linear equations within<br>
> the vernacular of C/C++.<br>
<br>
</div>AFAIK, PETSc has C, Fortran, and C++ interfaces.<br>
If I remember right, you can choose between C++ or C when<br>
you compile it.<br>
You can choose between real (for most linear algebra )<br>
or complex (a must for FFTs, and FFTW support) scalars.<br>
You can also choose which<br>
linear algebra packages you will use (which you can compile separately,<br>
or let PETSc handle - it handles many of them),<br>
which BLAS/LAPACK to use (say, Goto BLAS,<br>
Intel MKL, AMD ACML, or let PETSc install its own version).<br>
<br>
I am not a PETSc user, but some people here used PETSc to solve inverse<br>
and forward problems, problems with sparse and dense matrices, etc.<br>
As it often happens when science (not computer science)<br>
is the main goal, once the "it works" stage is reached,<br>
the code starts to be used to produce papers,<br>
more and more features are added,<br>
and it is never again improved for efficiency,<br>
never goes beyond the prototype phase in this regard.<br>
<br>
In terms of efficiency, the results didn't shine,<br>
in terms of science/results output they were very good.<br>
However, it may well be that the way the problem<br>
(not PETSc) was programmed led to the inefficient code.<br>
<br>
PETSc experts in the list:<br>
For the benefit of Peter, Ivan, and the list,<br>
would you kindly jump in and say something about PETSc<br>
efficiency and other features?<br>
Txs.<br>
<br>
Maybe PETSc is an overkill for you, but if the problem is<br>
that simple, why not use Fortran and regular LAPACK/PLAPACK?<br>
<br>
My two cents.<br>
Gus Correa<br>
<div class="im"><br>
> Thanks,<br>
> Peter<br>
><br>
> On Mon, Nov 22, 2010 at 11:42 AM, Gus Correa <<a href="mailto:gus@ldeo.columbia.edu">gus@ldeo.columbia.edu</a><br>
</div><div class="im">> <mailto:<a href="mailto:gus@ldeo.columbia.edu">gus@ldeo.columbia.edu</a>>> wrote:<br>
><br>
> Would PETSc be of use for you?<br>
><br>
> <a href="http://www.mcs.anl.gov/petsc/petsc-as/" target="_blank">http://www.mcs.anl.gov/petsc/petsc-as/</a><br>
><br>
> Gus Correa<br>
><br>
> Ivan Marin wrote:<br>
> > I'm interested also in the bindings for C++ and LAPACK/ScaLAPACK. I<br>
> > did some tests in the past for simple stuff, and worked, but very<br>
> > simple linear system solvers.<br>
> ><br>
> > Ivan Marin<br>
> ><br>
> > Civil Engineering Dept<br>
> > University of Minnesota<br>
> > 500 Pillsbury Dr. S.E.<br>
> > Minneapolis, MN 55455<br>
> ><br>
> > Laboratório de Hidráulica Computacional - LHC<br>
> > Departamento de Hidráulica e Saneamento - SHS<br>
> > Escola de Engenharia de São Carlos - EESC<br>
> > Universidade de São Paulo - USP<br>
> ><br>
> > <a href="http://albatroz.shs.eesc.usp.br" target="_blank">http://albatroz.shs.eesc.usp.br</a><br>
> > +55 16 3373 8270<br>
> ><br>
> ><br>
> ><br>
> > 2010/11/17 Peter St. John <<a href="mailto:peter.st.john@gmail.com">peter.st.john@gmail.com</a><br>
</div>> <mailto:<a href="mailto:peter.st.john@gmail.com">peter.st.john@gmail.com</a>>>:<br>
<div class="im">> >> I'm going to dabble a bit with writing some C/C++ with a linear<br>
> algebra<br>
> >> library. My fortran is rusty (although I did some work<br>
> translating fortran<br>
> >> to C in this millennium :-).<br>
> >> The wiki suggests: LAPACK++, clapack, and Armadillo. Would this<br>
> choice have<br>
> >> any real bearing on compiling against MPI later?<br>
> >> Just if anyone is happy with any combination (LAPACK++ and<br>
> OpenMPI, say) I'd<br>
> >> be glad to hear it.<br>
> >><br>
> >> I just installed (k)ubuntu, which seems fine for my purposes so<br>
> far. I do<br>
> >> miss fvwm, and the default "single click to invoke an<br>
> application" is<br>
> >> totally nuts but I fixed it.<br>
> >> Thanks,<br>
> >> Peter<br>
> >><br>
> >> _______________________________________________<br>
> >> Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a><br>
</div>> <mailto:<a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a>> sponsored by Penguin Computing<br>
<div class="im">> >> To change your subscription (digest mode or unsubscribe) visit<br>
> >> <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
> >><br>
> >><br>
> > _______________________________________________<br>
> > Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a><br>
</div>> <mailto:<a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a>> sponsored by Penguin Computing<br>
<div class="im">> > To change your subscription (digest mode or unsubscribe) visit<br>
> <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
><br>
> _______________________________________________<br>
> Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a><br>
</div>> <mailto:<a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a>> sponsored by Penguin Computing<br>
<div><div></div><div class="h5">> To change your subscription (digest mode or unsubscribe) visit<br>
> <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
><br>
><br>
<br>
_______________________________________________<br>
Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a> sponsored by Penguin Computing<br>
To change your subscription (digest mode or unsubscribe) visit <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
</div></div></blockquote></div><br>