The blas/lapack build works fine, but when passing the arguments to PETSC
configure, it ignores totally the BLAS/LAPACK libraries, and as a result when
running the tests all the links to blas/lapack are missing.
Any idea ?
From: Roy Stogner [royst...@ices.
It triggered the error the dbg mode once I started looking for it and
compiled in dbg mode. I first tested adapativity_ex4 with multiple
processors to make sure that there is nothing wrong with FEClough or
FEHermite.
Ata
On Tue, Jun 26, 2012 at 11:52 AM, Roy Stogner wrote:
>
> On Tue, 26 Jun 201
On Tue, 26 Jun 2012, Ataollah Mesgarnejad wrote:
OK. So the constants in 4th order equation that I solve are solution
of another non-linear system. The problem was for some reason I was
not using the right ghosted solution vector in this particular
subroutine therefore the error on the partitio
OK. So the constants in 4th order equation that I solve are solution of
another non-linear system. The problem was for some reason I was not using
the right ghosted solution vector in this particular subroutine therefore
the error on the partition boundaries.
As always thanks for the constant help
On Tue, 26 Jun 2012, Ataollah Mesgarnejad wrote:
> Just wanted to let you know I found the error in my own code and as
> far as I can tell FECLough and FEHermite are both doing nicely.
Thanks very much for the update!
Is there any quick way you could summarize the error for us? In the
past we'
Roy,
Just wanted to let you know I found the error in my own code and as far as
I can tell FECLough and FEHermite are both doing nicely.
Thanks,
Ata
On Fri, Jun 22, 2012 at 12:30 PM, Roy Stogner wrote:
>
> On Fri, 22 Jun 2012, Ataollah Mesgarnejad wrote:
>
> Yes. I actually calculated the magn
In parallel, each process only stores a subset of the full vector, so
you can't access all vector components on every processor. You need to
check that the vector index is between
NumericVector::first_local_index() and NumericVector::last_local_index()
(these functions return a different result
Hi!
I have been trying to access the structure that stores the nodal solution
with the operator "()".
For exemplo, let fem_sys be a transiente system. So I have been trying to
access its values as (*fem_sys.solution)(i) where i is the index.
It works perfectly in a serial run but if I do a paral
On Tue, 26 Jun 2012, Sylvain Vallaghe wrote:
> I'm actually only trying to see if I'm hitting the machine precision
> in my computations, so a longdouble version would be enough.
> I tried to compile a longdouble version of PETSC 3.1, but it
> wouldn't build a special blas/lapack for it when pass
I'm actually only trying to see if I'm hitting the machine precision in my
computations, so a longdouble version would be enough.
I tried to compile a longdouble version of PETSC 3.1, but it wouldn't build a
special blas/lapack for it when passing --download-c-blaslapack.
How do you get a blaslap
This might be more than just a lark for me too. It looks like long
double solver support has moved from "requires a tricky blas/lapack
build" in PETSc 3.1 to "requires a tricky blas/lapack build and a
couple edits in PETSc config python scripts" in PETSc 3.2 to "requires
a tricky blas/lapack bui
11 matches
Mail list logo
