El mar, 4 mar 2025 a la(s) 4:25 p.m., Pierre Jolivet (pie...@joliv.et)
escribió:
>
>
> On 4 Mar 2025, at 11:17 PM, Emmanuel Ayala wrote:
>
>
> Well, I just tested the configuration with --download-f2cblaslapack
> --download-blis
> and the direct solver cholesky with mumps becomes 10x faster :D
>
:-)
> On Mar 4, 2025, at 5:17 PM, Emmanuel Ayala wrote:
>
>
> Well, I just tested the configuration with --download-f2cblaslapack
> --download-blis and the direct solver cholesky with mumps becomes 10x faster
> :D
>
> Thanks!
>
> El mar, 4 mar 2025 a la(s) 3:21 p.m., Emmanuel Ayala (juan...
> On 4 Mar 2025, at 11:17 PM, Emmanuel Ayala wrote:
>
>
> Well, I just tested the configuration with --download-f2cblaslapack
> --download-blis and the direct solver cholesky with mumps becomes 10x faster
> :D
If your problem is big enough (and 700k unknowns should be by far big enough),
y
Well, I just tested the configuration with --download-f2cblaslapack
--download-blis
and the direct solver cholesky with mumps becomes 10x faster :D
Thanks!
El mar, 4 mar 2025 a la(s) 3:21 p.m., Emmanuel Ayala (juan...@gmail.com)
escribió:
> Thanks for the notes.
>
> El mar, 4 mar 2025 a la(s) 2:
Thanks for the notes.
El mar, 4 mar 2025 a la(s) 2:22 p.m., Barry Smith (bsm...@petsc.dev)
escribió:
>
>
> > On Mar 4, 2025, at 2:02 PM, Emmanuel Ayala wrote:
> >
> > Hello everyone.
> >
> > I'm trying to solve a linear system (which comes from 3D FEM with
> structured DM mesh) with a direct sol
> On Mar 4, 2025, at 2:02 PM, Emmanuel Ayala wrote:
>
> Hello everyone.
>
> I'm trying to solve a linear system (which comes from 3D FEM with structured
> DM mesh) with a direct solver. I configured petsc installation with mumps
> (–download-mumps –download-scalapack –download-parmetis –dow
OK, I see.
Thank you very much for the quick response and explanation.
Regards.
El mar, 4 mar 2025 a la(s) 1:27 p.m., Jed Brown (j...@jedbrown.org) escribió:
> Yup, this is expected for direct solvers: numeric factorization for a 3D
> problem scales as O(n^2) while solving with the factors is o
Yup, this is expected for direct solvers: numeric factorization for a 3D
problem scales as O(n^2) while solving with the factors is only O(n^{4/3}). So
you expect it to be slow to factor and fast to solve.
Iterative methods with good preconditioners can be O(n), and tend to be the
choice for la
Hello everyone.
I'm trying to solve a linear system (which comes from 3D FEM with
structured DM mesh) with a direct solver. I configured petsc installation
with mumps (–download-mumps –download-scalapack –download-parmetis
–download-metis, --download-hwloc, without ptscotch) and I have the
followi
