[CIG-SHORT] Pylith time stepping viscoelastic

[Matthew Knepley]

On Wed, May 11, 2016 at 4:02 PM, Francisco Delgado <fjd49 at cornell.edu>
wrote:

> Hi Tabrez, it's a 3D simulation and it's slow in the sense that it will
> take more than 1 day to run the 4 years simulation. The only thing I plan
> to change is the chamber time pressurisation, so the stiffness matrix will
> not change. Hence, wow can I change the solver?
>

--petsc.pc_type lu
--petsc.pc_factor_mat_solver_package mumps/superlu

  Thanks,

    Matt


> Thanks
>
> On Wed, May 11, 2016 at 3:55 PM, S Tabrez ALI <stali at geology.wisc.edu>
> wrote:
>
>> Which part is currently slow and is this a 2D or 3D case? You can perhaps
>> try a sparse direct solver. The first solve will be slow but the rest
>> should go fast (assuming the stiffness matrix is not changing).
>>
>>
>> Tabrez
>> ------------------------------
>> *From:* CIG-SHORT <cig-short-bounces at geodynamics.org> on behalf of
>> Francisco Delgado <fjd49 at cornell.edu>
>> *Sent:* Wednesday, May 11, 2016 1:26:25 PM
>> *To:* cig-short at geodynamics.org
>> *Subject:* [CIG-SHORT] Pylith time stepping viscoelastic
>>
>> Dear Pylith developers
>>
>> I'm running a viscoelastic simulation of a pressurised magma chamber
>> surrounded by a viscoelastic shell with a viscosity of 2*1e16 Pa s, a shear
>> modulus of 30 GPa. The latter figures imply a stable time step of 0.0042
>> years which will produce a very slow simulation over a period of 4 years in
>> terms of computing time. Hence I wonder if there a way to increase the
>> simulation speed.
>>
>> Thanks
>>
>> --
>> --------------------------------------------------------
>> Francisco Delgado
>> PhD student in Geological Sciences
>> Cornell University
>>
>> _______________________________________________
>> CIG-SHORT mailing list
>> CIG-SHORT at geodynamics.org
>> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>>
>
>
>
> --
> --------------------------------------------------------
> Francisco Delgado
> PhD student in Geological Sciences
> Cornell University
>
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/cig-short/attachments/20160511/1c869cb9/attachment.html>