for the non-overlapping version probably this one will be more apt: https://hal.archives-ouvertes.fr/hal-01244511/document
On Saturday, June 8, 2019 at 7:03:59 PM UTC+2, Tom Mathew wrote: > > Dear all, > Thanks again for the reply. I was planing to follow actually this work > https://arxiv.org/abs/1711.03789, where there is overlap. Though there > exists non-overlapping versions too > http://www.montefiore.ulg.ac.be/~geuzaine/preprints/ddm_double_sweep_preprint.pdf > I am a PhD student working in a thermal-fluids lab trying to model > electromagnetic scattering for radiative transport via indefinite Maxwell's > equations. My triangulations mostly comes from tomographied (micro-scale > imaging techniques) STL files from industrial applications. > I am more interested in the finite element-linear algebra world, so this > attempt is just for me to introduce myself as my helloworld to the linear > algebra scientific computing community. I want to explore a little bit the > underlying thing abstracting iterative Krylov methods, and I have at hand > an indefinite Maxwell's problem, for which every state of the art linear > algebra packages seems near obsolete. > Acknowledging that, everything is going fine for my thesis with FreeFEM++, > and thanks to its ffddm framework, which seems quite efficient for my > scattering problems. I have tried the hypre AMS quite briefly, regretting > my poor mathematical background (I didn't knew it was only meant for > definite Maxwell's problems. Only came to know after implementing it with > MFEM. It was quite good on my workstation, but couldn't scale at the > cluster). > So seeing for myself after the state of the art packages couldn't help me, > I was flattered to give it a try, may be in my free time, as something > which I aspire. > > Regarding your message, I guess I need sometime implementing my parallel > solver in dealii. I need sometime also to undrestand how dealii functions. > After I'll try to understand the linear algebra classes and try to inherit > may be. > I think the right direction for me is to try overlapping DDM, as > https://arxiv.org/abs/1711.03789, discusses the influence of "generous > overlap" (I couldn't read it completely though, it's highly mathematical). > Thanks very much for your reply, I will get in touch if I am stuck at > some point. > Best regards, > Tom > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/4de8da84-ce1f-4d6f-b197-9fd7de345b73%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
