MultiThreading and Direct Solvers

The next step towards an efficient Hybrid Solver is to optimize the computation of the Schur complement on each subdomain. Actually this part of the algorithm is the slowest, and it can be very slow.

First of all the reordering is special, since the degrees of freedom on the interior must be numbered first. I use the CAMD ordering by Tim Davis and al., that provide satisfaying enough orderings.

After that, I am concentrating on the code that computes the Schur complement. There is much work to do here. At the begining I was thinking about using MUMPS for this, since it has a subroutine for Schur complement computation and it is multithreaded. By googling around I found there is may be a (slightly) better solution, that would be to implement a sparse Cholesky solver based on the Direct Acyclic Graph (DAG) of the tasks. The computational tasks and their dependencies are expressed as an acyclic graph which is used to organize the thread hierachy to compute the Cholesky decomposition. It seems to be faster than MUMPS on multicore processors.

Comments

Shear waves, medecine and brain

Yesterday evening, too bored by what TV was proposing to me, I decided to watch a conference of Mathias Fink , a french researcher working on multidisciplinary application of waves. Specially shear waves. Here is a brief summary of his talk. In solids, waves have two principal components: compression waves (P-waves for primary) moving in the direction of propagation, and shear waves (S-waves, for secondary) that make ripples in the plane orthogonal to that direction. Since compression waves propagate in the direction of propagation, they move faster than shear waves. Usually ultrasound equipment in medicine only use compressional waves. But since human tissues have a high bulk modulus, the P-wave speed is relatively constant (around 1580 m/s). Human tissues are very stiff if you apply isotropic constraints on them (like pressure of water). However M. Fink and his colleagues proposed a new way to investigate human tissues by first sending a strong compressional wave in ...

We're not playing dices !

Software programming and design is an amazingly complex task. Specially when it concerns numerical applications, that generally require optimization to get results in a reasonable time. For that purpose the basic pattern is usually to write the code as simple as possible, debug it and when it works to begin the optimization process. This asks for some nerves, and patience, two things I usually don't have in real life , but for programming yes. This design pattern makes sense because we are doing things really sequentially. You first wrote some c++ class and then add some feature progressively, and when you come up with essentially different concept (I mean a concept that should be well separated from the first one), you write an other class and so on... Usually the class-writing process follows the solving process you have in mind. For instance you have to make some initialization on your model first, then you declare some variables for the computation, the computation then f...

Circles Packing and Kepler Theorem

By a bored Saturday afternoon I decided to make a little code for my amusment to test the sphere packing theorem. So what I did is to try to pack as many small circles into a big one, by an incremental method which doesn't converge but still gives convincing results. The animation showed here is a vizualisation of this incremental method which tries to minimize overlap one by one. I also tried with circles of various radius. What strikes me is the waves that propagates from circles that have trouble to insert themselves among other. We can see them propagating from one side to an other of the green circle (sorry for colorblind people, those are default gunplot colors).

New Horizons

Search This Blog