<div dir="ltr">Thanks guys. I will definitely spend sometime and consider these packages. The truth is I really like the structure of PETSc and I want to stick with it as much as possible.<div><br></div><div>If I can use any of these packages to replace for my data structure in the grid generation phase, I'll probably be able to port most of my code from serial into parallel. Hopefully it will be good enough (in the long term) that I can eventually publish it and be of help to the community.</div>
<div><br></div><div>Thanks,</div><div>Mohammad<br><br><div class="gmail_quote">On Sat, Oct 30, 2010 at 10:06 AM, Pearl Flath <span dir="ltr"><<a href="mailto:pflath@ices.utexas.edu">pflath@ices.utexas.edu</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">Mohammad,<br>You could also look into p4est, which also interfaces with the deal II library (and deal II has wrappers for calling either PETSc or Trilinos). <br>
<br>"The p4est software library enables the dynamic management of
a collection of adaptive octrees, conveniently called a forest of
octrees. p4est is designed to work in parallel and scale to
hundreds of thousands of processor cores." <a href="http://p4est.org/" target="_blank">http://p4est.org/</a> <br><br>Best,<br><font color="#888888">Pearl Flath</font><div><div></div><div class="h5"><br><br><div class="gmail_quote">
On Sat, Oct 30, 2010 at 10:32 AM, Aron Ahmadia <span dir="ltr"><<a href="mailto:aron.ahmadia@kaust.edu.sa" target="_blank">aron.ahmadia@kaust.edu.sa</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204, 204, 204);padding-left:1ex">Mohammad,<br>
<br>
I am sure some of the other users and developers here will have<br>
different opinions on the correct way to approach this.<br>
<br>
It sounds like you may benefit even more from investigating several of<br>
the packages that manage meshes and grids on parallel architectures.<br>
There are some very general toolkits for managing adaptive grids and<br>
meshes out there, one could start with Sieve or deal.ii. If one of<br>
these packages is suitable for you, I strongly suggest you consider<br>
reusing as much of their frameworks as possible to avoid "rewriting<br>
the wheel" so to speak.<br>
<br>
Cheers,<br>
Aron<br>
<br>
On Sat, Oct 30, 2010 at 9:07 AM, Mohammad Mirzadeh <<a href="mailto:mirzadeh@gmail.com" target="_blank">mirzadeh@gmail.com</a>> wrote:<br>
> Aron,<br>
> Thanks for the quick reply. It's really great that PETSc has such an awesome<br>
> community.<br>
> Anyway, I am working on adaptive Cartesian grids for which I use<br>
> Octree/Quadtree data structures. Naturally, then, I have components like<br>
> cells, nodes, neighbors, child/parent, etc and my whole domain is consisted<br>
> of arrays of these types. That is, if I happen to have 100 cells and 200<br>
> nodes, for example, I create an array for the whole domain by calling,<br>
> Array<Cell> *CellArray = new Array<Cell> [100];<br>
> Array<Node> *NodeArray = new Array<Node> [200];<br>
> Now the problem is I want to be able to distribute this in parallel and have<br>
> an array of cells or nodes. I understand that one of doing this is to change<br>
> my data structure such that is consistent with PETSc only accepting double.<br>
> I was hoping I could prevent that by using a package that allow for<br>
> templates. That being said, I am not an expert on PETSc by any measure! As a<br>
> result I highly appreciate any ideas and comments if you think this is<br>
> possible to do with PETSc.<br>
> All the best,<br>
> Mohammad<br>
><br>
> On Fri, Oct 29, 2010 at 10:44 PM, Aron Ahmadia <<a href="mailto:aron.ahmadia@kaust.edu.sa" target="_blank">aron.ahmadia@kaust.edu.sa</a>><br>
> wrote:<br>
>><br>
>> Dear Mohammad,<br>
>><br>
>> As a user of PETSc for the last 8 years, since my days as an<br>
>> undergraduate, and now as a professional staff scientist at a<br>
>> supercomputing center, I can say with some confidence that there are<br>
>> no codes like PETSc in C++ or any other language in terms of quality<br>
>> of implementation, documentation, and support. Can you tell us a<br>
>> little more about your current implementation? It is true that PETSc<br>
>> does not support multiple types in the same build, but you do get your<br>
>> choice of floating-point values and real or complex types. Also,<br>
>> PETSc has several C++ components within it, and one of the supported<br>
>> ways of building it is in 'C++' mode, see -c-language in the configure<br>
>> options.<br>
>><br>
>> If you insist on departing us (we'll miss you), I suggest you look at<br>
>> Sandia's Trilinos package: <a href="http://trilinos.sandia.gov/" target="_blank">http://trilinos.sandia.gov/</a><br>
>><br>
>> Good Luck,<br>
>> Aron<br>
>><br>
>> On Sat, Oct 30, 2010 at 7:54 AM, Mohammad Mirzadeh<br>
>> <<a href="mailto:m.mirzadeh@engineering.ucsb.edu" target="_blank">m.mirzadeh@engineering.ucsb.edu</a>> wrote:<br>
>> > Dear all,<br>
>> > I was wondering if anyone can refer me to a package similar to PETSc but<br>
>> > that is written in c++? Right now I have a large code written in c++ for<br>
>> > doing CFD simulations that I need to transform from serial to parallel.<br>
>> > Initially I was thinking of PETSc and tried using it but found that<br>
>> > PETSc is<br>
>> > written in C and thus does not allow to have arrays (in parallel) of<br>
>> > arbitrary type. I have a big data structure and it is much easier for me<br>
>> > to<br>
>> > retain the current structure and form of the code. As a result I was<br>
>> > wondering if you guys know of any similar package in C++ ?(in the sense<br>
>> > that<br>
>> > it can provide with efficient linear solvers in parallel while hiding<br>
>> > most<br>
>> > of MPI from the user)<br>
>> > I could think of HYPRE but then again I am not sure it is written in<br>
>> > C++.<br>
>> > Thanks,<br>
>> > Mohammad<br>
><br>
><br>
</blockquote></div><br>
</div></div></blockquote></div><br></div></div>