<div class="gmail_quote">On Thu, Jan 6, 2011 at 09:46, Hamid M. <span dir="ltr"><<a href="mailto:spam.wax@gmail.com">spam.wax@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div id=":t6">1- Can I use PETSc to build/populate my LHS matrix on different nodes<br>
of a cluster (as opposed to constructing it on a single node and then<br>
distributing it) ?<br></div></blockquote><div><br></div><div>You should always do this.</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;"><div id=":t6">
<br>
2- Are there optimized parallel solvers for dense matrices in PETSc ?<br></div></blockquote><div><br></div><div>The current implementation uses PLAPACK.</div><div><br></div><div><br></div><div>You might also consider using FMM (see <a href="http://barbagroup.bu.edu/Barba_group/PetFMM.html">http://barbagroup.bu.edu/Barba_group/PetFMM.html</a>) since explicit storage of dense matrices cannot scale.</div>
</div>