<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Mon, Feb 13, 2017 at 5:51 AM, Barletta, Ivano <span dir="ltr"><<a href="mailto:ibarletta@inogs.it" target="_blank">ibarletta@inogs.it</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div><div><div><div><div><div><div><div><div><div><div><div><div><div><div><div>Hello PETSc users<br><br></div>I was wondering if you've got any suggestion for the issue<br></div>mentioned in the topic. <br><br></div>I'm solving a linear system on my unstructured grid that has its<br></div>own partitioning. When I set up the linear system in PETSc I <br></div>use the model partitioning whose grid points might not be<br></div>evenly balanced (generally is not) over the processors. <br><br></div>So far I use VecScatter and IS to re-order the solution vector<br></div>with the application numbering. <br><br></div>What I aim at, now, is to let PETSc to decide the linear system <br></div>partitioning with PETSC_DECIDE local dimension and<br>see if I get some gain in performances, but in <br></div>this case I have to scatter between vectors that, at any rate, have <br></div>different parallel layouts (one of the requirements of <br></div>VecScatter is to have 2 Vecs with same layouts)<br></div></div></div></div></div></blockquote><div><br></div><div>That is exactly what VecScatter does. It does not require the same layout.</div><div><br></div><div>Example:</div><div> proc 0 proc1</div><div>Vec A 0 1 2 3</div><div>Vec B 0 1 2 3</div><div><br></div><div>You create a VecScatter A --> B with indices</div><div> proc 0 proc 1</div><div>from 0 1 2 3</div><div>to 3 2 1 0</div><div><br></div><div>and it will reverse the vector into the new layout.</div><div><br></div><div> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div><div><div><div></div>Does anyone of you have ever faced a similar problem? <br></div>any suggestion about how to circumvent this?<br><br></div><div>Feel free to ask for further information if my explanation is<br></div><div>not clear enough<br></div><div><br></div>Thanks in advance<span class="HOEnZb"><font color="#888888"><br></font></span></div><span class="HOEnZb"><font color="#888888">Ivano<br></font></span></div>
</blockquote></div><br><br clear="all"><div><br></div>-- <br><div class="gmail_signature" data-smartmail="gmail_signature">What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead.<br>-- Norbert Wiener</div>
</div></div>