<html><head></head><body><div style="color:#000; background-color:#fff; font-family:times new roman, new york, times, serif;font-size:16px"><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">Apologies for being not sufficient clear in my previous message. </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr"><br></div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">I would like to be able to Galerkin coarsen A^h to obtain A^H </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">and to separately Galerkin coarsen M^h to obtain M^H. </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr"><br></div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">So, yes, the way in which I currently (partially) understand your </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">description of the new DMCreateMatrices would do the job. </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr"><br></div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">What is a sensible way to proceed? </div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr"><br></div><div id="yui_3_16_0_1_1468907431186_61044" dir="ltr">Thanks, Domenico. </div><div class="qtdSeparateBR" id="yui_3_16_0_1_1468907431186_61079"><br></div><div class="yahoo_quoted" id="yui_3_16_0_1_1468907431186_61094" style="display: block;"> <div style="font-family: times new roman, new york, times, serif; font-size: 16px;" id="yui_3_16_0_1_1468907431186_61093"> <div style="font-family: HelveticaNeue, Helvetica Neue, Helvetica, Arial, Lucida Grande, sans-serif; font-size: 16px;" id="yui_3_16_0_1_1468907431186_61092"> <div dir="ltr" id="yui_3_16_0_1_1468907431186_61091"> <font size="2" face="Arial" id="yui_3_16_0_1_1468907431186_61090"> <hr size="1" id="yui_3_16_0_1_1468907431186_61368"> <b id="yui_3_16_0_1_1468907431186_61098"><span style="font-weight:bold;" id="yui_3_16_0_1_1468907431186_61097">From:</span></b> Lawrence Mitchell <lawrence.mitchell@imperial.ac.uk><br> <b id="yui_3_16_0_1_1468907431186_61096"><span style="font-weight: bold;" id="yui_3_16_0_1_1468907431186_61095">To:</span></b> domenico lahaye <domenico_lahaye@yahoo.com> <br><b id="yui_3_16_0_1_1468907431186_61089"><span style="font-weight: bold;" id="yui_3_16_0_1_1468907431186_61088">Cc:</span></b> petsc-users@mcs.anl.gov<br> <b><span style="font-weight: bold;">Sent:</span></b> Thursday, July 21, 2016 11:25 AM<br> <b><span style="font-weight: bold;">Subject:</span></b> Re: [petsc-users] Regarding ksp ex42 - Citations<br> </font> </div> <div class="y_msg_container" id="yui_3_16_0_1_1468907431186_61147"><br>[Reintroducing petsc-users in cc]<br clear="none"><br clear="none">On 21/07/16 10:18, domenico lahaye wrote:<br clear="none">> Thanks Lauwrence. <br clear="none">> <br clear="none">> Does the fact that the coarse level preconditioner M^H should be<br clear="none">> constructed <br clear="none">> by Galerkin coarse (rather then rediscretization) cause additional<br clear="none">> wrinkles? <br clear="none"><br clear="none">Do you want to rediscretise A, but use a galerkin coarse grid M?<br clear="none"><br clear="none">If so, that is currently unsupported in PCMG: In PCSetUp_MG (mg.c,<br clear="none">line 660 or so):<br clear="none"><br clear="none">if (mg->galerkin == 1) {<br clear="none"> /* Currently only handle case where mat and pmat are the same on<br clear="none">coarser levels */<br clear="none"> ...<br clear="none">}<br clear="none"><br clear="none">I guess if you're managing the creation of the coarse grid operators<br clear="none">yourself via KSPSetComputeOperators and a putative (new)<br clear="none">DMCreateMatrices then you'd have the flexibility to do separate things<br clear="none">for A and M (including, I think, galerkin coarse M). Since you have<br clear="none">access to the DM hierarchy inside your compute operators.<br clear="none"><br clear="none">Make sense?<div class="yqt6117309937" id="yqtfd07502"><br clear="none"><br clear="none">Lawrence<br clear="none"></div><br><br></div> </div> </div> </div></div></body></html>