<div class="gmail_quote">On Fri, Mar 9, 2012 at 11:05, Barry Smith <span dir="ltr"><<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div id=":x8"> I will suggest the following approach. Attach various L to the original Jacobian with a different name for each. In your case L_displacement and L_pressure, have PCFIELDSPLIT propagate them to the Schur matrix. Then have -prefixfromfieldsplit_pc_type <span class="il">lsc</span> -prefixfromfieldplit_lsc_name L_displacement to tell <span class="il">LSC</span> which one to use.<br>
<br>
Barry<br>
<br>
Not perfect since the user could name the wrong <span class="il">lsc</span> for the Schur index sets used but it is better than nothing.</div></blockquote></div><br><div>It's also shameful that it doesn't compose (how does the user know to compose under the name "fieldsplit_outervariable_mg_levels_4_fieldsplit_pressure_LSC_Lp"?) so I just used the field name corresponding to the Schur complement (so the user would compose under "pressure_LSC_L" and "pressure_LSC_Lp"). This looks at both the primary and secondary matrices for LSC_L and LSC_Lp.</div>
<div><br></div><div><a href="http://petsc.cs.iit.edu/petsc/petsc-dev/rev/61b7754a64fa">http://petsc.cs.iit.edu/petsc/petsc-dev/rev/61b7754a64fa</a></div>