<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
  </head>
  <body>
    In our FEA code we perform the rotations at the local level, before
    assembly so that it is easy to apply the boundary conditions, then
    unrotate locally after solution to get the usual Cartesian
    components.  Somehow this seems more efficient than doing this
    globally, but perhaps I am missing something.<br>
    -sanjay<br>
    <pre class="moz-signature" cols="72">

</pre>
    <div class="moz-cite-prefix">On 5/31/21 9:33 AM, Matthew Knepley
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAMYG4GmjmbTg8_6TU9FGEfkwooRrfeJtiwtxaYwg-GDAQF2MAQ@mail.gmail.com">
      <meta http-equiv="content-type" content="text/html; charset=UTF-8">
      <div dir="ltr">
        <div dir="ltr">On Mon, May 31, 2021 at 11:12 AM Stefano Zampini
          <<a href="mailto:stefano.zampini@gmail.com"
            moz-do-not-send="true">stefano.zampini@gmail.com</a>>
          wrote:<br>
        </div>
        <div class="gmail_quote">
          <blockquote class="gmail_quote" style="margin:0px 0px 0px
            0.8ex;border-left:1px solid
            rgb(204,204,204);padding-left:1ex">
            <div dir="ltr">Mike
              <div><br>
              </div>
              <div>as long as P is a sparse matrix with compatible rows
                and cols (i.e. rows(P)= cols(A) = rows (A)) , MatPtAP
                will compute the result. </div>
            </div>
          </blockquote>
          <div><br>
          </div>
          <div>Stefano and Mark are correct. This will work.</div>
          <div><br>
          </div>
          <div>I implemented the same thing in my code in a different
            way. I put this transformation into the mapping between
            local and global vector spaces. The global degrees of</div>
          <div>freedom are the ones you want for boundary conditions
            (normal and tangential to the boundary), and I eliminate the
            ones that are constrained. The local degrees of</div>
          <div>freedom are the normal Caresian ones, and these are used
            for assembly. The map is used when I
            execute DMGlobalToLocal() and DMLocalToGlobal(). There is an</div>
          <div>example of me doing this in SNES ex71, Poiseuille flow in
            a tilted channel.</div>
          <div><br>
          </div>
          <div>  Thanks,</div>
          <div><br>
          </div>
          <div>      Matt</div>
          <div> </div>
          <blockquote class="gmail_quote" style="margin:0px 0px 0px
            0.8ex;border-left:1px solid
            rgb(204,204,204);padding-left:1ex">
            <div class="gmail_quote">
              <div dir="ltr" class="gmail_attr">Il giorno lun 31 mag
                2021 alle ore 16:52 Mark Adams <<a
                  href="mailto:mfadams@lbl.gov" target="_blank"
                  moz-do-not-send="true">mfadams@lbl.gov</a>> ha
                scritto:<br>
              </div>
              <blockquote class="gmail_quote" style="margin:0px 0px 0px
                0.8ex;border-left:1px solid
                rgb(204,204,204);padding-left:1ex">
                <div dir="ltr">
                  <div dir="ltr"><br>
                  </div>
                  <br>
                  <div class="gmail_quote">
                    <div dir="ltr" class="gmail_attr">On Mon, May 31,
                      2021 at 9:20 AM Michael Wick <<a
                        href="mailto:michael.wick.1980@gmail.com"
                        target="_blank" moz-do-not-send="true">michael.wick.1980@gmail.com</a>>
                      wrote:<br>
                    </div>
                    <blockquote class="gmail_quote" style="margin:0px
                      0px 0px 0.8ex;border-left:1px solid
                      rgb(204,204,204);padding-left:1ex">
                      <div dir="ltr">Hi PETSc team:
                        <div><br>
                        </div>
                        <div>I am considering implementing a skew roller
                          boundary condition for my elasticity problem.
                          The method is based on this journal paper: <a
href="http://inside.mines.edu/~vgriffit/pubs/All_J_Pubs/18.pdf"
                            target="_blank" moz-do-not-send="true">http://inside.mines.edu/~vgriffit/pubs/All_J_Pubs/18.pdf</a></div>
                        <div><br>
                        </div>
                        <div>Or you may find the method in the attached
                          Bathe's slides, pages 9 -10.</div>
                        <div><br>
                        </div>
                        <div>Roughly speaking, a (very) sparse matrix T
                          will be created which takes the shape [ I, O;
                          O, R], where R is a 3x3 rotation matrix. And
                          the original linear problem K U = F will be
                          modified into (T^t K T) (T^t U) = T^t F. In
                          doing so, one can enforce a roller boundary
                          condition on a slanted surface.</div>
                        <div><br>
                        </div>
                        <div>I think it can be an easy option if I can
                          generate the T matrix and do two matrix
                          multiplications to get T^t K T. I noticed that
                          there is a MatPtAP function. Yet, after
                          reading a previous discussion, it seems that
                          this function is not designed for this
                          purposes (<a
href="https://lists.mcs.anl.gov/pipermail/petsc-users/2018-June/035477.html"
                            target="_blank" moz-do-not-send="true">https://lists.mcs.anl.gov/pipermail/petsc-users/2018-June/035477.html</a>).</div>
                      </div>
                    </blockquote>
                    <div><br>
                    </div>
                    <div>Yes, and no. It is motivated and optimized for
                      a Galerkin coarse grid operator for AMG solvers,
                      but it is a projection and it should be fine. If
                      not, we will fix it.</div>
                    <div><br>
                    </div>
                    <div>We try to test our methods of "empty" operators
                      , but I don't know if MatPtAP has ever been tested
                      for super sparse P. Give it a shot and see what
                      happens.</div>
                    <div><br>
                    </div>
                    <div>Mark</div>
                    <div> </div>
                    <blockquote class="gmail_quote" style="margin:0px
                      0px 0px 0.8ex;border-left:1px solid
                      rgb(204,204,204);padding-left:1ex">
                      <div dir="ltr">
                        <div><br>
                        </div>
                        <div>I assume I can only call MatMatMult &
                          MatTransposeMatMult to do this job, correct?
                          Is there any existingly PETSc function to do
                          T^t K T in one call?</div>
                        <div><br>
                        </div>
                        <div>Thanks,</div>
                        <div><br>
                        </div>
                        <div>Mike</div>
                        <div><br>
                        </div>
                      </div>
                    </blockquote>
                  </div>
                </div>
              </blockquote>
            </div>
            <br clear="all">
            <div><br>
            </div>
            -- <br>
            <div dir="ltr">Stefano</div>
          </blockquote>
        </div>
        <br clear="all">
        <div><br>
        </div>
        -- <br>
        <div dir="ltr" class="gmail_signature">
          <div dir="ltr">
            <div>
              <div dir="ltr">
                <div>
                  <div dir="ltr">
                    <div>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><br>
                    </div>
                    <div><a href="http://www.cse.buffalo.edu/~knepley/"
                        target="_blank" moz-do-not-send="true">https://www.cse.buffalo.edu/~knepley/</a><br>
                    </div>
                  </div>
                </div>
              </div>
            </div>
          </div>
        </div>
      </div>
    </blockquote>
    <br>
  </body>
</html>