<html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    <p><br>
    </p>
    <div class="moz-cite-prefix">Il 15/07/21 14:15, Matthew Knepley ha
      scritto:<br>
    </div>
    <blockquote type="cite" cite="mid:CAMYG4GnZrAcRaipgh9XNUKpoBxP1PSobVdMipkpF72t14OPMhg@mail.gmail.com">
      
      <div dir="ltr">
        <div dir="ltr">On Thu, Jul 15, 2021 at 6:39 AM Matteo Semplice
          <<a href="mailto:matteo.semplice@uninsubria.it" moz-do-not-send="true">matteo.semplice@uninsubria.it</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>
              <p><br>
              </p>
              <div>Il 12/07/21 17:51, Matthew Knepley ha scritto:<br>
              </div>
              <blockquote type="cite">
                <div dir="ltr">
                  <div dir="ltr">On Mon, Jul 12, 2021 at 11:40 AM Matteo
                    Semplice <<a href="mailto:matteo.semplice@uninsubria.it" target="_blank" moz-do-not-send="true">matteo.semplice@uninsubria.it</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>
                        <p>Dear all,</p>
                        <p>    I am experimenting with hdf5+xdmf output.
                          At <a href="https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.xdmf.org%2Findex.php%2FXDMF_Model_and_Format&data=04%7C01%7Cmatteo.semplice%40uninsubria.it%7Cfb0d540c87a64e4ea51d08d9478a4e9b%7C9252ed8bdffc401c86ca6237da9991fa%7C0%7C0%7C637619481622608977%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=sOQskFoMTHdOwjOec0d9npdw%2BiAVbar1SXBuU%2BH8xKc%3D&reserved=0" originalsrc="https://www.xdmf.org/index.php/XDMF_Model_and_Format" shash="COJ2W8d94F8Av8Nl7hPOHkHbcNK26tdZzSiNxaZ45Ahdx0RWbRHwMJ8I39rv8PF9wl1VQ1m6te6nGQSAJ58Qeu0aDwsR7wBYrdVDTM2bg58Fg5GuzCwYumyQatQx5JlygQhDGoEgHNtx2Y8cRBExojXzmIoesjWQnlLgo+lYCDc=" target="_blank" moz-do-not-send="true">https://www.xdmf.org/index.php/XDMF_Model_and_Format</a>
                          I read that "XDMF uses XML to store Light data
                          and to describe the data Model. Either HDF5<a rel="nofollow" href="https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwww.hdfgroup.org%2FHDF5&data=04%7C01%7Cmatteo.semplice%40uninsubria.it%7Cfb0d540c87a64e4ea51d08d9478a4e9b%7C9252ed8bdffc401c86ca6237da9991fa%7C0%7C0%7C637619481622618934%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=akFyfHE4BULSscevnO0MxjMhEByT3USoJFQKsikJSzg%3D&reserved=0" originalsrc="https://www.hdfgroup.org/HDF5" shash="umHxF55MUp5QVOFJQe5ottGXTW2kCspZ4TCWI15XANTRKsp+09sIxgq1dk+HLLYVp6Jr2mw+uebXPk92w8PXWB0W838N9oQyGGlnwehDPk0G5MzfhCYAZFTw+i7qKsXEGCttI40J68EzzXxL4kBCm6PdIQLE+wYwWlGsvcuh+bQ=" target="_blank" moz-do-not-send="true">[3]</a>
                          or binary files can be used to store Heavy
                          data. The data Format is stored redundantly in
                          both XML and HDF5."</p>
                        <p>However, if I call DMView(dmda,hdf5viewer)
                          and then I run h5ls or h5stat on the resulting
                          h5 file, I see no "geometry" section in the
                          file. How should I write the geometry to the
                          HDF5 file?<br>
                        </p>
                        <p>Here below is what I have tried.</p>
                      </div>
                    </blockquote>
                    <div>The HDF5 stuff is only implemented for DMPlex
                      since unstructured grids need to be explicitly
                      stored. You can usually just define the structured
                      grid in the XML</div>
                    <div>without putting anything in the HDF5. We could
                      write metadata so that the XML could be
                      autogenerated, but we have not done that.</div>
                  </div>
                </div>
              </blockquote>
              <p>Thanks for the clarification. It shouldn't be hard to
                produce the XML from my code.<br>
              </p>
              <p>Just another related question: if I call VecView in
                parallel with the HDF5 viewer, I get a single output
                file. Does this mean that data are gathered by one
                process and written or it handles it smartly by
                coordinating the output of all processes to a single
                file?<br>
              </p>
            </div>
          </blockquote>
          <div>This is slightly more complicated than you would expect.
            We have two implementations, one which uses MPI-IO, and one
            which sends</div>
          <div>data from each process to 0, which writes it out. It
            turns out that MPI-IO is sometimes poorly supported or badly
            implemented, so you need</div>
          <div>the fallback.</div>
        </div>
      </div>
    </blockquote>
    <p>Thanks!<br>
    </p>
    <p>On my machine I am compiling from the git repo with
      --download-hdf5, so I have some control, but on clusters I prefer
      to use the available petsc.<br>
    </p>
    Is there a simple way to check which implementation is begin used in
    a run? <br>
    <p>Matteo<br>
    </p>
    <br>
  </body>
</html>