<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><div><div><blockquote type="cite"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex; "><div id=":lw"></div></blockquote></div></blockquote><blockquote type="cite"><blockquote type="cite"><div class="gmail_quote"><blockquote class="gmail_quote" style="margin-top: 0px; margin-right: 0px; margin-bottom: 0px; margin-left: 0.8ex; border-left-width: 1px; border-left-color: rgb(204, 204, 204); border-left-style: solid; padding-left: 1ex; position: static; z-index: auto; "><div id=":lw">What I need is to have two different global ids over each interface entity,<br>one global id created by the owner and one created by the (unique) sharer,<br>where the same intereface entity is not owned, and then communicated to the owner.<br></div></blockquote><div><br></div></div></blockquote>Is this because you are putting different data in those locations?</blockquote></div><div><br></div><div>Yes, my goal is to create a PETSc ghosted vector to communicate the shape function evaluations</div><div>at the interface quadrature points. </div><div>Since I'm working with agglomerated elements, this way I can avoid agglomerated elements ghost cells</div><div>meaning a lot of (an unpredictable lot of) layers of sub cells.</div><div><br></div><div>By the way, I've read your last paper on Journal of Scientific Computing about efficient nonlinear solvers. Very interesting!</div><div>The low order preconditioner with AMG seems to work great.</div><div><br></div><div>Greetings</div><div>Lorenzo</div><div> </div><div><br></div><div><br></div><div>On Mar 7, 2012, at 12:31 AM, Jed Brown wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite"><div class="gmail_quote">On Tue, Mar 6, 2012 at 15:06, Lorenzo Alessio Botti <span dir="ltr"><<a href="mailto:ihabiamx@yahoo.it">ihabiamx@yahoo.it</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div id=":lw">What I need is to have two different global ids over each interface entity,<br>
one global id created by the owner and one created by the (unique) sharer,<br>
where the same intereface entity is not owned, and then communicated to the owner.<br></div></blockquote><div><br></div><div>Is this because you are putting different data in those locations?</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div id=":lw">
<br>
I should be able to create an id (global_id_1) for interface entities,<br>
such that the interface entity has a different id on the owner and the sharer,<br>
modifying a bit the assign_global_ids function.<br>
Thats the easy part.<br>
Then I need to communicate global_id_1 (the owner communicates to the sharer and the sharer to the owner) and store it in global_id_2.<br>
I guess this is what you mean when you say "store in another tag on all sharers".<br>
<br>
This is what happens with ghost entities ids where global_id_2 is simply the global id of the ghost cell. Thats why I say that I'd need a ghost<br>
interface entity. Might tag_reduce help me?</div></blockquote></div><br><div>Tim, this is the operation I have mentioned to you before and that came up again last week. I can explain in more detail tomorrow if you have time.</div>
</blockquote></div><br></body></html>