AO is not scalable. I am pretty sure this is stated explicitly in the<br>
documentation. We are working on a parallel reordering construct,<br>
but it is not yet finished.<br>
<br>
Matt<br><br><div><span class="gmail_quote">On 7/6/06, <b class="gmail_sendername">Laslo Tibor Diosady</b> <<a href="mailto:diosady@mit.edu">diosady@mit.edu</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I have an array of values which are in my application ordering and I wish<br>to reorder these to the petsc ordering. To do this I use<br>AOApplicationToPetscPermuteInt, or PermuteReal, which works really well in<br>serial but not so much in parallel.
<br><br>In parallel I want to have it so that each processor has it's local<br>portion of the array, so that the sum of the lengths of the arrays on each<br>processor add up to the sum of the indices in the AO ordering.<br>
<br>However when I use AOApplicationToPetscPermute I get a segmentation fault,<br>which I've traced to the fact that AOApplicationToPetscPermute assumes<br>each processor has an array which is the number of indices in the AO
<br>ordering as opposed to only the length of the local portion of the AO.<br><br>The permutations that I am doing are all local to each processor, but my<br>AO has been defined globally since elsewhere in my code I need access to
<br>off processor entries.<br><br>Do you have any suggestions on how I can solve this problem with the least<br>amount of headache.<br><br>Thanks,<br><br>Laslo<br><br></blockquote></div><br><br clear="all"><br>-- <br>"Failure has a thousand explanations. Success doesn't need one" -- Sir Alec Guiness