<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><blockquote type="cite"><p>How is the matrix assembled?</p></blockquote>In a deterministic way.<br><blockquote type="cite"><p>How different? </p></blockquote><div>Not very much. The last residual can vary in its trailing digits, e.g. from 6.5819e-7 to 6.5820e-7</div><blockquote type="cite"><p>You should usually not expect deterministic associativity. </p></blockquote><div>Ok, I see. Thanks!</div><br><blockquote type="cite"><p>If your results are sensitive, you may be able to make it deterministic (with extra work and choice of package), but you should probably reformulate the problem</p></blockquote><blockquote type="cite">
<div class="gmail_quote">On May 30, 2012 9:21 AM, "Hui Zhang" <<a href="mailto:mike.hui.zhang@hotmail.com">mike.hui.zhang@hotmail.com</a>> wrote:<br type="attribution"><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; ">
<br>
> I suspect there are some randomness from external matsolverpackage. I use a parallel matrix<br>
> as preconditioner, and the preconditioner is exactly LU factored, then I saw the KSP (gmres)<br>
> residuals are different from different runs.<br>
<br>
Besides, for sequential matrix, there is no such phenomenon, only for parallel matrix.<br>
<br>
><br>
> Is it true of randomness from matsolverpackage? Thanks a lot!<br>
<br>
</blockquote></div>
</blockquote></div><br></body></html>