<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Sat, Feb 1, 2014 at 1:16 PM, David Liu <span dir="ltr"><<a href="mailto:daveliu@mit.edu" target="_blank">daveliu@mit.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div dir="ltr">I see. Does that s^2 memory scaling mean that sparse direct solvers are not meant to be used beyond a certain point? I.e. if the supercomputer I'm using doesn't have enough memory per core to store even a single row of the factored matrix, then I'm out of luck?</div>
</blockquote><div><br></div><div>Yes exactly</div><div><br></div><div> Matt</div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="gmail_extra">
<div class="gmail_quote">On Fri, Jan 31, 2014 at 9:51 PM, Jed Brown <span dir="ltr"><<a href="mailto:jed@jedbrown.org" target="_blank">jed@jedbrown.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div>David Liu <<a href="mailto:daveliu@mit.edu" target="_blank">daveliu@mit.edu</a>> writes:<br>
<br>
> Hi, I'm solving a 3d problem with mumps. When I increased the grid size to<br>
> 70x60x20 with 6 unknowns per point, I started noticing that the program was<br>
> crashing at runtime at the factoring stage, with the mumps error code:<br>
><br>
> -17 The internal send buffer that was allocated dynamically by MUMPS on the<br>
> processor is too small.<br>
> The user should increase the value of ICNTL(14) before calling MUMPS again.<br>
><br>
> However, when I increase the grid spacing in the z direction by about 50%,<br>
> this crash does not happen.<br>
><br>
> Why would how much memory an LU factorization uses depend on an overall<br>
> numerical factor (for part of the matrix at least) like this?<br>
<br>
</div></div>I'm not sure exactly what you're asking, but the complexity of direct<br>
solves depend on the minimal vertex separators in the sparse<br>
matrix/graph. Yours will be s=60*20*6 (more if your stencil needs<br>
second neighbors). The memory usage scales with s^2 and the<br>
factorization time scales with s^3.<br>
</blockquote></div><br></div>
</blockquote></div><br><br clear="all"><div><br></div>-- <br>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>