<div>This reminds me of a similar issue I had. What approaches do you take for large dense matrix multiplication in MPI, when the matrices are too large to fit into cluster memory? If I hack up something to cache intermediate results to disk, the IO seems to drag everything to a halt and I'm looking for a better solution. I'd like to use some libraries like PETSc, but how would you work around memory limitations like this (short of building a bigger cluster)?
<br><br><br><br> </div><blockquote style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;" class="gmail_quote">>> I don't speak fortran natively, but isn't that array
<br>>> approximately 3.6 TB in size?<br>><br>> Oops, forgot to put the decimal in the right place.<br>><br>> 9915^3 * 8 bits/integer / 1024^3 bytes/GB = 907 GB.<br>><br>> It could be done with a 64 bit kernel. Too big for PAE.
<br><br>Yeah, if you had a box with several hundred memory slots....<br><br>Which I say only semi-sarcastically. They sound like they're coming,<br>they're coming. Who knows, maybe they're here and I'm just out of
<br>touch.<br><br>If it is a sparse matrix, them just maybe one can do something on this<br>scale, but otherwise, well, it's like telling mathematica to go and<br>compute umpty-something factorial -- it will go out, make a herioc
<br>effort, use all the free memory in the universe, and die valiantly<br>(perhaps taking down your computer with it if the kernel happens to need<br>some memory at a critical time when their isn't any). Large scale<br>
computation as a DOS attack...<br></blockquote><br><br clear="all"><br>-- <br>Peter N. Skomoroch<br><a href="mailto:peter.skomoroch@gmail.com">peter.skomoroch@gmail.com</a><br><a href="http://www.datawrangling.com">http://www.datawrangling.com
</a>