As far as I know, PLT Scheme currently does all of its calculations in double-precision. But, I can control the stored representation of the underlying representation now. It's that structure than can allow us to code efficient butterfly addressing for ffts, etc for the numerical analysis code. We can worry about compiler efficiency in the future.<br> <br><div class="gmail_quote">On Thu, Aug 13, 2009 at 10:01 PM, Robby Findler <span dir="ltr"><<a href="mailto:robby@eecs.northwestern.edu">robby@eecs.northwestern.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;"> I think Doug is saying that the user might want, say, 128 bits of<br> precision in their floating point numbers and asking what to do if<br> mzscheme's floats don't support that. (Or do mz's floats support all<br> those different sizes?)<br> <font color="#888888"><br> Robby<br> </font><div><div></div><div class="h5"><br> On Thu, Aug 13, 2009 at 10:55 PM, Eli Barzilay<<a href="mailto:eli@barzilay.org">eli@barzilay.org</a>> wrote:<br> > On Aug 13, Doug Williams wrote:<br> >> That is the default. But, we want to user to be able to control the<br> >> representation for numerical analysis reasons, too. So, we're<br> >> allowing both.<br> ><br> > But if you don't use external libraries, then you're better off just<br> > checking the inputs, but use Scheme values directly.<br> ><br> > --<br> > ((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:<br> > <a href="http://barzilay.org/" target="_blank">http://barzilay.org/</a> Maze is Life!<br> </div></div><div><div></div><div class="h5">> _________________________________________________<br> > For list-related administrative tasks:<br> > <a href="http://list.cs.brown.edu/mailman/listinfo/plt-scheme" target="_blank">http://list.cs.brown.edu/mailman/listinfo/plt-scheme</a><br> ><br> </div></div></blockquote></div><br>