On Fri, Jul 16, 2010 at 6:49 PM, Robert Myers <span dir="ltr"><<a href="mailto:rbmyersusa@gmail.com">rbmyersusa@gmail.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="gmail_quote"><div>Whether you like it or not, and whether, even more important, you understand it or not, the architecture of a computer has a big impact on the kinds of models you can implement with any degree of efficiency on the computer. Hanging lots of flops on inadequate bandwidth is popular because you can get a high ranking on the Top 500 list with a minimal expenditure, but it also means that you can't do an important calculation like a global FFT at anything like reasonable efficiency. Flops are almost free. Bandwidth is expensive. Unfortunately, what is almost free (flops) can be nearly useless for, say, an FFT unless you have the bandwidth to move data around globally fast enough.</div>
<br></div></blockquote><div><br>You need to make friends with a guy named Edward Lorenz. Bandwidth has nothing to do with it. Do you even know what this term means?<br> <br></div><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
<div class="gmail_quote"><div>I don't think name-calling and labeling ever helps.</div></div></blockquote><div><br>Then stop doing it.<br> </div></div>-- <br>Chris<br><br><br><br>