<div class="gmail_quote">On Thu, Sep 13, 2012 at 11:05 AM, Casey DuBois <span dir="ltr"><<a href="mailto:casey@grlug.org" target="_blank">casey@grlug.org</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hey GRLUG,<br>
<br>
Let's say you needed 3000-5000-Cores, and 6-10TB of RAM for 3-6 months.<br>
<br>
What are the options?<br></blockquote><div><br></div><div>The nearest thing I'm aware of is the National Center for Supercomputing at the University of Illinois. But it's not something I'm studied on.</div><div>
<br></div><div>A lot rides on how the job being done, and how the nodes need to communicate. If work units are distributed and collected over, e.g. HTTP, you could try Amazon EC2 or some other cloud services, but I don't know if they have the capacity.</div>
<div><br></div><div>You might prod, e.g. LiquidWeb. When we were there, they had a lot of unfilled space, and you might come to an arrangement that fills another wing...</div></div>-- <br>:wq<br>