I dug up this pdf from Nvidia: <a href="http://www.nvidia.com/docs/IO/43395/tesla_product_overview_dec.pdf">http://www.nvidia.com/docs/IO/43395/tesla_product_overview_dec.pdf</a> <br>
Since I can't imagine coding a graphics card while it serves my X :-) I
supposed one might put the PCIE card in a box with a cheap SVGA for the
least-cost CUDA experiment (one GPU, 128 "thread processors" per
GPU). The deskside is 2 GPU with 3 GB, the rackable is 4 GPU with 6 GB;
they have PCIE adapter cards to talk to your workstation.<br>
<br>
I think one plan might be like the GraPE (Gravity Pipe); maybe one of
the rackables alternating with a CPU MB acting as net host and
fileserver, so each (2-board) node has 4 GPU for array computing.<br>
Peter<br>
<br>
P.S. incidentally, while I was browsing Nvidia, I spec'd out a fantasy
gaming rig. $23K, 256GB of solid state "disk", 3x 1GB video cards, 180W
just to liquid-cool the CPU :-) Maybe next year.<br><br><div><span class="gmail_quote">On 6/19/08, <b class="gmail_sendername">John Hearns</b> <<a href="mailto:john.hearns@streamline-computing.com">john.hearns@streamline-computing.com</a>> wrote:</span><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Wed, 2008-06-18 at 16:31 -0700, Jon Forrest wrote:<br> <br>> Kilian CAVALOTTI wrote:<br> <br> > I'm glad you mentioned this. I've read through much of the information<br> > on their web site and I still don't understand the usage model for<br>
> CUDA. By that I mean, on a desktop machine, are you supposed to have<br> > 2 graphics cards, 1 for running CUDA code and one for regular<br> > graphics? If you only need 1 card for both, how do you avoid the<br>
> problem you mentioned, which was also mentioned in the documentation?<br> <br>Actually, I should imagine Kilian is referring to something else,<br> not the inbuilt timeout which is in the documentation. But I can't speak<br>
for im.<br> <br><br> <br> <br> > Or, if you have a compute node that will sit in a dark room,<br> > you aren't going to be running an X server at all, so you won't<br> > have to worry about anything hanging?<br>
<br> <br>I don't work for Nvidia, so I can't say!<br> But the usage model is as you say - you can prototype applications which<br> will run for a short time on the desktop machine, but long runs are<br> meant to be done on a dedicated back-end machine.<br>
If you want a totally desk-side solution, they sell a companion box<br> which goes alongside and attaches via a ribbon cable. I guess the art<br> here is finding a motherboard with the right number and type of<br> PCI-express slots to take both the companion box and a decent graphics<br>
card for X use.<br> <br><br> <br> ><br> > I'm planning on starting a pilot program to get the<br> > chemists in my department to use CUDA, but I'm waiting<br> > for V2 of the SDK to come out.<br> ><br>
<br> <br>Why wait? The hardware will be the same, and you can dip your toe in the<br> water now.<br> <br><br> <br> _______________________________________________<br> Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org">Beowulf@beowulf.org</a><br>
To change your subscription (digest mode or unsubscribe) visit <a href="http://www.beowulf.org/mailman/listinfo/beowulf">http://www.beowulf.org/mailman/listinfo/beowulf</a><br> </blockquote></div><br>
<br />--
<br />This message has been scanned for viruses and
<br />dangerous content by
<a href="http://www.mailscanner.info/"><b>MailScanner</b></a>, and is
<br />believed to be clean.