<div class="gmail_quote">On Fri, May 4, 2012 at 8:58 PM, Oliver Stueker <span dir="ltr"><<a href="mailto:ostueker@gmail.com" target="_blank">ostueker@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi all,<br>
<br>
This week Nvidia has also announced the GeForce GTX 690 with two<br>
Kepler Chips [1].<br>
That will make a total of 3072 CUDA cores @ 915MHz to 1019MHz and will<br>
perform like two GTX 680s in SLI mode [2], but more energy efficient.<br>
<br>
I'd really be interested if parallelization would be a bottleneck for<br>
the Kepler chips compared to the Fermi ones. Has anyone already run<br>
some benchmarks on Kepler?<br></blockquote><div><br></div><div>If you stick the GTX 690 into PCi-E 3.0, it should have a performance quite close to 2x GTX 680 on PCI-E 2.0. Don't expect to gain more than 5% from having GTX 680 on PCI-E 2.0 vs PCI-E 3.0.</div>
<div class="gmail_quote"><br></div>Cheers,<br clear="all">--<br>Szilárd<br><div> </div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Best,<br>
Oliver<br>
<br>
<br>
[1] <a href="http://nvidianews.nvidia.com/Releases/NVIDIA-Unveils-GeForce-GTX-690-Dual-Graphics-Card-Combines-World-s-Fastest-Gaming-Performance-With-Sleek-Sexy-Design-7c1.aspx" target="_blank">http://nvidianews.nvidia.com/Releases/NVIDIA-Unveils-GeForce-GTX-690-Dual-Graphics-Card-Combines-World-s-Fastest-Gaming-Performance-With-Sleek-Sexy-Design-7c1.aspx</a><br>
<br>
[2] <a href="http://www.geforce.com/whats-new/articles/article-keynote/" target="_blank">http://www.geforce.com/whats-new/articles/article-keynote/</a><br>
<div class="HOEnZb"><div class="h5"><br>
On Wed, May 2, 2012 at 10:16 AM, Peter C. Lai <<a href="mailto:pcl@uab.edu">pcl@uab.edu</a>> wrote:<br>
> On 2012-05-02 06:13:04PM +0200, Mirco Wahab wrote:<br>
>> Hello Peter,<br>
>><br>
>> Am 02.05.2012 17:44, schrieb Peter C. Lai:<br>
>><br>
>> > You can wait for ivy bridge, then stick some Kepler GPUs (nvidia gtx<br>
>> > 680) in it. That should max the performance. asm is pretty much stagnant<br>
>> > for general purpose procs since core2 came out.<br>
>><br>
>> Is this true?<br>
>><br>
>> To my knowledge, the Fermi GPU (GF-110, eg. GTX-580) is<br>
>> a 16 processor (streaming multiprocessor, SM) system, each<br>
>> processor having 32 cores running at 1.5 GHz, and the<br>
>> Kepler-1 (GK-104, eg. GTX-680) an 8 processor system<br>
>> with 192 cores per processor at 1GHz.<br>
>> Because on the Fermi each SM has more L1 cache than<br>
>> each SM on the Kepler-1 and because it might be harder<br>
>> to saturate 192 cores in a compute-scenario, I'd expect<br>
>> the 580 (GF-110) to be significant(?) faster in the<br>
>> next Gromacs (4.6). Maybe somebody tested this already.<br>
>><br>
>> (Here in Germany, I can by GTX-580/3GB for ~325€ + Tax.)<br>
>><br>
>> Regards,<br>
>><br>
>> M.<br>
><br>
> You might be. I do have a collaborator in France who chooses to use GTX 580<br>
> with NAMD and is eschewing the Keplers. I think I saw a bench somewhere<br>
> for Kepler compute; not sure if I saw that much difference (but you will<br>
> be paying the price though). In that case stick with Fermis for now :)<br>
--<br>
</div></div><div class="HOEnZb"><div class="h5">gmx-users mailing list <a href="mailto:gmx-users@gromacs.org">gmx-users@gromacs.org</a><br>
<a href="http://lists.gromacs.org/mailman/listinfo/gmx-users" target="_blank">http://lists.gromacs.org/mailman/listinfo/gmx-users</a><br>
Please search the archive at <a href="http://www.gromacs.org/Support/Mailing_Lists/Search" target="_blank">http://www.gromacs.org/Support/Mailing_Lists/Search</a> before posting!<br>
Please don't post (un)subscribe requests to the list. Use the<br>
www interface or send it to <a href="mailto:gmx-users-request@gromacs.org">gmx-users-request@gromacs.org</a>.<br>
Can't post? Read <a href="http://www.gromacs.org/Support/Mailing_Lists" target="_blank">http://www.gromacs.org/Support/Mailing_Lists</a><br>
</div></div></blockquote></div><br>