<div dir="ltr"><br><div>It appears to me that the numbers posted on that page for the card you are testing are with ECC off? I know you are asking the question "what if", but the current test isn't even apples-to-apples. </div><div><br></div><div>We want best price-performance for our codes. We have not gone down the road of CUDA because that would lock us out of other technologies (still hoping that OpenACC and OpenMP merge). We have never ruled out AMD, but we have never seen an argument to try it.</div><div><br></div><div>Craig</div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Mar 4, 2015 at 12:07 PM, C Bergström <span dir="ltr"><<a href="mailto:cbergstrom@pathscale.com" target="_blank">cbergstrom@pathscale.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">This is part shameless self promotion, but also meant to stir some discussion..<br>
<br>
Take a look at some of the recent top results for SPEC ACCEL<br>
<a href="http://spec.org/accel/results/accel_acc.html" target="_blank">http://spec.org/accel/results/accel_acc.html</a><br>
/* It may not seem like much, but I can say that is the visible result<br>
to a heck of a lot of work */<br>
<br>
I have to word this carefully, but after that was submitted we fixed a<br>
cooling issue and the score went higher.. expect better public results<br>
in 2 weeks. (Anyone want to party at GTC?)<br>
<br>
Frankly, it's visible we still aren't the best performance on some<br>
codes.. Those issues basically fall into two camps and we're working<br>
on it.<br>
-----------<br>
Having said this..<br>
Would anyone build an AMD GPU cluster if<br>
1) The GPU was 30-40% faster than a K40<br>
2) It cost less<br>
3) OpenACC/OpenMP4/CUDA programming models were high performance and supported<br>
4) More oncard ram<br>
-----------<br>
The s9150 is a top end card and priced relatively in the same ballpark<br>
as the competition..<br>
<br>
Forgive me for sounding like a salesperson.<br>
If you don't need all the ram on the card - there's the w8100 (with<br>
fan on the card) - $1050 on Newegg! (Why do I feel like an<br>
infomercial)<br>
<br>
I have 2x w8100 in a server for benchmarking and they are nice GPU.<br>
Not as fast as a K40, but beats other cards.<br>
_______________________________________________<br>
Beowulf mailing list, <a href="mailto:Beowulf@beowulf.org" target="_blank">Beowulf@beowulf.org</a> sponsored by Penguin Computing<br>
To change your subscription (digest mode or unsubscribe) visit <a href="http://www.beowulf.org/mailman/listinfo/beowulf" target="_blank">http://www.beowulf.org/mailman/listinfo/beowulf</a><br>
</blockquote></div><br></div></div>