On Fri, Jul 29, 2016 at 2:31 AM, Kilian Cavalotti <[email protected]> wrote: > On Thu, Jul 28, 2016 at 9:06 AM, C Bergström <[email protected]> wrote: >> Not so fast! Are you talking half, single or double precision? > > All of them! :) > The S9150 is 2.53 TFLOPS FP64 (double precision), 5.07 TFLOPS FP32 > (simple precision) > The GP100 (which you're right is probably a better comparison) is 5.3 > TFLOPS FP64, 10.6 TFLOPS FP32 and 21.2 TFLOPS FP16. > > Of course, those are theoretical values and real-world application > benchmarks will probably yield different ratios, but still, it's about > twice the theoretical oopmh. :)
AGAIN you aren't being honest - *squinty eyes* The original quote of 2x was using a comparison against a GTX1080 - Quote some numbers for that. The GP100 isn't really fair because mere mortals don't have access to it.. I'd like to see a show of hands for people outside of a tiny circle who do.. AMD's s9300 is meant to compete against the Pascal class hardware, but they haven't released the single asic part yet. Lets see those numbers.. I guess you're trying to be funny, but don't cross the line into be a trolloomph.. :) _______________________________________________ Beowulf mailing list, [email protected] sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
