I would not draw too many conclusions, the SpecAcc is just telling you the quality of the OpenACC compiler and the quality of the porting. For example, if you look at the results for CloverLeaf ( I am familiar with this application and have other reference points), you have: AMD/Pathscale: 3.13 specaccel_peak NVIDIA/PGI: 3.45 specaccel_peak
Keeping the HW constant and changing the software ( adding CUDA C and CUDA Fortran to the mix) will give you for the 3840x3840 grid the following average times per cell (measured in 10^-8s): OpenACC loops: 1.92 OpenACC kernels: 1.78 CUDA Fortran; 1.33 CUDA C: 1.25 Timing is on a K20c, but we are interested in the relative performance. Cuda C/Fortran in 30% faster. There is also an OpenCL implementation of CloverLeaf but I don't have the results. It is probably in the same ballpark. This is a "simple" CFD code with regular access pattern, a directive base porting gives you decent results. You could try to run the OpenCL code on the AMD card and see how far the Pathscale compiler is from it, but I am expecting something similar. OpenACC is an interesting option for people looking for high level programming, but you usually pay a penalty. How big is the penalty will depend on a lot of factors and it is very difficult to generalize. M On Wed, Mar 4, 2015 at 12:26 PM, C Bergström <cbergst...@pathscale.com> wrote: > On Thu, Mar 5, 2015 at 3:10 AM, Craig Tierney - NOAA Affiliate > <craig.tier...@noaa.gov> wrote: > > > > 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. > > SPEC does allow you 1:1 comparisons. In this case we're not yet > showing the gains I know we can achieve. I'm mostly trying to stir the > pot to see the level of interest. > > Here's NVIDIA's best published result > http://spec.org/accel/results/res2014q1/accel-20140303-00018.html > compared to ours > http://spec.org/accel/results/res2015q1/accel-20150218-00045.html > > The specific Intel CPU is less a factor if you're concerned about > that. I could put this card in the exact same system NVIDIA used and > show some decent performance. (That 3.8Ghz boost in fact may help more > than anything) > _______________________________________________ > Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing > To change your subscription (digest mode or unsubscribe) visit > http://www.beowulf.org/mailman/listinfo/beowulf >
_______________________________________________ Beowulf mailing list, Beowulf@beowulf.org sponsored by Penguin Computing To change your subscription (digest mode or unsubscribe) visit http://www.beowulf.org/mailman/listinfo/beowulf
