If you look at GB3's integer sub-tests per-core, you'll see that the A7 and Z3770 (not Merrifield, btw, which is clocked at only 89% of BYT) trade-blows in the non-Crypto benchmarks. In the memory and the FPU benchmarks, there's absolutely no contest: A7 does much better by a fairly huge margin, even if you look at the ICC v.s. GCC/MSVC++ numbers.
So, I guess to you FPU performance is completely irrelevant on a phone/tablet.
Thank you for your opinion :-)
As soon as Intel apologizes for: - Poorly forecasting PC demand during 2013 - Guiding down in 2012 multiple times - The "contra-revenue" surprise after telling us Bay Trail would be very profitable in tablets - DCG growth would be 9% during 2013 and enterprise growth at ~8% in 2014 only to miss/guide down about 2 months later - Forecasting $5.5B in revenues in Other IA during 2013 (up from $4B) only to actually decline during 2013 - XMM 7160 being extremely late - The cancellation of '6331' - etc.
Then I'll "apologize" for pointing out that Silvermont seems to not do very well in GB3 against A7. I've made many mistakes and I've openly admitted to failure/defeat and apologized...but this doesn't look like one of those times - at least not yet.
That being said, you seem to ignore all of the good I've done/written for/about Intel over the last two years. I've countered countless articles claiming absurd things about X86/Intel, and yet an article simply presenting the results of a well-known, relevant benchmark is "irredeemable".