“When Apple announced that its iPhone 5s featured a 64-bit ‘desktop class’ architecture, the press focused intensely on the ’64-bit’ aspect of the chip design,” Ashraf Eassa writes for The Motley Fool. “While the move to ARM’s new, clean 64-bit instruction set certainly helped, the real ‘magic’ was actually in some of the extremely smart design choices that Apple made when architecting the Cyclone CPU found inside of the A7. Thanks to a deep investigation by Anandtech, we now have even more evidence that Apple blew away the competition.”
“In the vast majority of benchmarks, Apple’s 1.3 GHz Cyclone core offered performance that was roughly in line with a 2.4 GHz Intel Silvermont and meaningfully higher than the various ARM Cortex A15 and Qualcomm Krait-based processors per processor core,” Eassa writes. “Of course, Qualcomm and Intel are able to put four of these cores in roughly the same thermal envelope that Apple was able to put two of its cores, so the design trade-off was really ‘more cores’ versus ‘more performance per core.'”
“Given that the vast majority of mobile applications likely can’t make use of more than one core, let alone four cores, the design decision made absolute perfect sense. Indeed, it likely delivers the best user experience,” Eassa writes. “Additionally, while the Android handset vendors need to differentiate among themselves with whiz-bang marketing features, Apple need not get involved in that sort of thing, although the “64-bit” marketing point didn’t hurt. It can make the soundest technical decisions free from marketing constraints.”
Read more in the full article here.
Apple’s 64-bit Cyclone microarchitecture detailed – March 31, 2014