Early Geekbench benchmarks of the iPhone 5s seemingly prove Apple’s claim that the new 64-bit A7 chip provides two times greater CPU and GPU performance over its predecessor, the A6 package inside the iPhone 5. Specifically, the iPhone 5s Geekbench score (the new version of the Geekbench app supports 64-bit chips) is close to the early-2010 Mac mini model.
But how do benchmark scores translate into real-world performance in your favorite apps? Blogger Lex McFarley did an interesting head to head test using the iMovie app running on both an iPhone 5s and a previous-generation iPhone 5. Perhaps unsurprisingly, the new phone rendered the video in half the time as the iPhone 5…
As noted by Macsfuture, the A7 chip has a Geekbench score of 2552, which is “is close to the score of the early-2010 version of the Mac mini”.
Here’s the iMovie test.
The entire project runs ten minutes 55 seconds long.
It took the iPhone 5s about six minutes to render the project in full HD resolution – that’s almost two times faster than real-time rendering, which includes the time needed to copy the resulting movie back to the Camera roll.
Apple says the A7 supports OpenGL ES version 3.0 to deliver the kind of detailed graphics and complex visual effects “once possible only on Mac computers, PCs, and gaming consoles”.
The difference is amazing. Take the imaginary worlds in games, for instance. Textures and shadows look more true to life. Sunlight reflects off the water. The whole experience feels much more realistic.
A more in-depth speed comparison of the two phones follows below.
Created by Tom Rich, the following video not only covers the general speed, but gaming performance of the 64-bit and 32-bit processors and much more.
Apple now has the first 64-bit smartphone on the market running the first 64-bit operating system with all stock apps optimized for 64-bit computing.
It should be noted that while navigating the handset and launching apps appears smoother on the iPhone 5s, these common tasks don’t appear any faster versus the iPhone 5 due to timed UI animations in iOS 7.
Had Apple removed these transitions altogether, the speed increase would have been more noticeable. Of course, the difference in speed become readily apparent when running graphics-intensive apps and high-end games optimized for the A7 and the 64-bit platform, such as Infinity Blade III (pictured top of post).
Here’s Infinity Blade III demo from the keynote.
http://www.youtube.com/watch?v=Wnh98ELfR04
By all accounts, Apple has likely licensed the ARMv8 architecture for the CPU.
There’s little doubt left that the A7 runs a 100 percent customized CPU core, alongside what’s probably Imagination Technologies’ PowerVR Series 6 GPU code-named Rogue.
Apple holds a ten percent stake in Imagination.
The iPhone 5’s A6 chip was a break from Apple’s previous strategy of incorporating ARM’s CPU blueprints into its A-series chips on an as-is basis.
@dujkan Completely plausible. Likely a big jump in polys even if they go with a lower-end Rogue.
— Jeremy Horwitz (@horwitz) September 6, 2013
Instead of using either the ARM Cortex A9 or ARM A15 CPU core for the A6 chip, Apple had licensed the ARM v7 instruction set and produced a fully customized dual-core CPU core named Swift.
As per usual, the industry is expected to close the gap some time next year, giving Apple plenty of lead time to bank on its “the fastest smartphone” claim. Right after the keynote, Samsung co-CEO Kim Yoo-chul said future Samsung phones will have 64-bit processing functionality, although “not in the shortest time”.