The iPhone XS and XR come with Apple’s new A12 Bionic chip. This has an upgraded neural engine over the A11 in the iPhone X and iPhone 8. While Apple only told us the difference in the number of transistors, it was difficult to see if this was put into practice efficiently. Would this really make a difference when it comes to real world implementations?
Sebastiaan Van den Branden at In The Pocket wrote about the dramatic speed of the iPhone XS. His company is working on a machine learning program to recognize body parts. So far it can identify faces, arms, and hands. In their experimentation with the technology, they found the iPhone was able to return a result in a fraction of the time of the iPhone 8.
Rate of detection on the iPhone 8. Speeds are between 21 and 23 ms.
Detection time on the iPhone XS: between 2.69 and 2.85 ms.
In The Pocket’s developers found the iPhone XS’s A12 Bionic chip was consistently nearly 10 times faster with machine learning apps than the iPhone 8. To users, this is the difference between image recognition that seems instantaneous and image recognition that seems to take a fraction of a second.
The iPhone 8 is fast. Under 30ms would still feel nearly instantaneous to most users. However, the iPhone XS is extremely fast. This is machine learning power in our pockets like we’ve never had before. I can’t wait to see the creative ways we put it to use.
Source: Sebastiaan Van den Branden, In the Pocket (Medium)