A15 benchmarks show a 21% CPU performance gain over A14 but there’s a caveat

The fact that Apple failed to mention the actual performance gain of the new A15 Bionic chip over its predecessor during the California streaming event is causing quite a stir in the internet space. Earlier today, a report by Semianalysis stated that there were no major improvements made to the CPU on the new chip.

According to the report, the reasons behind the underwhelming CPU performance of the A15 boil down to the human factor. Apple lost a number of important employees recently including the man who designed the company's A-series chips from the A7 through A12X, Gerard Williams III.

And then Hell broke loose. Geekbench scores started to leak left and right. A Geekbench Metal score (measuring the performance of the GPU) has leaked, showing a result of 14216. The alleged device (iPhone 14.2 AKA iPhone 13 Pro) scored 55% better than the 9123 tallied by the iPhone 12 and the A14 Bionic SoC.

Another report from AppleInsider painted a different picture of the CPU performance. The tech site looked at CPU statistics uploaded by Geekbench users and found out that iPhone 13 managed an average single-core score of 1730, and an average multi-core score of 4621, roughly 10% and 21% increase respectively compared to the iPhone 12's score in the same tests.

This result has to be taken with a grain of salt, though. Users have posted comments showing yet another Geekbench test score of the iPhone 13, which supposedly offers a more realistic real-world result and a better comparison with the A14 Bionic.

iPhone 12 Pro Max - iOS 14.71 - Geekbench 5.4.1 - Sept. 15th 2021

A14 @ 2.99 GHz
Singe Core Score: 1605
Multi Core Score: 4294

iPhone 14,2 - iOS 15 - Geekbench 5.4.1 - Sept. 15th 2021

A15 @ 3.22 GHz
Singe Core Score: 1728 (+7.6% over A14)
Multi Core Score: 4695 (+9.3% over A14)

If we look at the scores above, it’s clear that the A15 Bionic has a higher clock rate than its predecessor. The much more humble increase in performance (8-9%) could be attributed to the higher frequency of the new chip.

Our take

It’s really had to judge the performance of a chip that’s yet to be officially released. Yes, we understand that it’s strange, to say the least, for Apple to omit the CPU performance gain of the A15 over A14.

The company never misses an opportunity to show how much better its new products are compared to the previous generation. That being said, Geekbench scores are a bit unreliable at this point in time.

For all we know, those scores might have come from early samples or they could be plain fakes. The only way to get the real-world Geekbench scores is to wait for the production models to hit the stores and the reviewers subsequently.

Does CPU performance matter all that much?

And then there’s the real-world advantage of a 10% synthetic benchmark score increase. We’re talking about the fastest mobile chipset anyway - the A14 Bionic still tops benchmark charts, and no one would call the iPhone 12 slow.

Furthermore, the way Apple optimizes its ecosystem and apps, in general, helps tremendously with the iPhone’s real-world performance. Sometimes people are obsessed with numbers, and “only” 10% increase seems to be the end of the world.

And to be honest, the CPU and GPU computational power in modern smartphones is way higher than most use cases call for. So, at the end of the day, we suggest you don’t pay too much attention to synthetic benchmark scores (leaked) and wait for tech reviews.