However much weight you put into benchmarks, one thing we know is that devices of a certain model should compare well amongst themselves. It’s what we call consistency. But once you start pitting chips on mobile phones to chips on laptops, you’re not really getting meaningful results.
iPhone 8, iPhone 8 Plus and iPhone X variants have been getting tested left and right on Geekbench 4. All of them have the Apple A11 chipset manufactured by TSMC and all of them perform amazingly on some level. Single-core scores settle in the low 4,000s while multi-core scores are generally in the low 10,000s. This is unprecedented for the mobile side of things — the multi-core tests beat out the mid-9,000s that a dual-core Kaby Lake Intel Core i5 can pull off. You know, the 13-inch MacBook Pro introduced last year and upgraded this year.
Analysis from Daniel Matte’s Tech Specs blog gives a lot of credit for the score juicing to the transition from 32- to 64-bit architecture.
“I know many people will raise an eyebrow at this, but I would encourage you to completely ignore Geekbench,” Matte said. “And I can’t emphasize enough how almost everything written online about Apple’s CPUs is wrong.”
In the real world, instructions per cycle on the A11’s two big cores may amount to 15 percent over the A10’s. The 10nm FinFET process that has been raising hell this past year haven’t been producing massively improved and sustained clock reads.