The Apple iPhone X' all new A11 Bionic chip smokes Android's best: A closer look reveals why

Apple’s **new iPhones** will be powered by an all new chip, the A11 Bionic. As usual, the chip smokes the competition and yet again raises the bar as far as performance is concerned. [caption id=“attachment_4039221” align=“alignleft” width=“380”] Apple A11 Bionic[/caption] The A11 Bionic chip is the first hexa-core chip from Apple, though in some ways, one might even consider it a 10-core chip. Details are scant, but the little that we’ve gathered from Apple’s announcements and leaked information tells us that the chip features a 6-core CPU, a 3-core GPU and an M11 motion co-processor. The 6-core CPU comprises of 2x high-performance cores and 4x energy-efficiency cores. This is two more than on the previous generation. The GPU, which handles all the graphics processing tasks, is the first graphics chip that Apple has designed. Previously, Apple used GPUs from Imagination Technologies. When news broke that Apple would no longer use Imagination Technologies’ chips, the company’s stocks fell by 70 percent and **it’s now up for sale** . The information we’ve found so far suggests that the A11 Bionic uses 64-bit ARM cores that use the ARM v8-A instruction set. These are said to be TSMC-built 10 nm chips. The previous A10 chip was built on a 16 nm process, meaning that the A11 is far more efficient. The 10 nm process also means that the chip can be clocked higher, resulting in higher performance. [caption id=“attachment_4039229” align=“alignnone” width=“1280”] The Metal 2 and Core ML APIs will enable powerful AR experiences[/caption] The idea of using high-performance cores in conjunction with low-performance is not new. In fact, it’s the cornerstone of mobile chipset design today. ARM popularised it as the big.LITTLE architecture, where high-performance cores kick in for demanding tasks, leaving routine and less demanding tasks to the energy-efficiency cores. This results in tremendous power savings. Apple appears to have changed things up a bit, however. The company is apparently using a second-generation controller that permits the usage of all six cores simultaneously, resulting in even higher performance. Judging by the architecture of the A10 chips from the **iPhone 7 Plus** , this is a design that makes sense for Apple as the cores can share memory directly with each other, reducing latency and improving performance. According to Apple, the CPU is 25 percent faster than the A10 and the GPU is 30 percent faster. Apple also claims that the performance cores are 25 percent faster than their counterparts on the A10, and that the energy-efficiency cores are 70 percent faster than their counterparts. [caption id=“attachment_4039235” align=“alignnone” width=“1280”] The A11 Bionic’s Neural Engine can perform 600 billion operations per second[/caption] The good stuff doesn’t stop there. Two new features were introduced with iPhone X: **Portrait Lighting** and **Face ID** . We’ve written extensively about both these features **here** and **here** . Suffice it to say that these tasks are computationally intensive. A traditional CPU will struggle to process that amount of data at speed. To streamline this process, Apple has updated the Image Signal Processor (ISP) on the chip with support for “lighting estimation, wide colour aperture and advanced pixel processing”. For Face ID, a “Neural Engine” capable of 600 billion operations per second has been added. This chip, claims Apple, incorporates several machine learning algorithms that speed up the process of biometric authentication via Face ID. As mentioned in our Face ID explainer, all Face ID authentication and processing happens on the device. The graphics units also supports Apple’s Metal 2 graphics API and Core ML, a machine learning API that can enable AR (augmented reality) experiences and support native ML functions in apps. [caption id=“attachment_4039247” align=“alignnone” width=“1142”] An iPhone X Face ID-enabled animoji[/caption] In total, the A11 Bionic chip is made up of over 4.3 billion transistors. For comparison, the previous generation A10 chip housed around 3.3 billion transistors and one of Android’s best processing platforms, the Qualcomm Snapdragon 835, is home to 3 billion. We’re not electronics engineers so we can’t comment on whether the maxim, “bigger is better” applies here, but initial benchmark figures certainly reaffirm that notion. Looking at Geekbench 4 scores, the A11 Bionic simply thrashes the competition from every competing mobile device on the market except Apple’s own 10.5-inch iPad Pro. Android-powered devices like the Samsung Galaxy S8+ aren’t even comparable. This helpful chart we’ve compiled from Geekbench 4 data (actual data, not leaks and rumours) should give you a fair idea of what to expect from the new iPhones.

Apple A11 Bionic vs the world: Geekbench 4 scores Infogram