Revision as of 20:13, 12 September 2017

Template:mpu A11 Bionic is a 64-bit hexa-core ARM performance mobile system on a chip introduced by Apple in late 2017 as a successor to the A10.

Overview

Introduced during a keynote event on September 12 2017, the A11 Bionic features six 64-bit CPU cores. Fabricated on TSMC's 10 nm process, the A11 consists of two high-performance big cores which are said to be 25% faster than the A10 and four high-efficiency little cores said to be 70% faster than A10. The chip features Apple's 2nd generation "performance controller" that allows the system to use all six cores at once and is said to offer up to 70% performance improvement in multi-threaded workloads.

In prior models, Apple was using PowerVR GPUs designed by Imagination Technologies. The A11 is the first chip to make use of Apple's in-house designed GPU. The GPU is a tri-core design which is said to deliver 30% higher performance while using half the power as the A10. The GPU has been optimized for 3D visuals specifically for gaming apps and is optimized around Apple's own Metal 2 hardware-accelerated graphics API.

The A11 also includes a new neural engine processor for accelerating image recognition, an ISP, and a Secure Enclave Processor (SEP).

ISP

The A11 incorporates Apple's image signal processor (ISP) for camera related applications. The ISP is said to deliver faster auto-focus in low-light scenarios and incorporates new pixel processing for sharpness and textures. The A11 ISP also incorporates new hardware for multiband compression/expansion noise reduction capabilities.

Integrated Graphics

The A11 incorporates an Apple-designed GPU. The GPU is a three-core design which is said to deliver 30% higher performance while using half the power as the A10. As with the previous microprocessor, the A11 incorporates a video encoder designed by Apple. Among other things, the encoder introduces two major features: faster video frame rates and real-time image analysis.

Apple stated that the encoder is now capable of real-time image and motion analysis. This is done by divided the screen into individual tiles (on the iPhone 8 this works out to 2 million tiles) which are analyzed each second for things such as textures and edges in order to predict whether the tile is displaying things such as grass, water, sky or movement. Based on those properties the encoder can optimize the video encoding algorithms for quality and compression level.

The A11 supports faster video frame rates in order to improve slow motion videos. The A11 supports up to 240 FPS at 1080p resolution which is double the previous frame rate.

Neural Engine

The A11 chip incorporates a new neural engine which implements a set of machine learning algorithms designed to accelerate image recognition. The engine, a dual-core design, capable of performing real-time processing at 600 billion operations per second. This is the second mobile processor on the market to feature such processing unit after HiSilicon's Kirin 970.

Utilizing devices

• iPhone 8
• iPhone 8 Plus
• iPhone X

This list is incomplete; you can help by expanding it.