A15 Bionic Processor - Phamox Tech

Table of Contents

Apple A15 Bionic Processor

The A15 Bionic is another chipset from Apple. This chip is Apple’s fifth-generation Bionic System on a Chip (SoC) which means all of the components.

So, the CPU, GPU, memory, i/o chip, Thunderbolt controller, Apple T2 are all on the same dye or package like a sandwich.

The American consumer products tech giant designed the 64-bit ARM-based System on a Chip (SoC) to power its new iPhone 13 model.

And just like other chipsets from the company, A15 Bionic was also built on TSMC’s 5nm process.

The Apple A14 chip was manufactured by TSMC on their first-generation 5nm fabrication process, N5. This makes the A 14 the first commercially available product to be manufactured on a 5nm process node.

The transistor count has increased to 11.8 billion, a 38.8% increase from the A13’s transistor count of 8.5 billion.

However, the new A15 Bionic processor by Apple utilized up to 15 billion transistors.

The number of transistors increased in A15 and a higher transistor count primarily means a newer, and more effective processor.

Also, the A15 chip has new features like an ISP (Image Signal Processor), video encoder, video decoder, display engine, and security module. These features were integrated into the chip to deliver specific needs of the Apples’ iPhones efficiently.

At the end of the day, Apple claimed that the A15 Bionic Processor was the fastest smartphone chipset available on the globe today.

Apple A15 Bionic Design

The A15 Bionic is a chip designed to optimize every aspect of the 2021 iPhone 13 series and guarantee a superior user experience.

Apple actually designed the A15 chipset with a focus on efficiency that leads to high performance. As a result, the company has taken the lead in the industry competition.

The A15 is a leap forward for Apple in terms of pure performance and majorly in EFFICIENCY!

The leap forward and bounds are not what we have seen in years when Apple moved to fully custom cores. Or when it added cores or benefited from process shrinks.

Actually, Apple has been remarkably consistent with the basic Bionic architecture design over the last few years.

Apple moved from Ice Storm high-efficiency cores and Firestorm high-performance cores on A14 Bionic to Avalanche and Blizzard.

Basically, the A15 chip is a 6-core CPU with two high-performance cores and four high-efficiency cores.

Also, the new A15 chipset has a 4/5-core GPU, 16-core Neural Engines cores, and much more specific silicon features.

A15 design was indeed a great innovation by Apple because it makes the entire iPhone 13 system deliver more and high performance in power efficiency.

A15 Bionic Components And Features

The A15 Bionic chipset comes with a bunch of much more specific silicon features.

The features were put in place to support the much more specific experiences that Apple wants to deliver with each and every new iPhone.

Apart from the E-core and P-core, the A15 Bionic features an ARMv8 Instruction Set.

Also, it features 16-Core Neural Engine, Machine Learning Controller, Secure Enclave, and Advanced Image Signal Processor.

In addition to core power improvements, there is a new display engine, new video encoder and decoder, and twice the system cache.

The 6-Core CPU

The new A15 Bionic chip comes with a 64-bit 6-core CPU with ARMv8 engineering. The 6-core CPU is divided into two major performance parts. The two parts are 2 High-Performance cores (or P-cores) and 4 High-Efficiency cores (or E-cores).

Specifically, the Blizzard core is the High-Efficiency core on A15 Bionic while the Avalanche core is the High-Performance core.

The anadtech report about the new cores (both P-cores and E-cores) performance around the cores frequencies is as shown below:

Just looking at the cores, the highest CPU frequency per core is 3,240MHZ in the A15 chip. This is greater than the 2,998MHZ maximum CPU frequency per core in the A14.

The two-performance core cluster CPU registered an 8% increase in single-core performance. But when both performance cores were active, there is a 10% average improvement in single-core performance.

Apple A15 4/5-core GPU

Apple normally customized its GPU just to increase performance and power efficiency. It does this especially for new features like the depth effect, and computational photography.

The first Apple’s Bionic chipset, A11 came with a custom 3-core GPU. But the A12 Bionic chip took Apple’s custom GPU to 4-cores with “Apple G11P”.

The subsequent Bionic A13 and A14 kept that exact 4-core GPU configuration.

However, Apple kept the 4-core GPU configuration for this year’s A15 in iPhone 13 and 13 Mini by disabling the GPU 5th-Core.

And for iPhone 13 Pro and 13 Pro Max along with iPad Mini 6, Apple went for a full 5-core GPU configuration.

The Apple 4/5-core GPU on A15 runs graphics-intensive tasks like advanced Games smoothly.

The 4/5-core GPU enables more eye-catching visuals and lighting effects. And Apple claimed that it delivers up to 30% Faster graphics versus its leading competition.

Moreover, A15 Bionic GPU comes with Lossy Texture Compression that allows half of the memory for the same resolution texture. In other words, Lossy Texture Compression means to double the resolution for the same memory size.

The new A15 Bionic Lossy Compression saves the memory up to 50% with negligible loss in visual fidelity. Besides, the new Lossy Texture Compression helps to save bandwidth.

Apple also introduced support for Sparse Depth and Stencil Textures and a new SIMD shuffle and fill instructions.

The 32MB A15 Bionic System Cache

Apple increased the size of the system cache from 16MB in A13 and A14 to 32MB in the A15.

Other SOC vendors also used these types of last-level cache approach like 3MB SLC on the Snapdragon 888 with 3MB SLC or Exynos 2100 with an estimated 6-8MB SLC.

However, the new 32MB development on Apple’s A15 is enormous compared to other implementations.

This has helped to improve the speed even more. Furthermore, latency testing shows that the system cache has been expanded and improved the chip’s power efficiency.

Because it allows memory access to remain on the same silicon for an extended period of time rather than looking for DRAM.

However, Apple does not mention changes to the L2 cache of the performance core which was raised by 50%.

The L2 cache is from 8MB to 12MB and of course the same size as the version found in the M1. Moreover, it does not serve four performance cores and now serves two instead.

The access latency also increases from 16 cycles to 18 cycles from A14 to A15 respectively.

Moreover, the A15 is capable of 1 cycle access to the cache line in the L1 cache which A14 has to go through 3 cycles.

On the efficiency cores side, Apple maintained the cache sizes at the same 64KB L1D’s along with 4MB shared L2’s.

But Apple upgraded the L2 Translation Lookaside Buffer (TLB) to 2048 entries that cover up to 32MB.

Again, Apple-designed efficiency cores in A15 Bionic to access DRAM faster than what it has on A14.

The DRAM access latencies on A15 are approximately 130ns now compared to +215ns on the A14 chip.

The 16-Core Neural Engine

In the early days of hardware-accelerated Machine Learning, Apple depend completely on the GPU.

However, they want all the processes in a faster and more efficient way than what the general-purpose GPU could ever do.

So for this reason, Apple introduced the 2-core Neural Engine in 2017 with the A11 Bionic.

The Neural Engine was designed to handle the massive processing required for all the new features on the iPhone better.

Specifically, to process all the new algorithms and adversarial neural networks behind the features like Face ID on the iPhone 10 effectively.

The 2-core Neural Engine in A11 Bionic can perform 600 billion operations per second.

Afterward, Apple moved to 8-core Neural Engine in 2018 with A12 Bionic that performs 5 trillion operations per second.

A13 Bionic came out in 2019 with (Machine Learning Accelerators) AMX blocks added to the CPU.

Also, Apple added a machine learning controller that can dispatch tasks to the AMX, the GPU, and the ANE (Neural Engine) in real-time.

That seems to be a game-changer and Apple’s secret sauce.

Later in 2020, Apple increased the Neural Engine count from 8 to 16 to increase efficiency.

So they have a 16-core Neural Engine with the A14 Bionic chip that can perform 11 trillion operations per second.

Now with the new A15 Bionic chip, Apple maintains a 16-core Neural Engine and increased operations per second by over 40%.

So, A15 Bionic is capable of performing an amazing 5.8 trillion operations per second.

New A15 Neural Engine Functionality

The Neural Engine (ANE)is specialized hardware built for a set of machine learning algorithms.

The ANE hardware is the cornerstone for the earlier Bionics Chipset Machine Learning processes and remains the same on A15.

Apple’s Silicon team worked together with the Apple hardware and software teams to bring the Neural Engine to life.

And with the 5.8 trillion operations per second, it means iPhone 13 series are capable of faster Machine Learning computation experiences.

Apple is leaning on its 16-core Neural Engine for so many of the A15 and iPhone 13 features like Siri and live texting camera to perform better.

Moreover, the efficiency and performance increment of the Neural Engine brings the leading capabilities in iPhone’s video and image processing.

The Neural Engine on A15 powers Apple’s new features like Cinematic Mode and Photographic Styles in iPhone 13 series.

ISP (Image Signal Processor)

Inside the A15 Bionic is an Apple-designed Image Signal Processor (ISP) that plays a big role in enhanced computational photography capabilities.

Also, the new Apple’s ISP also brings huge improvements to camera quality advancements.

Apple started adding ISP from the A5 chipset for things like auto white balance, focus, face detection, and all the image processing basics.

The ISP in the new A15 Bionic comes with a Smart HHDR 4 feature that gives it the capability to semantic rendering for multiple people.

That means it can render the same shot to several people at the same time separately and individually.

The new iPhone 13 series also uses a new NVME based storage controlling system to record and save your capturing without slowing down your iPhone process.

Apple’s also been using custom encode and decode blocks for H.264, then H.265, aka HEVC then VP9 for YouTube.

With the A15 and its ISP, Smart HDR 4 is able to render multiple people in

and save your capturing without slowing down your iPhone process.

Final Note

Specifically, Apple uses the A15 Bionic chipset on its entire iPhone 13 series system to deliver more high performance and power efficiency.

And with different tests, the A15 Bionic chip has outperformed all its rivals in terms of performance.

In particular, tests show that the A15 Bionic outperforms the A14 bionic. And this will be the most powerful and quickest CPU that the company has ever built.

Apple has gained several advantages over its competitor by designing its own chip by itself. With this custom chip design approach, Apple has priorities in its products’ features and release timing.

Also, Apple now has full control to integrate its own software.

We trust that you found the information on “A15 Bionic” educative and informative. Please do not hesitate to share it with anyone who might also benefit from this article.