The smartphone market has come a long way as companies like Apple, Huawei and Snapdragon have developed chip sets with advanced capabilities allowing smartphones to perform new tasks and process it faster too. The main aim this time was to develop chips which can provide for machine learning and artificial intelligence on devices.
Apple just announced their latest chipset, the A11 Bionic, calling it ‘The most powerful and smartest chip ever in a smartphone’. This six core CPU consists of 4.3 billion transistors and the six cores can run simultaneously providing 70% more raw computing power. Apple also boasts a computational power of 600 billion operations per second.
Now apart from this they also claimed that the A11 provides better battery life compared to its predecessor and that models with the A11 can last up to two hours longer. The A11 Bionic chipset is also integrated with an Apple-designed GPU with a three-core design that delivers up to 30 percent faster graphics performance than the previous generation. But what truly makes the A11 Bionic better than most of its competition is the neural engine which Apple has integrated into it. This neural engine lets the SoC process tasks related to Artificial intelligence. The iPhone X uses this neural engine for features like the Face ID facial recognition, better image signal processing and for the animoji feature. Artificial intelligence demands a great deal of computational power and the A11 provides for it.
But Apple isn’t the only company which has developed such a chipset. Huawei announced their new Soc, the Kirin 970 before the A11 which is their best chipset till now. With the Kirin 970 Huawei entered the 10nm league competing with the A11 Bionic and Snapdragon’s 835. The Kirin 970 also houses a Neural Processing Unit (NPU) which will drive the 970’s mobile artificial intelligence platform enabling cloud based AI and on-device AI to run alongside each other, faster than ever before. Huawei hasn’t launched a virtual assistant of its own yet but have great plans for artificial intelligence. More will be revealed in October when the company will announce the Mate 10, the first phone to use the Kirin 970.
The Kirin 970’s dedicated Neural Processing Unit is basically a piece of hardware that is very good at running neural networks and when compared to the 970’s CPU, the NPU delivers up to 25 times the performance with 50 times greater efficiency, meaning it performs the same tasks faster and while consuming less power. Richard Yu, the CEO of Huawei shared the results of an image recognition test in his presentation during the launch of the Soc in which the Kirin 970 processes 2000 images per minute, which is about 20 times faster than how long it would take the CPU to handle the same workload.
Looking at the rest of the chip, it is being manufactured by TSMC using a 10nm process. It is an octa-core processor, with a 12-core GPU, dual-ISP and a high speed Cat 18. LTE modem. The CPU is similar to that of the Kirin 960, with four ARM Cortex-A73 cores and four ARM Cortex-A53 cores, but this time clocked at 2.4 GHz and 1.8 GHz respectively. The Kirin 970 is also the first commercial SoC to use the Mali-G72, the latest GPU from ARM. According to Huawei, its implementation of the G72 will make the Kirin 970 20 percent faster than the Kirin 960, but yet 50 percent more power efficient. ARM’s Mali G-72 GPU used in the 970 delivers a 25 percent improvement to energy efficiency and a 20 percent saving on performance density, when using the same processing node as a G71 design. The G-72 uses the same bifrost architecture used in the previous high end Soc G-71. In terms of performance, SoC designers could immediately put this 25 percent energy saving towards additional performance while sticking within previous power budgets.
Huawei decided to set the Kirin 970 as an open platform for mobile AI, meaning the chipset is open to developers thus allowing them to find new and innovative uses for the AI processing capabilities of the Soc.
But even then calling these chipsets the best becomes difficult with the advent of yet another chipset developed by Qualcomm earlier this year. Successor to the Snapdragon 821, which was the pinnacle of Android smartphone chipsets in 2016 and used in many of the top end smartphones of the time, the Snapdragon 835 offers significant differences over the 821. Physically, it’s much smaller, which is a result of the push into a 10nm (nanometre) production process from the 14nm process of the 821.
Qualcomm says that, on average, the 835 will use around 25% less power, which will lead to a direct increase to battery life. The Snapdragon 835 uses an updated version of Qualcomm’s Kryo CPU, known as the 280. The CPU has eight cores and is based on ARM’s ‘big-little’ design. The four “big” cores run at 2.45GHz for snappy performance, while the four “little” cores run at 1.9GHz for those background tasks that don’t require huge power. The Adreno 540 GPU used in the 835 is known to be 25% more powerful than its predecessor in terms of its 3D performance so games run much faster and smoother now. 4K UHD Premium and HDR10 video and Quickcharge 4 are also supported now. Quickcharge 4 can provide up to 25% faster charging along with enhanced efficiency too. The Snapdragon 835 now supports Electronic Image Stabilization 3.0 and uses enhanced algorithms for better camera performance.
Now the Snapdragon 835 was able to compete well with processors like the Kirin 960 and the A10 Fusion in terms of computing power and was also able to beat its competition in terms of graphics thanks to its Adreno GPU, but when compared to the next gen Kirin 970 and Apple’sA11 Bionic, it’s difficult to call the Snapdragon 835 the best chipset.
So it comes down to the Kirin 970 and the A11 Bionic where the A11 Bionic definitely seems to have an edge over the former and stands on top of the list of SoCs produced this year.