The Kirin 980 — Huawei’s First 7 nm Processor SoC
The power behind Huawei’s flagship smartphones is the Kirin 980 processor. It is also its first commercial AI chip that uses the 7 nm fabrication process. The Kirin is a 64-bit high performance mobile ARM SoC processor with 4 big Cortex-A76 cores that operates at up to 2.6 GHz along with 4 little Cortex-A55 cores operating at up to 1.8 GHz. Other features of this chip include an LTE modem with 1.4 Gbps support, Mali-G76 GPU and LPDDR4X-4266 memory support.
Microarchitecture Specs
ARMv8
The ISA (Instruction Set Architecture) uses the ARMv8 instruction sets. The ARMv8 introduces 64-bit capabilities. It is the 64-bit variant of the ARMv8 that is used in Huawei’s high end P30 Pro. A cryptography acceleration engine has been added to the instruction set for increased security. This provides support for 2 types of algorithms, AES and SHA (SHA-1 and SHA-256).
MALI-G76
The Mali-G76 GPU provides support for Direct X12, Open CL1.2, OpenGL ES 3.2, OpenVG1.1 and Vulkan 1.0. This is a 3rd generation Bifrost architecture based chip that is for complex graphics and ML (Machine Learning) computation. It contains 10 execution units at 750 MHz.
Wireless
The following wireless features are supported:
- LTE Modem
- DL: Up to User Equipment (UE) Category 21 (Cat.21)
- Downlink up to 1.4 Gbps (4x4 MIMO + 256QAM 3CC CA = 1.2 Gbps, 2x2 MIMO + 256QAM + 1CC = 200 Mbps)
- UL: Up to User Equipment (UE) category 18
- Uplink of up to 200 Mbps (2x2 MIMO, 256-QAM, 1x20MHz CA)
- Wi-Fi 802.11 ac
- Bluetooth 5
- NFC
- GPS / A-GPS / GLONASS / BDS
Memory
The LPDDR4X-4266 does not support ECC. The maximum memory is 8 GiB. There is 1 controller that use 4 channels with with a width of 16 bits. The maximum bandwidth it can support is 31.78 GiB/s (Double 15.89 GiB/s Quad 31.78 GiB/s).
Synopsis
As an AI chip, it will be competing with the likes of Apple’s A12 and A13 Bionic, Qualcomm’s Snapdragon 855 and 865. A revelation of the Kirin 980 is its octa-core cache configuration that splits the processing between big and little cores. Huawei also claims it is the the world’s first dual NPU design and the world’s first chipset to support LTE Cat.21.
The chip was manufactured at TSMC using a 7 nm process like that used on Apple’s A12 Bionic. An NPU is going to be important in upcoming chip designs due to the advanced applications that are now in demand. Things like face recognition, object recognition, object detection, image segmentation and intelligent translation will require neural processors. Apps that push the boundaries of AI to new levels is sure to require more processing units. It has a 20% improved performance from the 10 nm process used in previous chips.
It is the efficiency claims that Huawei makes that is another significant feature of the Kirin 980 chipset. The power efficiency is up by 40% and the overall energy efficiency of the Kirin 980 increased by 58% (based on testing). The efficiency is accomplished by the core architecture design which uses 4 small efficiency cores with 2 super cores and 2 large cores. This means less power used does not drain the battery as much. Longer battery life allows users to have more power to their devices.
I did not see support for 5G on the website or from other sources. Future versions will most likely have it. The connectivity features for networking on this device supports most all industry standards. Its most innovative feature is the support for Cat.21 that allows for 1.4 Gbps maximum downloads and 200 Mbps uploads.
The main criticism for the Kirin 980 is the graphics performance. This is when it comes to gaming apps on the devices. These were based on benchmarks carried out on the Mali-G76. Despite that, many of Huawei’s claims are legit but real world use of devices will tell a better picture. The Kirin 980 will not likely beat the Snapdragon 865 or A13, so it is ideal for the lower end of the top tier flagship market.
