The HiSilicon Kirin 980 system-on-chip that’s likely to power Huawei’s upcoming Mate 20 lineup of Android flagships was detailed in a new leak out of China, with sources close to the semiconductor manufacturer claiming the module will feature four Cortex-A55 cores for energy-efficient performance and another quad-core cluster using Cortex-A77 cores meant to handle all computational heavy lifting. The latter configuration is said to be clocked at up to 2.8GHz, with a custom, in-house GPU supposedly also being part of the package. The graphics chip in question will reportedly be up to 50-percent faster than the Adreno 630 GPU used by Qualcomm’s ultra-premium Snapdragon 845 SoC, though its exact technical specifications remain unclear.
The Kirin 980 is said to be produced on TSMC’s 7nm FinFET process, thus being the world’s first chipset to utilize that manufacturing technology. Much like last year’s Kirin 970 debuted with a dedicated neural processing unit specifically designed for handling machine learning and general artificial intelligence applications, the Kirin 980 is likely to feature another such module, with the new report indicating Huawei’s new NPU will use solutions created by a company called Cambricorn which will allow it to perform a maximum of five trillion calculations per a single watt of power.
As the Kirin 970 was announced at the 2017 iteration of IFA, its successor is also likely to debut at this year’s edition of the Berlin, Germany-based trade show. Rumors about the chip have been circulating the industry for months now and while the Huawei Mate 20 series should be the first product family to feature the new silicon in the final quarter of the year, the same module should also be found inside the company’s P30 flagship line expected to launch in late winter or early spring. As Huawei managed to get its last high-end chip to drive a car, the new hardware is also expected to deliver state-of-the-art capabilities pushing the boundaries of contemporary mobile technologies.