Home > News content

Self-developed Da Vinci architecture NPU support! A picture of Huawei Kirin 810

via:博客园     time:2019/6/21 17:18:17     readed:259

At Huawei's nova5 press conference this afternoon, he Gang, president of Huawei's terminal mobile phone line, unveiled a new artificial smartphone chip.

According to the official introduction,Kirin 810 is manufactured by 7-nm process. Compared with 8-nm process, energy efficiency and transistor density are increased by 20% and 50% respectively. At the same time, Huawei's self-developed Da Vinci NPU was adopted for the first time, and AI runs up to 32,280 points.

自研达芬奇架构

For CPU, Kirin 810 adopts a new system-level AI FM scheduling technology.2 6 size nuclear architecture with two large cores (2.27GHz) based on Cortex-A76 developers and six Cortex-A55 small cores (1.88GHz)GPU upgrade to Mali-G52 customization to support Kirin Gaming technology.

In photography, Kirin 810 integrated detail enhancement (DE) module.Support the latest generation of automatic white balance algorithm (AWB) and AR feature point cloud computing acceleration, ISP performance and algorithm double improvement.

In addition,Kirin 810 supports dual-card dual-VoLTE, realizes stable and extremely fast mobile communication connection under various complex communication scenarios. At the same time, Kirin 810 introduced a self-developed intermediate operator format, which greatly enhanced the compatibility of Huawei HiAI and accelerated the landing of more AI applications.

The following is an official summary of Kirin 810.

自研达芬奇架构

<#comment><#comment>

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments