Self developed chip V2 made its debut, and the performance of vivo X90 series was comprehensively improved

Vivo series mobile phones can be said to be the industry model this year. All kinds of new products, not only mobile phones, but also systems, images, and chips. The X80 series mobile phones at the beginning of the year were widely praised, and the X90 series launched at the end of the year was also highly anticipated. Recently, a blogger reported that the performance of the vivo X90 series has been comprehensively improved, and the self-developed chip V2 has made its debut, and will conduct in-depth cooperation with ZEISS. The images and performance will be greatly improved!

Self developed chip V2 architecture iteration brings extreme computing power and energy efficiency ratio

Since 2021, vivo has pioneered and adhered to the path of self-developed chip technology, constantly polishing the underlying core competence that no one has and continues to iterate. This effort has also been recognized by the market through the V1 and V1+two generations of self-developed chips.

This time, the self-developed chip V2 opened a new chapter in mobile imaging with a new iterative AI-ISP architecture.

The self-developed chip V2 has brought about comprehensive improvement in compatibility and functionality, and greatly upgraded the on-chip memory unit, AI computing unit, and image processing unit. Put forward FIT dual core interconnection technology, establish a new high-speed communication mechanism between self-developed chip V2 and Tianji 9200 flagship platform, enable two chips with completely different architectures and instruction sets to complete dual core interconnection synchronization in 1/100 second, and achieve optimal coordination and high-speed collaboration of data and computing.

Thanks to the near memory DLA (vivo self-developed AI deep learning accelerator) module and the large capacity dedicated on-chip SRAM (high-speed and low-power cache unit), the self-developed chip V2 rematched the computing power capacity, computing power density and data density, greatly improving the capacity and computing speed of the on-chip cache. Compared with the DDR external memory design commonly used by NPUs, the SRAM data throughput power consumption theory can be reduced by 99.2% at most, and the energy efficiency ratio can be improved by 200% compared with traditional NPUs.

The structure design of FIT dual core interconnection and near memory DLA enables the AI-ISP architecture of self-developed chip V2 to be established, and complements the ISP algorithm and computing power with the platform chip NPU to achieve the ultimate image processing effect and energy efficiency ratio.

Vivo's new generation of self-developed chips and other new technologies will be carried onto the Vivo X90 series. It's really not easy. The development of domestic mobile phones is really difficult. New systems, new images and new chips will not disappoint you.