Electric Bike,Electric Fat Bike,E Bicycle,Electric Mountain Bike Wuxi Yizhiling Locomotive Development Co. LTD , https://www.yizhilingev.com
What is the current status of AI chip technology development?
In the past 2018, many chips have sprung up, and various computing power and various application scenarios have been produced. The number of floating-point operations per second ranges from 1T Flops to 16T Flops, from the cloud to the edge. It has multiple modes such as single chip and SoC integration.
There are two ways to implement an AI chip:
The first is called the front-end solution. The AI ​​chip is placed in each camera, and the data is processed in real time and then output to the back-end server.
The second is called the back-end solution. The AI ​​chip is placed in the back-end server, and all the original data is summarized here for centralized analysis.
According to the HIS research report, the IPC SoC chip shipments have a compound growth rate of 55.9% from 2013 to 2018. This SoC chip that integrates intelligent analysis and network transmission is a battlefield worthy of all efforts.
In the field of SoC, Huawei HiSilicon has formed an oligarch in the sense of fact. In just a few years, it has increased its domestic market share from 37.3% to over 74%, and has pulled down TI and Anba. Currently, in the network camera (IPC) SoC chip market, Hisilicon is a big one.
In addition, various cross-border AI chips are rapidly infiltrating into security surveillance, and Intel has invested heavily in end-to-end, as well as hardware and software. In addition to the acquisition of NervanaSystems and Altera, it also acquired artificial intelligence companies such as Movidius and Itseez. After the establishment of the Artificial Intelligence Division in March 2017, Intel acquired Mobileye for $15 billion and established a $1 billion AI Innovation Fund.
Many well-known security companies in China are using NVIDIA GPUs, Hikvision uses NVIDIA GPUs and JetsonTX1 platforms, and NVIDIA's "Smart" series of NVIDIA Tesla P4 GPUs, Yushi Technology Cloud Intelligent Analysis The server supports 80 NVIDIAJetsonTX1 processors. (Author: Guo Hui)