Huawei / Apple Is Eyeing Qualcomm "hard And Soft" Layout Ai Market

Handset  chip maker Qualcomm has now launched three generations of platforms  aimed at artificial intelligence (AI) and continues to focus on AI  applications and the hardware and software needs that they require. As  AI continues to grow in the handset industry, its competitors are  gradually increasing. Mobile phone brands such as HUAWE and Apple have  also been involved in the development of this sector, and related  solutions (such as Kirin 970 A11 Bionic). In this regard, Qualcomm said that compared with Huawei, Apple and  other mining hardware AI development strategy, the company still take  the "hardware and software" approach, in response to AI development.

Gary  Brotman, director of product management at Qualcomm Technologies, a  Qualcomm subsidiary and Qualcomm Technologies, Inc., said Qualcomm has  released three generations of platforms aimed at AI and has been  concerned with AI application cases and the hardware and software  requirements they drive. For the time being, the company does not think it is necessary to use  proprietary AI hardware in its current use case, of course, that does  not mean Qualcomm will not consider the development of dedicated AI  hardware in the future.

Brotman  further explained that the current challenge for the industry as a  whole is that AI develops very rapidly and algorithms evolve almost  daily. Therefore, the use of dedicated hardware platform has a certain risk,  because in the development of the early need to predict the development  of the field of AI in the next 18 months trends and hot spots, and the  corresponding technology into a dedicated hardware platform.

In  other words, the general hardware from the design and production to the  market, the cycle takes about 18 months, and the AI algorithm evolves  very quickly, when the hardware really on the market, the previous  solution may no longer fit. In addition, the development of dedicated AI hardware will also increase the corresponding chip costs.

Brotman  shows that most developers now use custom run-time layers in their AI  algorithms, but the specialized hardware is not efficient enough to  handle these custom layers, so developers have to call other modules  such as GPUs or DSPs deal with.

However, on the software side, software tools from either ARM or other vendors offer very high levels of flexibility. Developers  not only have access to the underlying hardware, but they also show  that they have the flexibility to take full advantage of hardware  capabilities through programming in some situations, such as the  Snapdragon platform's DSP, to reduce the risk of developers altering  algorithms. Therefore, Qualcomm in the future development of AI, will still be a  combination of software and hardware to deal with artificial  intelligence needs.

For  the future layout, Brotman revealed that in the coming months, the  company will release a benchmark system for AI performance, focusing on  the following three aspects. The  first is the performance of the AI processing case, including speed  and number of frames; followed by power consumption; and finally  precision. Qualcomm hopes to strike a balance between those three points, and if  blindly pursuing the efficiency of processing, its accuracy may decline,  so I hope to have a more balanced evaluation system.

Brotman  said that, in fact, compared with the benchmark test, the application  of the case is the real touchstone, AI performance can be the most real  situation.