Shulou(Shulou.com)11/24 Report--

Since the head of Invida, Huang Renxun, put forward that "calculation power is the new horsepower", calculation power will replace mechanical parameters such as horsepower and torque and become an important standard to measure the power of automobile products, which has been gradually accepted by the industry.

Once upon a time, the Qualcomm 8155 cockpit platform was called the "car gauge chip ceiling" and became a yardstick for measuring the technological level of a smart car.

When entering the 2023 "smart car 3.0 era of competition for the first year" node, the configuration of the inner volume to a higher level. At the time of the large-scale production and landing of 8155 cars, Qualcomm 8295 chips have been launched on the car robot Ji Yue 01 and will be delivered this year, while zero running, Zhiji and other car companies have also followed suit.

Advanced 5nm technology, 8 times the horror of 8155 computing power, how to enable smart cars, is very likely to subvert our current imagination of smart cars.

Some industry insiders speculate that the Ji Yue 01, which is equipped with Qualcomm 8295, 35.6in screen, fully offline voice and high-order intelligent driving, may be another "dark horse" of the 300000 level. If Ji Yue 01 can hit less than 300000, it will make the already internal volume of the car market more competitive.

From horsepower to computing power, how does the chip become the core competitiveness of the car?

In China's new energy market, consumers' preference for intelligence is much higher than expected. According to data from third-party market research firm Jundi J.D.Power, the intelligent ability of cars has become the highest ranked factor in Chinese consumers' purchase decisions.

This also has a profound impact on the actions of car companies, such as the ideal as an intelligent follower, but also after the basic stability of the business, began to increase intelligence on a large scale this year. And more car companies actually show a deep "intelligence anxiety", quickly upgrade the chip is one of them.

Behind this change is the change of the logic of automobile competition. In the real era of smart cars, the intelligent ability of cars is no longer just cool on the big screen, but also the "computing power" behind the big screen and the functions based on super-powerful computing power.

Qualcomm's performance in the smart cockpit is showing its dominance in the field of smartphone chips. After the exploration of the first generation Snapdragon 602A platform and the second generation Snapdragon 820A platform, the performance explosion was achieved on the third generation cockpit platform Snapdragon 8155A, and firmly occupied the core position of the intelligent vehicle cockpit chip.

It is no exaggeration to say that in today's smart car market, without 8155 support, cockpit systems will greatly reduce the attractiveness of potential car owners. The sight of car companies "saying that they must be called 8155" is also very similar to the old days of competing for the first opportunity of Qualcomm chips in the smartphone era.

As mentioned earlier, the 8295 chip, which has an eight-fold increase in computing power of nearly 8155, is about to be delivered on the car robot Ji Yue 01.

Obviously, if a smart car can be launched with a whole new generation of high-end chips, it means that the smart car itself is a technology leader. Secondly, for manufacturers, the first launch carrying "grab" more time advantages, can advance competitors to carry out intelligent function research and innovation, leaving more intelligent expansion space for models, let smart cars on the market that is to take the lead.

So how will the 8295 chip enable smart cars? Can the industry's advanced computing power really solve the "intelligence anxiety" of all car companies?

Why is it so important that the same chip can make a big difference in the "math adjustment" of smart cars?

In the era of intelligent car, computing power, algorithm and data are the three key factors that constitute the intelligent ability of automobile. If the mechanical structure of the car is the "skeleton" of the vehicle, then the data can be called "blood", the algorithm is "brain", and the arithmetic is "muscle". Several major elements cooperate with each other to form an excellent smart car.

In other words, the intelligent ability of the car, in addition to the bigger and stronger "computing power", is also closely related to the data, algorithms, and the quality of the whole vehicle, especially the algorithm ability.

If the core data of the smart car is missing, the algorithm lags behind, or there is a deficiency in other intelligent hardware, even if it has a "violent computing power" of up to 1000 TOPS, it cannot be called a real smart car, but can only be a "muscular and simple-minded" monster. (TOPS is the unit of computing power of a computer. 1TOPS represents a processor that can perform 1 trillion calculations per second.)

In a word, you can't build a real smart car just by math. In other words, a high-power chip is a good thing, but it is just like reference materials. If you don't buy more, the better your grades will be. If the car company itself does not have the corresponding intelligent support such as software capabilities, even if it uses the same chip, it can only produce a large number of wasted "invalid computing power".

In recent years, the traditional mainframe factory has continuously increased its investment in intelligence, and has paid a lot of "tuition fees" in order to compete with the new forces, but limited by its deep-rooted traditional attributes and weak software capabilities, the effect of intelligent transformation has improved somewhat. But it also seems to be "stumbling".

For example, although the models of an independent factory are equipped with the top 8155 chips, there are still some problems, such as stuttering, lack of applications, slow opening time, low resolution and so on. The actual intelligent experience in the cockpit is not satisfactory, let alone in the lead.

Pincha even reported that most car companies are actually "wasting" Qualcomm 8155 chips, and even the most basic fluency has not yet been done, let alone to give full play to the potential of 8155, it may be redundant to use the previous generation of 820A. Under such circumstances, most car company officials announce that they are actually playing "billboards" when carrying 8155.

Nevertheless, this is not to say that the excessive pursuit of on-board chip computing needs to be criticized. After all, if smart cars do not have the "muscle" of computing power, they will not be intelligent at all.

So, how to let the well-developed "muscle" burst out of its due strength? Of course, it depends on the smart "brain".

Take, for example, the first car robot Ji Yue 01 with a Qualcomm 8295 chip. In fact, the anthropomorphic analogy of intelligent cars above is the same as the idea of developing car robots.

In the product concept of automobile robot, highly intelligent automobile robot is divided into three parts: "brain", "nerve" and "torso". Among them, the ability related to intelligent driving and intelligent cockpit is the "brain", the intelligent architecture JET is its "neural network", and the whole vehicle structure and power always become the "torso".

JET intelligent architecture is the underlying foundation to create the shape and ability of automotive robots. It can achieve "global integration", which is equivalent to opening up the second pulse of the governor and maximizing the use of 8295 of the super-high computing power. In addition, the basis of the algorithm determines the intelligence of the brain. Ji Yue 01 wisdom module and wisdom drive have Baidu AI, Apollo algorithm enabling. Baidu AI has more than 10 years of research and development accumulation, and Apollo itself is also a L4 autopilot algorithm.

So the effects, such as ultra-smooth interactive experience, ultra-delicate high-definition visual experience, offline voice as low as 500ms recognition speed, and even Qualcomm 8295 can take over when the smart driving system fails, fully ensuring safety.

Without exception, traditional car companies are making every effort to make intelligent transformation, and new forces are also spending a lot of money on intelligent research and development. To be fair, almost no car company has 10 years of AI technology accumulation, the lack of powerful AI algorithm is one of the reasons why many car companies can not give full play to the chip computing power.

By comparing the applications of 8155 and 8295, we find that the next generation of smart cars use the underlying logic of chip computing. Fundamentally, the product should be user-oriented, if it is only blindly strong on the high-power chip, but does not play the value of the AI ability, it can only intangibly allow users to increase the cost of buying cars.

For car companies that use 8295 chips later, only by practicing the "internal skills" of AI, taking this "required course" and maximizing the potential of high-power chips can the real universal benefits of science and technology be realized.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.