Home > News content

ARM CEO Interview: Small Low Power Processors and Big Dreams

via:博客园     time:2019/1/16 18:32:08     readed:247

Today, ARM is not only a smartphone processor IP manufacturer, its ambition is making it break through its own borders and March in the direction of TV, autopilot, notebook, Internet of Things, etc. More importantly, in the embrace of Softbank, the development of the whole company has also undergone interesting changes. During CES 2019, foreign media Venturebeat interviewed Arm CEO Simon Segars, and Lei Feng compiled the interview unwillingly.

ARM

Note: The latest image processing chip of ARM can realize HDR.

ARM is currently in good working condition. Since Softbank acquired the company for $31 billion in September 2016, the company has added thousands of engineers and more than 6,000 employees.

ARM is headquartered in Cambridge, UK. Its business is to design energy efficient processors. Meanwhile, the global semiconductor market of $450 billion is moving in this direction. Softbank estimates that by 2025 we will have a trillion networking devices, thanks to the Internet of Things, which makes every object intelligent and interconnected.

ARM is taking steps to ensure that its chip design can be used in these areas, and it is also strengthening its partnership with chip manufacturing partners worldwide to ensure that these friendly manufacturers (valued at $10 billion) that produce ARM processors can meet the future needs of ARM customers.

ARM

Note: Simon Segars, CEO of ARM, at CES 2019.

VentureBeat: The idea of integrating technology with ordinary products and transforming them on the Internet of Things seems to be a big theme for you this year.

Simon Segars: Yes. By today's standards, by paying our partners $5, you can buy a chip with relatively low processing power, but by adapting the right sensors, you can meet all the computing needs you need.

VentureBeat: Did you know the areas of your product? Have you calculated how many designs they ultimately have?

Segars: Oh, No. In things like this, they may have bought a chip through a distributor, so even our licensees don't know what kind of product it will become. As far as I know, ST or other distributors who sell thousands of standard products through distribution channels may be more involved in product design than anyone else in the entire supply chain. We did make some estimates of the number of different SOCs based on ARM. I don't remember the exact number, but it's a big number.

VentureBeat: What's interesting about ARM on CES this year?

Segars: Things like this always interest me. For example, take these very cheap chips and use them to do creative things. As you said, ordinary devices are enhanced by relatively low-cost technologies. I always like to see people do that.

This year's theme seems to be about 5G technology, which is a consumer show, so you will have a lot of questions about what changes 5G will bring to people. But for me, 5G is the foundation, for many products delivered in 2019 and beyond. This is a very critical moment for us.

Many physical devices collect data and assist AI technology. There seems to be another competition about who has the best AI technology. Home appliances, digital assistants, these ongoing developments are very interesting. Then there is the popularity of the product itself.

VentureBeat: I just walked through the wearable equipment exhibition area. Now there are many unknown Asian companies making wearable devices. It shows how low the cost has become and to what extent technology can be popularized. It's not just Fitbits, a cool Western startup. And there are people you've never heard of.

ARM

Segars: Again, by today's standards, this seems like a cliche, but you can buy a low-cost microcontroller with the right sensors, and designing such a product is just a matter of programming. Automotive products are another growing area of consumer electronics. I heard they had to drive to another exhibition area because the normal place was full. Unfortunately, I don't have time to walk around the exhibition.

Segars: Mobile phones have been a driving force in the technology industry for many years. We have always been closely related to it. It builds our roadmap. For a long time, the technology we developed for mobile devices has found ways to enter other fields. I always like to focus on the TV industry on CES to see how large they are and how complex the user experience of TV is. This is driven by chip technology for mobile devices. Today, television is like a smartphone, with a very large screen. From one technology to another, it's simple and rough.

Generally speaking, the growth of mobile device business promotes the progress of higher performance, higher integration, sensor and efficiency technology. Everything else is like this. For example, P&G's cosmetic equipment is the beneficiary of all mobile technology. You see it spread to other areas. Modern technology is penetrating into every field. 5G will provide its interconnection technology. Cars can also be interconnected. It takes advantage of all the advantages of mobile technology to create a complex user interface in the car, and over time, it will promote computing power and ultimately achieve automatic driving.

Mobile phone business has always been a huge driving force in this industry. When I see people talking about how the growth rate of mobile business has become flat, I think it's wrong to draw a picture of the end of mobile phone business around this point. The mobile phone business has brought a lot of things, and it has become ubiquitous. Even in laptops, we can now see ultra-low-power, ultra-long battery life laptops running Windows 10 systems, providing instant network connections through built-in modems. Similarly, it acquires many advanced technologies from mobile devices and applies them to different devices.

ARM

VentureBeat: In the market, do you measure success by your share?

Segars: Yes. In Windows 10 devices, it is still in its early stages, but a few device manufacturers have begun to offer products. I've been using Lenovo's Yoga C630, which is a great device. There's also a two-in-one laptop from HP. They are two in one and very useful. Batteries will last forever. You can use it just by opening the screen. It was a great experience.

VentureBeat: Many big manufacturers have a Chromebook product line, and now they also have Windows 10 laptops based on ARM processors.

Segars: I think it's similar and different. Chromebook is a different category of use. I think Windows 10, which can be connected at any time, is a different case, but it is an extension of the current case. People are used to lifting laptop lids and wondering how they can connect to the Internet. Should I pair my laptop with my mobile phone? Did I find Wi-Fi? Do I believe in Wi-Fi network security? The technology that has been connected to the Internet has eliminated all the contradictions. It's changing the way you look at that device.

VentureBeat: I also saw Kohler's smart toilet.

Segars: What kind of intelligence does it have?

VentureBeat: It can sense you're coming and lift the lid. It can accomplish self-cleaning. It also has lights, and you can program some tasks using matching applications.

Segars: Toilet network. You don't need a lot of electronic devices to do that. But I hope to see this when I decorate my house. But it doesn't do medical analysis or anything like that. Do you know there are such cases?

ARM

Segars: Yes, Renee did a good job in Ampere. AWS deployed Annapurna Labs laboratory equipment, these artificial intelligence devices, in the real sense, brought 41% cost savings. Just on Monday, Huawei just launched their server devices. You can see what they have achieved. That's it. A lot of things will happen in the field of ARM servers.

VentureBeat: Strategically, how much do you value this particular market?

Segars: That's important to us. When we think about how computing develops, we think about many different terminal devices, whether they are P&G products or mobile phones. It spread like wildfire. Network and data center is an important computing category for us. We've been investing in that roadmap. At TechCon, where we last met, we launched Neoverse, a product specifically for this purpose. We want to be involved because it's a fast-growing market.

I can foresee the interconnection and integration of the devices we deploy in the ARM-based data center and the devices we deploy in the automobile. This will be the basic technology we see elsewhere.

VentureBeat: Since the acquisition, ARM has added thousands of people. Which direction do they all have to go?

Segars: Actually, it's very comprehensive. More of us are committed to the core roadmap, providing design and design changes for different markets. Just a few days ago, we launched another component, A65AE, designed specifically for automobiles. It is a processor designed for automobiles with safety function and locking function. This stems from our evolving core engineering technology, which enables us to adopt core processors and create changes with the right functionality and performance levels for mobile, automotive and business.

ARM

Venture Beat: I usually describe the competition with Intel like this. Intel is top-down, you are bottom-up. Who do you think will win the game?

Segars: Well, I don't know who will win, but we see a lot of opportunities to apply energy-efficient processing technology to the silicon semiconductor devices that our partners will be producing. A large number of software can be adapted around the standard architecture. We are integrating all the expertise of our partners. Applications in this area are growing.

The annual Consumer Electronics Show seems to be more diversified than the previous year. Some years you'll see a big change, and some years it's growing in volume, but every year when I come here, I see people putting it together in different ways and integrating it more and more. I don't think this trend will slow down soon.

VentureBeat: Try to solve the most difficult problem in this way. Do you try to do it through the collaboration of small devices? Like a bee in a hive? Or is there another way?

Segars: Our approach is to turn high-performance, low-energy processors into the basis of many devices. We have a set of other IP that can be placed where the computing elements need to interact. We have done a lot of work on integration, such as a CPU, a GPU and a neural network accelerator, and so on. We will consider how data flows among security components, interconnection components, and I/O components. It creates a structure under which our licensors can mix, match and build using standardized interfaces.

We're trying to provide a lot of things that are common in devices. There is no real distinction between different types of end products. This enables our partners to devote R&D to the most differentiated things. But what we really need to do is to attract new ideas. We work with hundreds of companies and they come up with ideas that we can't think of.

VentureBeat: It's interesting that you're starting to take the initiative to push chip manufacturing technology forward. You said just now that Intel was supposed to be in charge of this job, but now the world is different. Someone has to take some responsibility, from 10 to 7 or whatever. This is some of your ideas, trying to achieve it?

Of course, that's true. We are now in a position to work with such companies as TSMC and Samsung. I was having lunch with E.S. Chung, the head of Samsung's industrial business, and he talked about the direction of the new transistor structure. How do we focus on these optimization designs? How do we rebalance pipes when the performance ratio between connections and memory and transistors changes during these transformations? How will we run the early equipment together to test and understand this?

ARM

VentureBeat: Does this look more like the responsibility of the integrated device manufacturer?

VentureBeat: You seem to have been focusing on design and manufacturing for factories. This is a good way to make a clear division of labor. But that doesn't seem to happen anymore. You need this mode of cooperation.

Segars: Yes, you know. If you go back 10 years, see how people describe how the industry developed. Initially, IBM did everything from product design to transistor and even device manufacturing to ASML-like devices, followed by a TSMC that made wafers. You have ARM and EDA companies that provide IP and design streams. Then you have a designer. This seems to be a very neat, segmented area of expertise where you can concentrate on getting the best products.

As everything becomes more and more complex, communication becomes more and more important. We are focusing on collaboration methods with OEMs, design companies and EDA companies. It provides what I think is the best of both worlds. In our case, we focus on the expertise we are good at, and EDA does what they are good at, but we work closely to ensure that when all these jobs are combined, there is no loss in the combination. The differences between jobs need to be minimized.

VentureBeat: Qualcomm Qualcomm, a long time ago, didn't care much about the process. Now they have to invest to push it forward.

ARM

VentureBeat: I always thought Sun Zhengyi's discussion about singularity was interesting. How does this affect your thinking as part of such a larger plan?

Sun's way of looking at the world is interesting, and he has invested a lot of money to realize this vision. It's fascinating to be one of them.

VentureBeat: I almost thought he would be rashly interested in Nvidia or other supercomputing companies, rather than finding ARM, because it doesn't matter on the road map.

Segars: Part of Sun's interest in ARM lies in the unprecedented wide application of our technology. In such a situation, I can spend a meeting with a person on autopilot, the next meeting with another person talking about cell phones, next meeting and a person talking about network construction. We are in an absolutely unique position to understand the future course of development in all these different areas. This really gives us some insight into the future.

VentureBeat: If you also think that beehive thinking is the way to achieve singularity, that makes sense.

Segars: Of course, yes.

Segars: I don't know if Amazon has killed the retail business. But they changed the whole industry, and now they know too much about people's retail habits. You can't buy all the food just through Amazon, but you can do it now through Whole Foods. It is a source of other data.

Softbank has changed a lot in the past few years since we were acquired. It's very important that companies understand what's happening in the world. How to use these long-term trends to invest in companies? More importantly, this is the current concept to promote the development of Softbank.

ARM

VentureBeat: Compared with these 30 years of investment, what do you think will most affect your daily investment decisions?

Segars: Daily investments are largely driven by the following factors: What extensions do we need in our portfolio to enable computing platforms for these applications to run on ARM? We are at a moment to complete the shift from mobile phone business-driven, in a world where technology from mobile phones is used in many other areas. What is the next big driver? The next major drivers are 5G networks, artificial intelligence and driverless. They all overlap. We are investing in computing platforms to drive growth in these areas, whether data centers, automated cars or Internet of Things devices. It's about generating data, it's about transmitting data, it's about processing data. Data is the driving force of next generation computing.

VentureBeat: Are you excited about what inspires your imagination? I saw some advances in prosthetic technology, which was exciting. It looks cool.

Segars: Yes. What's exciting is what people do with this technology. In my career, technology used to be difficult to use, expensive and slow to develop. Now it's easier to use. Programming language. I learned assembly language 30 years ago. Now you can program in Python and create a neural network with one line of code. Productivity has increased substantially. The threshold for the use of computers has been greatly reduced.

ViaVentureBeat

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