We met up with ARM president Simon Segars at Mobile World Congress to chat about the company's recent milestones -- 8-core big.LITTLE processors, the powerful yet efficient Cortex-A50 architecture (which we discussed with James Bruce last year), MediaTek's Cortex-A7 quad-core SoC (the first of its kind) and Samsung's octa-core Exynos 5 chip -- to name a few. Hit the break to watch our video and read the interview transcript.
Myriam Joire: Hi there, it's Myriam with Engadget and I am here with Simon Segars, the President of ARM. And we're here at Mobile World Congress in Barcelona and I wanted to talk to you a little bit about some of the things we've seen in the last few months. I mean, last time I talk to you guys it was with James after the multiple A50 announcements. And you know, obviously there's been a lot new thing since then including some octa-core processors and MediaTek announcing the first A7 quad-core. Let's talk a little bit about some of those new products.
Simon Segars: So I think what you're seeing there is breadth of different types of implementation that can be created around our cores. At one end, the A7 could be used for very low-cost devices. We're really excited about the volume, the size of the market opportunity that comes around that. And then at the other end of the spectrum, you've got big.LITTLE configurations, our 64-bit cores, which can take ARM into very high-performance computers. So I think what's really interesting is how many different types of end application can be built hitting different performance points and crucially different pricing points - just growing the breadth of the market that ARM can address.
MJ: So tell us a little bit: what is your vision in terms of the ARM group? I realize that you are providing technology that you license for people to create their own chips, right, but like you obviously have some direction that you provide along with that, you know, like a semi-conductor manufacturer who provides a data sheet as it were for using their chip. I'm sure you are also providing some advice to people who license your technology. How do you see these crazy octa-core that are made of four A15s and four A7s? How do you see that developing going forward? Are we going to see some other combinations of these cores?
SS: Well yeah, the say octa-cores... I wouldn't describe it as 'crazy,' of course.
MJ: I meant that in a good way, obviously.
SS: Yeah, I think you kind of hit it on the head there. We're providing technology that can be used in multiple different configurations. So at one end, we have very power-optimized cores like the A7. At the other end, very high-performance cores like the A15 and A57 in the future. That gives our partners the flexibility to build different configurations to hit different price and performance points. We don't give a lot of guidance on what the right answer is because we really want to leave that up to the market, up to our customers and their customers to work out what the right combination is. Really what we're focused on is the flexibility of the way you can put all these building blocks together and try to simplify that as much as possible.
MJ: Right, so have I see in front of you, here one of the Alcatel, I believe it's the Scribe?
SS: Yeah, so this one happens to be branded TCL. This has got the quad-core MediaTek chip in, the quad-core Cortex-A7 MediaTek chip in.
MJ: So that was the first quad-core A7 chip, right?
SS: Yes, I believe so.
MJ: And how is that significant do you think in terms of market right now? Obviously the price point is pretty amazing.
SS: Yeah, and I think that's really the significant thing. The level of performance that you get in that chip, they've integrated into their graphics, and video, and the modem. So it's a very highly integrated SoC with a quad-core A7, which is delivering a lot of performance. And in that quad configuration, depending on how many of the cores are switched on at any moment in time, what the frequency of voltage is set at, then you've got a very wide kind of dynamic range of performance and power that you can choose between depending on the complexity of end application that you're trying to run at any moment in time - so very flexible performance and highly integrated and a very low cost point. I think it's going to mean that we'll see some really useful and interesting smartphone being produced at very low cost points - sub-$100 - which is going to enable us to address very large developing economies.
MJ: So obviously with quad-core becoming affordable, the next question is, you know -- and octa-core coming to life with these, you know, combinations of core types being built into the same chip -- where do you see this going? Are we going to, ten year from now, have these like 16 core or like multi-core devices that are going to be prevalent or do you think that from a software engineering point of view it become a challenge and it will kind of top off at a certain comfortable level. You know the PC market for example in terms of Intel architecture seems to have comfortably settled around dual core. Do you think there is something like that is going to happen? What is your prediction in that area? Obviously you don't have control over that, right?
SS: Yeah, I mean I think what you see today, having eight cores on the same die seems like, to use your language, crazy. I mean it's a lot of processing capability in a phone, and on one level it is crazy to think, "Well, what is all that going to get used for?" But the great thing about phones is the fact that it's a very open platform. Developing software is very easy, very low cost; and as you put all this performance in, somebody will think of something to do with it. So eight cores seems like far more in the land of compute power today. Will we see 16, 32 in a few years time? I think two things are going to govern that. Basically the drive of Moore's Law driving cost down meaning that you can put all of those transistors in at very low cost. A few years ago having a quad-core in 28 nanometer at very low price point seemed just unrealistic. But with, you know... technology always gets cheaper, and it's amazing what can be done, and MediaTek have demonstrated that. So overtime I'm sure you'll see more and more cores. And there is a limit to how much parallelism you can get out of any one application, but I think you'll see the different processors being used for different tasks in parallel. Where the limit is of that I don't know. I mean when you think back, you know, the iPad is only three years old. Smartphones themselves are a relatively new product category, and I don't think we're at the end of people just working out what they can be used for by any means.
MJ: No, there's no doubt. In terms of really low power, high efficiency, low compute power ARM cores like for example what's in a device like the Pebble for example, what needs to happen in that area? Do you think there are some significant breakthroughs coming up?
SS: Well I think a lot of the technology building blocks are there. I think some of the chips we're going see built around very small, very low cost micro-controllers integrated with sensors. Again, I think we're kind of at the tip of that, and again, it's made possible due to the technologies all coming together. And building a chip to do that kind of thing is much less expensive than building a chip to do an eight-core application processor. So I think you're going to see a lot of new device coming out with people using these building block in different ways. And I think it's going to be pretty interesting to see what people do. And then, with low cost connectivity to your phone, and having your phone control that or aggregate the data and interact with cloud services, I think it's going to be pretty exciting to see what people think up, what the end applications are that never have been imagined before.
MJ: Yeah, I absolutely agree. Well listen, thanks for your time, Simon. It's been lovely getting your insight on, you know, some of the kind of forward vision towards what ARM can bring. I mean, you guys are pretty much ruling the world right now so it's an exciting time.
SS: It is an exciting time. Again, from our perspective it's about great technology and in enabling people to build all these exciting products. You know, we don't have a monopoly on great ideas for all of this. What's great is how many users can just interact with technology and build end devices. It makes for a fun time.
MJ: Great. Well thanks for you time again.
SS: Thanks a lot.
MJ: Cheers.
