 to join me in welcoming Renee Huss, Executive Vice President and Chief Commercial Officer of R, Renee. Good morning and me how? Welcome to our press briefing. Very much appreciate you taking the time to join us this morning. You know, as I was thinking about this show flying over here, this is my 10th Computex. It's not my 10th year in the industry, I've been around a little bit longer than that, but I was thinking about, 10 years ago when I first attended the show, what were some of the key themes, but more importantly, what were the devices I was bringing with me in my briefcase, not only were we using backpacks. And I remember the biggest thing that that show in 2006 was the fact that laptop computers were just starting to overtake desktop PCs as the primary computing device. And at that time Computex was really a show that was largely around the PC ecosystem, whether it was the OEMs or the OEMs and all the things that make Taiwan sort of the rich community that it is around this industry. And I was thinking back, I think the phone that I had with me at that time was a motorboat flip phone and a StarTac that if you wanted a text message to somebody, you had to be really, really good at knowing that the combination of keys that would sort of get your message out. So, you know, 10 years later to be here talking about the launches of new products that we're going to share with you today, really struck by the change that's happened in this industry and particularly as it impacts Computex and the show here and the things that we're doing. Obviously we're going to talk to you about today or some of the new products in the whole space which is a space that we continue to see huge growth and huge opportunity. And it goes without saying, and I think we all know this now that we have, obviously everyone in this audience for sure has a smartphone with them, I have two. So when you look at the use cases in terms of numbers, they're huge, right? And when you look at the numbers relative to what is going on with a smartphone, the numbers alone are just staggering. And it's really an output of what the smartphone means to us in our daily lives and how we use it. And again, if we just look back at a number of years and think that the amount of computing power that was in that laptop 10 years ago is far surpassed in terms of what's inside the phone today. It's pretty staggering. And with that, the smartphone is really the Swiss Army knife, if you will, of all things computing. It is a device that you can get sports scores on. It is a device that you can get a taxi. It is a device that you can obviously work on. It is a device that you can load the old make phone calls. So it is a platform that has so many different use cases that can be applied in so many different areas. It's really just, we think at the very beginning of some interesting new cases before that we're gonna talk about today. And with that, when we just look at just the numbers, three billion smart phones that are our power that are in the world today, if you do the math relative to whether it's in north of maybe seven billion people on the planet, three billion is a pretty large number. So you ask yourself, are we starting to saturate? Well, I have two, so maybe not. You can read the growth numbers here in terms of the population in terms of what's driving the use cases. And again, if you look at the demographics in terms of some of the younger audiences. But I think more important than just the numbers of people who use the phones, you start thinking about, well, what's gonna really drive growth in this market in the future? Is it more people who are just going to replace their phones or are there use cases that are going to really drive innovation? And that's what we're gonna talk about today with some of the new products that we're launching because we believe that right now we are looking at some really, really unique use cases around mobile computing that actually even 12 months ago were not something that we thought was possible to be done on the mobile platform that we're really super excited about. And I think one of the things to keep in mind again as we talk about these new use cases is the fact that building on top of this mobile platform is really a natural outcome. When you think about the software investment, when you think about the number of different partners in the ecosystem somewhere here with us today, it's obvious that the mobile platform is going to be the platform for growth in these engines. And it's really only in mobile that some of these use cases can get the kind of growth and adoption because the platform is so ubiquitous, it's used by so many folks. The developers are tuning applications to these platforms. So then when you think about these new use cases, it's really sort of a natural outcome that mobile will be where it's at. So some of the cases I'd like to talk about today, and again, Nandana, he goes through his presentation on the product side. We'll talk about a little bit why this is so important. And again, these are areas that even 12 months ago we might not be considering as real use cases for a handset. One is in the area of augmented reality. And I'll talk a little bit about what those use cases are. Gaming. Obviously, gaming is a use case that's existed on mobile, but when you start thinking about now the resolutions that are involved, the kind of performance that can be done, the kind of things that people do on their handsets with gaming, we're ushering in a brand new era, particularly when you give the fact that these kind of highest performance applications require low power, small footprints. We can do some very innovative things. And then virtual reality. Obviously, virtual reality is a very hot market. There's a lot of buzz around it. We're going to talk a little bit about some things that Farm is doing in that space. So first off, I'm going to talk a little bit about graphics virtual reality. This chart here that you take a look at on the left-hand side shows a little bit of the sub-systems that are required inside the handset to make graphics virtual reality go. And the first thing that sort of strikes you is that the area for the CPU and the GPU are on and highlighted. Doing virtual reality in a graphics space in the handset requires a lot of CPU horsepower. It requires a lot of GPU horsepower. It requires, obviously, memory in terms of all the data you're moving around. And it requires a lot from the display processor. So, non-done again, we'll go through a lot of detail on this, but graphics virtual reality and the use cases around it, we're bringing some of the subsystems to this world today that will address these solutions going forward. And it's very real. There's a demo back there. I hope you have a chance to use it and take a look. I would warn you that maybe hopefully you'll be less digested before you try that. It's pretty real. And this space is going to, we think, really launch a brand new set of really, really interesting sets of applications. And if you can do that around the global platform, again, where the developers have a very rich set of experiences that are important to tailor to the platforms, the growth will be really significant. Now, when you add on video onto that, again, if you look at this little chart that talks a little bit about the use cases, now you've got the CPU, which is helping do some of the work. GPU may be not so much work in this case, but you actually need a video engine, a video processor that can help you with all the work required for video virtual reality. You can read some of the modeling there from a 3D standpoint, but doing video on a mobile platform for VR takes a lot of work. There's a lot of energy required. There's a lot of performance required. There's a lot of use cases that demand a real high degree of performance. The mobile platform, again, is well-tuned for this. It's established. We have a lot of use cases that have already been proven. And this is an area where we think developers will be doing a lot of work in the future. So gaming, so when you think about gaming, you think about rich titles, triple-A titles, the very best titles that people want to run. Again, this is something that's now just finding its way to mobile. It existed in the high-end world, and desktop, and laptop, and now it's finding its way to mobile. And the reasons for that is a lot of innovation going on inside the GPU, where now you can have a tremendous amount of performance in a power envelope that doesn't drain your battery. So these triple-A titles that people are used to playing on high-end consoles or playing on laptop PCs, people want to play these on all their platforms. People don't want to be restricted to just a single platform to go on and use those cases. This is an area where having a mobile GPU that can be very, very efficient on power, very, very efficient on doing innovation relative to accelerating titles, whether it's around 3D, is a very, very significant use case. And another very interesting use case is around 4K video, and this is in the area of not only capturing images, but doing post-processing, or anything around image manipulation. And in fact, just last week or a week or so ago, we at ARM announced the acquisition of a company based in the UK called Catapult, that's doing work in this area and in this space that is processing. We think this is a huge growth area for the mobile platform. If you look at what's required from a performance standpoint, what's required from implementation area, you can see all the areas of that handsome area are illuminated. What does that tell you? There's a lot of processing that's required to go off and do this. And when you have these very, very large 4K images that can capture a very, very wide area, you also want to be able to do post-processing and editing of the images as you actually capture them. With this type of resolution, and this type of capability inside the handset, now in users you're gonna be able to do some really, really exciting things around 4K video and imaging. Augmented reality, again, it's really interesting that these are use cases, whether it's VR, or we're talking about post-processing or gaming, now augmented reality that you can do inside a handset. Augmented reality for those of you who aren't familiar with it, is the concept of adding information onto a existing scene. So let's say you're walking down the street, you've got your handset, and you want to put your handset up and find out where the next Starbucks is, and those of you who don't know me, now that's a very, very interesting case. With the augmented reality case, you can just put your handset up and it will map out for you in terms of what's ahead relative to locations. So you don't need to go into your handset, you don't need to type in where a Starbucks, you don't need to do anything via voice, you can just literally hold it up. This is not a new technology in terms of technology that's been implemented on higher-end platforms, and we've talked about it the past few years, but what is new, it's finding its way into mobile. And when you think about augmented reality as a use case of driving growth and driving innovation, it's a very, very real example of things that people will do. They'll find value from it. And again, just get another example of where the smartphone is platform to build this on. From a performance standpoint, you can also see that it requires a lot. There's a lot of heavy lifting that's required by all the elements of the sub-system, whether it's the CPU, whether it's the CPU, whether it's the video area. These are all areas where ARM brings a lot of leadership to this space. So it's, and again, now I'm talking more detail about this area, but it is a space that's very, very key for us. So with that, we are very, very bullish actually on the smartphone market. There are people who, full debate whether we've seen peak in this area, we actually think what drives growth is use cases and new areas of innovation. And we actually think in this world, we're seeing a tremendous amount of growth and unique applications. And the ones that we just talked about are actually some of the most exciting ones that I think we've seen in a very, very long time. It's gonna drive some real interesting opportunities. So just if you look backwards a little bit relative to what the performance benchmarks have looked like, and again, this is pre-MyFirst or post-MyFirst project, this is 2009, you look at some amazing performance gains in the platform. And I think this is important to look at on two vectors. One, it allows us to stand up here or at least allows me to stand up here and be proud about what the company's accomplished with all of our dedicated engineers and partners have done. But I think it also lays claim to, this is a platform that has a lot of innovation built inside of internal performance. So if you think about the use cases that I just mentioned, in 2009, obviously, you wouldn't be able to have any of those use cases be described as platform. Fast forward to now, and if you were to go maybe another five years forward, start looking at some incredible innovation that can be done on this platform. So, given the fact that there's a lot of work that's being done in innovating on these new engines, there's a lot of work being done in terms of pushing the process to get really, really good performance and really good area and power, these new use cases, I think, are just the tip of the iceberg relative to what's been drive road to this market. And if past history has shown us anything relative to the smartphone market, is that innovation can happen at an amazing click. So with that, I will now hand over to my partner, Nanda, and I told him I was going to try to pronounce his last name correctly. I still don't think I can, so I'm going to call him Nanda. And I'll let him take the next piece. Thank you. Thank you so much. Thanks, Kirene, for one not trying to butcher my name. My name is Nanda Manbully. I'm a leading marketing and strategy for the processors group. I also want to thank Kirene for kind of setting up both the opportunity and the challenge of the immersive use cases of tomorrow. But before we go there into the technology portion, what I'd like to do is take a moment to reflect on how far mobile visualization technology has come. The video you'll see demonstrates arms that enlighten mobile illumination technology running on a Mali GPU in a standard smartphone available today. Just look at how you have very realistic visuals, lighting, shading, et cetera. Let's take a look. This is what René would call the digestible format. If you want to actually have a full experience, you have the demonstrations here, but also we'll be showing that on our complex booth for the next three days. And what required literally a high-end graphics workstation and expensive software is now delivered as a completely untethered experience in a smartphone and, of course, running on something like a Mali GPU. Now, speaking of Mali GPUs, let's look at the numbers. We as a partnership have made Mali the highest shipping GPU. It's the number one GPU in the world today by numbers. Over 750 million units shipped in 2015, which is a 30% jump from 2014, and nearly a 15x jump from 2011. So we've come a really long way in a very short time. That's, again, part of how we work together. And from our standpoint, this success, we believe, is because we've been listening not just to the technology itself, but the needs of the market. It's not just about graphics performance. It's about solving system congestion issues. It's about putting compute together. It is about solving the software ecosystem issues. All the while looking at how to increase efficiency. And if you look at that, we look at the kind of approaching immersive experience needing another big step. And for that big step, we had to take a change and step in the architecture part. So Bicross is our next generation GPU architecture. I'll get to that in a little bit, but it's useful to see our evolution over the last six to seven years with our graphics architectures. So in 2010, it was a much more random in 720p with the Mid-Guard architecture. It's still a very, very popular product. The Mali 400 is both out of that. We started bringing in GPU compute, and larger resolutions 1080p and up with the Mid-Guard architecture. The Mali T600, Mali T700, 800 families are based off of that. And now, even though those higher resolution images that are coming in, it's still more of a consumption situation. For the next step, I think we need to go a little bit further. That's what Bicross brings. And remember, as we're trying to do this with more performance, with more capability, we still have to focus on efficiency because it's, after all, a mobile format. And Bicross brings with it a lot of innovations to support that. If we look at what we're trying to do, deliver more performance and more efficiency, we can classify that into three major areas. So the compute aspect of it, the system aspect of it, and the software. And just for making it a little bit interesting, we go left, the right, the center. So in terms of the compute aspect, firstly, Bicross brings in a more advanced scalar architecture to the instruction set. But we've focused on things that make it more interesting. Cross shaders enable us to compress more instructions together with less overhead. In this, we've driven word text shading will help us actually push a lot of the shading towards the end of it, actually reducing the amount of actual work we need to do on it. And then, of course, we've designed it with local registers so you can actually reduce the amount of wirelets as you build these GPUs. And especially in new process technologies, wirelets, as you know, are pretty important. Now moving to the right side, the new use cases are not just about pure GPUs. They are a lot more interactive with other components on the system on the chip. And so a lot of interaction with the CPU. And the Bicross-based components will be the first ones to have full heterogeneous capability. So you can use the same virtual memory system that the CPU uses and actually not worry too much about sharing data across. And more importantly, you have less of it growing off chip. And off chip is actually where a lot of memory effort and a lot of power is spent. And we're reducing that. And then finally, trying to make the software developers' life easier. Bicross was designed from the ground up for Vulkan. Which is the API behind a lot of the new content being developed. And with a very optimized back end for it, it makes the software developers, the game content, the developer's life a lot easier to bring content to market faster with the kind of performance and efficiency we need. Now with Bicross as the foundation, we can announce our first Bicross-based product, the Mali G71 GPU. It's targeted at delivering at least a 50% improvement in performance over today's leading premium GPU from our version, the Mali T880. Now, remember this. It is also the most scalable GPU. The 50% we talked about is based on a 16-shader configuration. But we know this is only gonna need more performance as we go up. And Mali G71 has been designed to go up all the way to 32-shaders. So there's a lot more upside in the performance on the Mali G71. We also expect the Mali G71 to first hit premium phones with SOCs designed in 10-metre technology. And that's where you'll see a lot of its key value requirements. But in spite of this, it is still independent of process technology. If we took the Mali G71 and compared it on the same process technology versus a Mali T880 today, it still delivers over 20% energy efficiency, meaning we're taking 20% less energy on the same process technology versus a Mali T880. It can, of course, scale up to 32 cores, but that's independent. Here's where it becomes that much more interesting. We take 40% less area on the same process to deliver the same level of performance. And as you can see, we need more and more of this compute. We need to make space to be able to fit more of this GPU capability in it. And that's very important. And last but not least, it reduces the external memory bandwidth by 20%. And again, as I said, the more you go off-chip, the more power you burn, we're trying to reduce all that as we get to these more immersive use cases. Now, we've compared it against our premium GPU. Let's compare it with things that you might recognize on the market, right? What we've run here is benchmarks based on the GFX venture of Manhattan, which are very immersive use cases or actually very big taxing on the GPU and compared it against laptops, mid-range laptops of 2015. And what you see is a 16-shader configuration of the Mali G71 matches the performance of a discrete laptop and versus an integrated GPU laptop. It has a comfortable lead. And the thing to note is, even if you look at the more compact laptops, you're talking about a 25-watt SOC versus a smartphone SOC that's four watts of realistically more than three watts. So delivering the same level of performance in almost an order of magnitude less power and that's really important. Now, if you look at what it can do now with this kind of performance, you can see where it's going in terms of augmented reality and virtual reality type applications. And as we get there, the constraints will really get more, what should I say, stringent. If you look at what you need in a situation like this, certainly for realistic VR experience, you need higher frame rates. 60 hertz is what you see today. You'll probably get to 120 hertz, especially in more motion-oriented virtual reality. Resolution goes to 4K pixels as they get closer to your eyeball rather than the screen. That becomes that much more important. And if you're talking about a truly interactive experience, you have to minimize the latency, let's say from the head motion to full-time delay for the real sense of realism that you expect from it. And putting it all together, when you've got everything close to your eyes, you start seeing edges, jagged edges a lot more, and that's where the anti-aliasing aspects of the Mali GPU come in very handy. These are done within the architecture extremely efficiently. Now that's a lot of performance improvement, right? Or that's what's demanded. But also now in a virtual reality situation, you're not designing it for one screen. You're designing it for two eyeballs so that it actually doubles the meaning for that performance. You have to do all that and stick within the same power budget because your phone size is not changing by much and the same thermal budget which is actually that much more interesting. And just to put it into context, and if you go to the display over there, you'll start seeing the phones as they had evolved. Look at six years ago, the HTC Nexus One. 12 millimeters thick, not bad, it's actually a very compact phone. You see the Huawei Mate 8 this year under eight millimeters thick. What was, by today's standards, a very basic gaming experience on the 12 millimeter phone to effectively a VR quality experience today on the PA. So you can imagine how much more you need to do to make it as efficient to fit it to those thin phones because you don't have any leeway in terms of the amount of heat you need to dissipate. We're talking about how the GPU is getting efficient to support that. We need the rest of the components of the chip and the CPU being a major one to do that as well. So to marry up the G71, we're also introducing the Cortex A73 which is our most efficient big arm CPU to date. It delivers over 30% improvement in performance in a constrained mode of budget. I'll get into that a little bit. It doesn't seem like a big amount, but it actually a lot. Of also to keep this going, we need to make it more energy efficient. So we think about it as 30% more energy efficiency to be able to deliver that kind of performance in the same budget as well. And in the target 10 nanometer technology, this premium CPU fully with level one caches, single CPU with a neon and separate will be less than 0.65 square millimeters. That's just a little over half the square millimeter of done. Now putting into context what this means in terms of actual delivered performance. We're trying to compare the performance of the last two or three generations. The Cortex A72, which is in the main tape, is still established as a leader in performance and efficiency. And in these graphs, what you're showing is the dark part of each of these bars is the amount of performance delivered in a 750 milliwatt envelope, which is what you need to be able to constantly sustain it. And to Rene's point, the interactive visual experiences are not going to have just bursty interaction. It needs a constant level of performance. So the A72 took a big leap over the A57 and 20 nanometer and then the Cortex A73 actually delivers another 30% boost over what the A72 could do. And if you just cap it over two years, that's nearly doubling the performance that you saw in 2015, all in a very sustained smartphone budget. So that's a pretty big jump from the last two years, but we've been doing it over the last four or five and that's why you're able to fit in to Rene's graph, the performance from 2009 to 2010, all the way up to here. So to give you an example, we've tried to show the level of performance up with that has gone on over the last few generations. We've purposely taken the comparisons against the Cortex A7 processor. One reason it's in some of the most popular smartphones, it's shipped over 2 billion units so far and a very competitive base level of performance. Compared to that Cortex A7 and 20 nanometer, you see where the Cortex A15 was about two times the performance, the A57 going a little over three and a half. The Cortex A73, each core will deliver about six times the performance off that Cortex A7. Now that's good, but it wouldn't be possible in the same budget if we weren't also reducing the power consumed for each task that it does. And what this tries to show you is the level of power savings doing the same tasks as the previous generation Cortex A7. And what you see in the blue line down there is that in 10 nanometer for the same task, it would actually consume less power than the Cortex A7 did in 28 nanometer. Now that seems okay, that comes from process technology, you expect that. But if you look at the purple line, you're still seeing a substantial drop even if you kept it in the same process technology. So in 28 nanometer, the Cortex A73 is still less than just 2X of what it was while delivering effectively what could be the success of the performance. So our focus is on not just delivering that performance uplift, it is also on delivering it efficiently so you can actually achieve these immersive use cases. Now it's not just about the CPU and the GPU, you actually need to put it together into any system, an SOC that actually works. So the first thing you need to tie together is that you have a coherent CPU, a coherent GPU, all with big, massive performance, need to have an optimized path to memory and an optimized path to communication between the two. And that's where the CCI 550 components come in. We announced that in the last year and it offers about 35, 38% improvement in latency going from the CPU or GPU to memory, but also more importantly, because of the coherency that being efficiently done, you save hundreds and millions of power because you're not growing off chip as much and you're efficiently communicating within a CPU and GPU that are interacting much more closely now. And you've got the base infrastructure in, now you put it all together. Our SOC components are designed with the CPU, GPU and other IP in mind so that when a partner takes it, they have been tested, you can start off at a much stronger footing and help to get a more efficient design to market quicker. And the infrastructure includes things like trust zone for security and now we've got more IP that will help us do more advanced vision, adaptive displays, assertive displays and more. And finally, underneath all that, we have the pop technology to help you achieve the key performance and power roles that you have to achieve your market position. So, where are we today with the signing of this? And we have 10 licensees in the Cortex A73-9 on the Mali G71. Obviously, not all of those are seen on this chart. And we're lucky to have some of the representatives from Isilic and we're honored to actually have them here. Also VD Tech and TSMC were all of the partners and recently worked together on demonstrating with Cortex A73 in a 10-nanometer test show together. So, it is getting real. In fact, we will see Silicon by any of your early next year devices certainly by 2017. And those are the devices that we believe are putting together your smartphone, which is the portal for your future use cases. So, Cortex A73 at least a 30% improvement in sustained performance delivered. Mali G71 at least a 50% improvement in performance delivered on the new Bifrost architecture. And we're focusing on the key technology trends that will give you the immersive experiences of virtual reality, triple-A gaming, and augmented reality. Thank you. Thank you, my friend, for introducing to us the latest CQ Cortex A73 and the premium IP Mali G71 graphics processor. I would like to thank Nanda for introducing our latest product and technology. And now we would like to proceed to the photo call and the ceremony today. We would now first like to invite Nade once again to come up. And also we would like to invite our ecosystem partners, and we are now going to invite our key partners to come up to the stage to introduce today's ceremony. First of all, welcome to the design and technology of the designer designer of the design technology of the VU Tech, Dr. Kaye Volozlo, former Senior Vice President of VU Tech, Dr. Lee Rui Jun, Director of Business Development, Division of TSMC. And here we have our key partners, and we are now moving on to the next stage, which is to introduce to the stage the first design technology of VU Tech, and also to introduce to the stage the second design technology of VU Tech, Development Division of TSMC, and Mr. Diao Yanqiu, the PGM of Turing VU High Silicon. So now, we would like to ask our guests to smile for our cameras and our journalists leading through partnership. We would like to invite the guests on the stage to smile for our camera reporters. Ma Chong, please give them a round of applause. Thank you, the guests on the stage. Please remain on stage for a few moments. We would like to invite you to join us for a toast. So now we would like to ask you to raise your glasses and again a big smile for our cameras. So raise your glasses once again. And also let us drift to a better connected future. Three, two, one, cheers. Thank you very much. So ladies and gentlemen, we will now open up the floor for Q&A. Please put your hand up if you have a question, and our staff will hand you a microphone. We have a question over here. Gentlemen in the black suit. I think we're all in the black suit, yes. Hi guys, Ian from Mantec. On one of your slides you showed respect to K-Numbers on the East Silicon. Is that assimilation or actually testing in houseways? Good question. So the actual projective numbers at the end are estimated, but based on actual assimilations that are connected in correlation with both RTL as well as silicon. Is, how accurate is the assimilation of the NG accurate or just the similar accurate? Sorry, what is that again? Are you doing them similar or are the G accurate assimilations? Not quite sure what you mean by that, but these are performance accurate numbers. I had a little question about your HGs support. I'm surprised to see an mention of HSA and the spec. If there's any integration of HSA or the speed server, your H20 and H20 are major partners. In terms of HSA, effectively we have, effectively it's kind of combined. So in terms of when the spec comes out to the house, this has a new state which spec you're writing for. I might have to defer that to Mr. Hutchinson over there. We should be able to, thank you. Thank you. Let's take another question in the back. Hello there, Leo, what are you doing today? You said that the G71 match is 20th in last, which 3G do you do? Yes, so this is based on the, I mean, Nvidia GeForce 9400 GPU alongside a Core i5. Thank you. Just a reminder to those of you who are asking questions, please identify your name and also the media on the internet, please mention your name and the media. Thank you. Hello, I'm the CEO of G71. I'm here to ask you a question. Because I've just heard that the products in the 2G market are all in the Snipe part. So we also know that the collaboration between the two companies has reached the 7G market. I don't know if there will be any expectations for the 7G market. This is the first question. The second question I would like to ask is that after the new GPU was released, what is the most recent GPU in the future? Is there any market that is expected in the next three years? The third question, thank you. My name is CY and I have two questions. So you talked about arts products making use of 10 nanometer technology that you're working with TSNC. To my knowledge, you used 7nm technology with TSNC previously. So will 7nm technology still be applied? That's the first question. The second question is, you talked about the new SGPU G71. So what are your expectations or projections for market development in the next three years? Okay, so I'm trying to take it step by step. So just wanted to clarify, we've been working with the lines of TSNC on 10nm for over two years now. So what we talked about with the test chip, as well as what is expected in production, 10nm going forward. In March, we talked about how we're working with TSNC partners to go into 7nm, which is obviously further out. We expect 7nm, will succeed 10nm, and we'll continue working with them for that. Now, the second question seems a bit more open-ended in terms of market development. I think the capabilities that we bring with the G71 opens up new use cases. They could still be done with phones, but you could see other devices coming off of that as well. So I don't think we have a clear projection of how that volume increases apart from the standard things that you can actually see from a lot of those publications on projections on VR-class devices. To one thing I might add onto that, too, is with the acquisition we made of apricot, it's very, very clear that we think that anything regarding the visual experience is extremely important for the mobile platform. So investing heavily in those areas is certainly an area we're looking forward to, and the apricot acquisition, I think, is a good indicator of that. Thank you. Yes? Hi, my name is Masaki Masahara, I'm from Ishigoshi, Japan. I have a question about the A73, my architecture point for you. So what is the main differences between A72 and A73? So could you give me a more detail? Sure, I'll try to keep it short, but in the GESC, we've actually gone slimmer on the Cortex A73. I've gone from actually a three-wide machine to a two-wide machine, but we focused on getting performance while getting more efficiency. So it has a much more improved memory system and a branch prediction capability, but it still uses all the big data path features that A72 had like Neon. So overall, we've gone slimmer, but delivered more performance while improving efficiency. Thank you. Thank you. I think I saw a hand go up to my right-hand side. It will skip from Ikea electronics. I have two questions. First question is very simple. GPU process generation is what? Is 10 nanometer high? GPU process may vary G71. The process technology is 7 nanometer, 10 nanometer is all right. So Mali G71 is independent of process technology. We expect premium devices, first we'll see it in 10 nanometer, yes, but it is equally capable of being implemented in 16 nanometer or in the future in the set. Next question is, I could find an important company name. Paracom. Paracom should be an important partner in the company. Why, lay down, you cannot disclose the company name. It's a good question. We're really only able to talk about companies that have agreed to be announced publicly at this time. Thank you. We'll take another question. Any other questions from the floor? Yes. I'm Hiromoto from Intercompute George. My question is about the CCI 550 Interconnect. You said it's a clean and coherent Interconnect. So does that mean a CPU core can smooth the GPU cache or CPU core just skip the end node? So it is fully coherent, so both can smooth each other? Thank you. Question in the back. Hello, I'm H.K. from Macro from South Korea and I want to ask you to Mr. Nanpane about VR and AR, about Mali. And my question is, do you have any talk on about any partnership about AR2? We are AR companies about the optimization of our performances or efficiencies about what to make VR and AR in performance. So there's no formal partnerships that you're talking about, but we collaborate very closely with our ecosystem and our ecosystems, our partners that actually build it as well as front-enders, software developers and developer community. So we are working with a lot of people in that community to make sure that we hold everything we deserve and also the tools we provide will make it more optimal. Any components about VR or AR companies? Constituent, okay? Well, maybe we can take this a bit offline because at least you could come up with. I'll do it online. There is a big list. All right, we'll take this online. Any other questions? All right, if not, let us thank once again Renee and also Nanpane. Thank you for joining us today. Thank you for joining us today. And of course, thank you to all of you for being here today. Thank you for joining us today.