 Do I just look at the camera because you're moving so yeah, I'm just trying to get nice angles with it. Please introduce yourself. Hi, I'm Suresh Thirashree. I'm the Vice President of Compute at Imagination. I look after imagination's compute products, which include CPUs, AI technology and our graphics compute GPU compute products and the associated software with it. What is the RAI technology you mentioned? Yes, so we can run accelerator neural network workloads on GPUs as well. And here at your booth, you're showing lots of hardware. We announced in late 2021 that we are developing RISC-5 CPUs. We launched our first product and in front of you, you have on an FPGA. It's an RTX-M 2200. It's a real-time embedded CPU. So it's running in this FPGA right there. It's running in this FPGA here. It's IDE, which is a Visual Studio code-based IDE, which is unique. And it's running a simple application here. So we have a customer who has already licensed it, the first CPU. So as and when they bring the chip out, that's when we will get our first silicon back. The real-time CPU, it's an 11-stage pipeline. It's a high-performance embedded CPU for real-time applications, nanometer, process node. So that's the kind of applications we are targeting are embedded, sort of helper cores, management, SOC management cores, power management cores for this implementation in a larger SOC. And anything which is to do with real-time deterministic behavior. So Imagination is famous for the GPU, TICMODE, around the booth. A lot of automotive partner demos. We have one with Telechips here, which is actually here you have a kind of, because it's hardware virtualized, we can run multiple OSes. Here you're seeing a Linux running alongside an Android. Is it an ARM core? The CPU in here would be ARM CPU. But it ships design, and then they use RVR. We are also partnering with other RISC-5 companies with our GPU. So here we have a RISC-5 board, which is a Star-5 CPU running within Imagination GPU. We are enabling the ecosystem. We are working with everyone in the ecosystem to make the RISC-5 more successful. Here, a T-head, Alibaba's T-head RISC-5 CPU with Imagination GPU, and it's running Android at the moment. We are actually the only GPU provider for RISC-5 and we are saying it's the GPU of choice for the RISC-5 architecture. All right, because you have the experience making the Android stuff to make it work on the RISC-5. No, so by default, our GPUs are OS or CPU agnostic. So it wasn't very heavy to port it to RISC-5. But of course, with our own internal RISC-5 development, we have all the in-house experience of developing these use cases. So this is a GPU company for almost three decades now, so you have real products out in the market for some time. Yeah, for some time. This is a single board computer that's available now. It's moving at a very, very fast pace. RISC-5 has disrupted the status quo. And we are seeing more and more practically everybody here in the RISC-5. For example, when you have a UI like this, which looks a little bit like automotive margins to this platform compared to previous platforms. It's not hard because GPU for graphics use cases are things like OpenGL, OpenGL ES. So we support those Kronos standards. So the APIs are all standard and it's easy to port. Nice. What do you show here? Compute use case, where we have a chip, a Renaissance R-car platform with the imagination GPU. Running alongside with our, here is a surround view use case, where the GPU is actually doing multiple functions. It's actually rendering the car here. It's stitching the neural network accelerator here. It's actually doing human detection as well on the scene. So it's all and these ones are showcasing you. Again, you can see reflections being reflected on from the camera on to the device to exaggerate. I think the gold color shows it very nicely. This is actually showing reflections. We can go right in the car and actually we are now showcasing the same sort of surround view in a cockpit environment. And all this will be running on one SS, but we have hooked it up with our neural network accelerator to give that a full surround view. This is an RZ, R-car, the R3 platform. All right. Anything else? Here are the booths you were talking about a lot. There's lots of discussions about the future. Yes, lots of discussions. Cool. All right. Thanks a lot. Thank you. I'll send you the video there. It was live. Nice to see you. Tonight, tomorrow, I'll get you there. So it was live. It was live streaming. I'll give you my card again. She wants something very cool. Otherwise, maybe I'll find on your website something about the latest. Whenever you have on the latest, press release. Maybe you can film me for a second. All right. See you here. Is it possible to bring it on the table out there? Oh, yeah, but you sent me an email. I did. Use yourself. I'm Steven Delant, I'm the operational manager of IMAG ICLink. Chip services. Anything from design to supply chain of A6, application-specific ICAs. So what's the latest technology you're showing here? We are able to provide customers with three nanometer technology, which is the latest technology from TSMC. We're a broker for TSMC, Value Chain Alliance partner. How do you make it work so small? IMAG is a research organization. They still live in Belgium and some other parts of Belgium in the world. Where we do research on new transistor concepts, new materials, new ways to make it work. So we're actually helping the families to make it work. I personally don't know to abstract the smallness of everything to do that, obviously. What are you showing here? We're on our booth now at Better World and it's hard to show things. A mask of a mask set of a TSMC chip. Several chips. You can see several chips on a mask. In this technology I think there's 36 masks that are used to create chips. So there's a mask before the wafer? So there's several masks that are used in a process where it's layer upon layer upon layer upon layer. That's creating the transistors and the interconnect between all those transistors. And somehow the mask is square and the wafer is round? In which part of the... The mask is repeated on the wafer. You can see that pattern here as well. So light is shown through the mask on the wafer. Where this photo is on the wafer and it's the mask. Why do you have a 6x5 here? What you see here is the reticle of the wafer. And for this chip it's optimizing the silicon area and the cost-hands. A very good mask on the 3nm tech. Millions and millions and millions of euros. It's a set of masks. In the 3nm it's in the order of 50-60 masks. That's a lot of money. Many millions. And somehow this mask has zero... It's a very important part of chip making. So yeah, there should not be any defects on the mask. Because otherwise every defect will translate into a wrong chip. How do your colleagues and yourself have the capacity to design something so complicated that just works? We're abstracting. It's by using abstraction. It's by not looking at transistors anymore but at logic gates. And not looking at logic gates anymore but at building blocks. That's the only way we as humans can grasp the millions and millions and millions of transistors that there are on one chip. Is it a game of software and the software just takes care of it? We wish. There's a lot of software. And the software takes care of a single transistor in a multi-billion transistor chip of course. There's a lot of software help. There's a lot of pre-built made blocks. But the human in that process is very important. We area. And that's a human thing. There are tools that are integrating AI as well in this process. They're pretty... but still the human is still the factor, right? I won't be able to zoom in on 3nm. I don't know. This is not 3nm. But the way they're designed fits the patient around the chip. Like put things in the right place so they don't... There's many things to take care of. The power mesh on the chip for example. All the right timing. And if that's not the case your chip will not work. Or it will not work intermittently, which is even worse. And when you do the right design the yield is mostly related to the process itself. And the layout of the chip. They work together. So you have to make a layout that is as well. At the end just before you're taping out, as we call it, go to production. Then we do a check. Which is a design rule check. Which... How do you define your role at IMEC? Are you very much working with the EDA industry? Or are you different... what is your... Well, obviously we are using a generation of new knowledge in the field. We are users of EDA. And we work with the whole ecosystem of the failure analysis companies etc. So what kind of discussions do you have here at the Invented World? What people come to ask you about? It's nice to see a lot of people. Future possible projects. New people or new customers that are coming with ideas to create new chips. Partners. A lot of themes like the Edge and IoT and low power microcontrollers and everything. And you have a role to play in all these different things. For us, it doesn't really matter which application it is. But obviously we see some applications being more sought after than others. There are these things that are... Because when you have three nanometers, there's so much space on your chip. And they just put the AI and stuff on the big part of the chip to take advantage of it. I don't think it's more like that. But the silicon real estate is defining your cost. How big the chip greatly defines your cost. So putting something on there just because you haven't seen it happen really. Alright, cool. Thanks a lot. Did you want to jump in? Playing cards. Was you doing that? No, he was not with iMac. If you made a video of Global then maybe... No, he just joined us one week and a half ago. Designer. Cool. That's also one thing that... I don't know, he was a blogger. And I'm very delighted to present you a couple of new products in terms of graphical user interfaces. So what are you showing here? In fact we have different products out there our customers are using. From consumer electronic appliances and industrial as well as medical. Alright. And so for example this one has iDent MX RT and then you make the PCB that makes it world-work. In fact known there. Hardware itself is done by our customers. We're providing the software, the GUI solution. In order to get some nice looking user interfaces and iGMIs running on such embedded systems. Alright. And nine. Which is one of the brand new microcontrollers from SD Microelectronics. And something which is very nice to create. 160 megahertz microcontroller. And just have a look what fancy facts you can do on the new Neocrom GPU. Available on the new U599. What is this? The Neocrom GPU is the successor of the previously added DMHD, the Chrome art. Which is a graphic accelerator built into the chip. That can offload the CPU while running on... And remember it's a Koopie and M7. But I'm a software guy so I'm not a hardware guy. It's a microcontroller. Absolutely. On the microcontroller. Correct. And then it makes all kinds of fancy. And while offloading the entire CPU. So, CPU can do other things. And are free. We're still there in order to get a nice responsive user experience for the customer. Nice. What are the stuff you're showing here? We're presenting our software vendor which provides software to different kind of customers and domains. And with our software stack embedded wizard we provide a very lean, versatile, scalable and fast solution for the entire embedded world. I see a lot of other ST solutions around here. Are there different kind of chips used? Yeah, sure. As I said before, they're the very famous H7 which is also microcontroller. But on the left we have the MP135 which is the new MPU coming from ST Microdotronics. Let's go check it out. So, it's the next gen. And we can achieve a very high frame rate by utilizing the ARM Neon Instruction Set. Which gives us a boost in terms of graphically rendering performance. What is the OS? There we are running Linux. Of course, on the microcontroller itself, we are running on Fiatus. Oh, the metal depends on the customers. How good is Yokto? How smooth does it run on these STM32s? To be honest, speaking, we just use what is there. So, it's not something our part. It's available. You also do a lot of different things here? Yeah. So, as I said before, we're independent software vendor. Hence, not only relying on ST, but also on other semiconductor manufacturers like NXP, RENASARS, MBIG. It's expressive. So, in fact, everywhere where you can find it this way is something where you can find an embedded result guy. What's one of the coolest new platforms you have here? Hold on, let me double check. Perhaps something for NXP. I don't know if it makes 8, but this is more or less cash car for NXP. Of course, that's the stack that's available since many, many months. This is the AZ-G2L, and this one is the HALA that we are using blurring and shading for presenting nice graphs on the display. Nice. How long have you been a wizard in embedded? It's been a while, and it's not the first generation that we are providing. In fact, it's already the fourth generation of your ITs providing to the market. Fourth generation of what? Of your iSoftware stack for the market. On the screen, we see our portfolio that we can do. Of course, we have the software stack, but your iDevelopment, and there we are showing customer cases in different domains, as well as the tooling itself, in order to accelerate customers to get their UI up and running on the target. Only all the time to update it all the time. Absolutely. They can update themselves later. Otherwise, we would get mad. No, we have very sophisticated stack out there. Great documentation. Of course, the port is there, but the main aspect is embedded. Because there are no third parties and no open source stacks used. Cool. All right. Thanks a lot. Thank you. How was the process? Please introduce yourself. Hi, my name is Johnny Slumpf, and I am a Senior Systems Engineer for Altia. Altia is an embedded tool chain expert. We're in over 100 million devices in the world. We run on all sorts of different systems, from the low end to the very high end, taking your graphics, whether that's 2D or 3D, of your choice and the display of your choice. Nice. What are some of the latest platforms you show here? So we have quite a few platforms. Over here, this is running on the Renaissance Rcar E3. So this is the Rcar E3. Is it powerful? Yes. A53 version? That's right. That's right. Yep, and running free... Charger. Yeah, up here we have a model of a car charger. It's running a real car charger inside. This is running on the latest TI satara chip. What are the other prototypes for now? These are prototypes, but running the full board inside. What is the advantage of having a touchscreen like that? What is it going to be able to do? Yeah, so you'll come in and change your charging. Of course, we're just showing off the UI that's presented to us in the most optimized fashion running on the hardware. Nice. This is our integrated cockpit for electric vehicles. It's running on a Qualcomm 8155P. This is actually run on a multi-visor. This is running an Android subsystem. Three windows here. This one's running 3D here. 3D, bring up different windows as well and rearrange them. That's a lot of graphics on one SoC. On one SoC. Yeah, the Qualcomm. And this is quite similar to the kind of running Altium. And this is a pretty big car to have such a big display. Some of them have Chinese displays. They do. The Cadillac earlier days from the very small displays and just kept growing and growing from between the cluster, the entire cluster becoming digital, adding in the multimedia section. The advantage here with the, we call it the SA815SP. Yeah. So it's powerful with one of the latest designs. That's right. And it maybe has baseband, including 5G and everything potentially. So when there's the next gen, you can port all your solutions to the next gen and just keep improving. Absolutely. And one of the advantages of Altia is that we are an operating system and processor and BSB agnostic. Behind you, it's a little bit busy right there. We tried to see if we could sneak in and film this kind of stuff. Sure. It's been busy here at Invented World? Yeah, very busy. You know, this is, I don't know, we've been in Invented World for maybe, I don't know, quite a few years now at least. It's been very busy. It's a great booth. We're fortunate enough to be near the entrance of hall 4 here. And having just a lot of... Right there, I see NXB. Yes. So on the end here, we have the NXB IMX 8M plus. So what's happening with this NXB here? Yeah. So all three of these N3D in real time. And then behind that, you'll find the SDM 32H7 and the SDM 30MP1 also running the same design. The SDM 32, this is the MP1 and it is actually running the human body rotation. Also, the option from Altea is you can just flip that over to run 2D, so simulated 3D on a lower end processor. On the other hand, for that processor... 2D stack but it looks... Yeah, but it appears to be 3D. Yeah, I mean, it appears to be 3D, right? Yeah, cloudware. Altea cloudware. In the modern days of distributed teams, it's often hard to get the hardware to those engineers and to ensure that the hardware up in a kind of a server farm in the cloud. We have, I think, seven different systems right now. And what that means is that you can... Design editor, that's our GUI creation tool. Drag in your graphics 2D or 3D and instead of deploying that to hardware that's sitting on your desk, you deploy it till the... Altea ploys it to the hardware up in the cloud. In your browser, it will bring up a video display, which is, you know, video capture of the screen. But in addition to that, it has a lot of versatility and very easily can switch between different platforms and get up and running literally on hardware in minutes instead of weeks, months, or even years in some cases. Nice. So it's how you work. Exactly right. Yeah, Altea has been in this industry for almost 30 years now. This is one of our latest innovations, Altea cloudware. And it also has 3.5-inch form factors. What is 3.5-inch? It's very compact. Yeah, it's very compact. And we have the different kind of skill and we follow by Intel average generations to make our motherboard. What do you mean by world leader? You ship the most? Yeah, we ship the most and we have the customer around the world. All right. What is the latest designs here? The most interesting you're showing. The latest generation is the Raptor Lake, which supports DDR5 with PCIe Gen5. And what's the target here at the embedded world? Many people use this kind of boards in the embedded world? Yeah, and lots of people use the IPC industry. So should we go this way or that way? Push on more. There. The system, devil. And we make it with... Compact size. And with wide temperature, wide voltage. So here and there goes an Intel 13th generation with all the RAM and everything. Yes, and we have a bridge IO and also have an expansion slot here. So customers can put different kind of the graphic card inside a different slot. Add the graphics cards right in there. And where is this going to go? Where does the application? Mostly using the automation. Automation. Like in the factory. Yeah. Maybe. Or AI-AOI application. And here you have smaller? Yeah, we also have the compact size for the smaller one. And also have the entry level for the controller. Boards and everything. Yeah, and this one is the entry level for alcohol. And it says Ubuntu certified. All of it? Yeah. Yeah. All right, so you work with the Linux, the Windows, the everything. Yeah. All right. And there's even these little boxes here. Here's. So mostly used, we have the Fender size and also have the Fender size for Intel and AMD solution. Do you work on the ARM chips? Do you work with any of these Qualcomm, NXP? We mostly it's 86. All of it. AMD also? Yes. Yes, we also some of the motherboards AMD but... Not here. Yeah. All right. Cool. Thanks a lot. Thank you. For you, this is one product that we've contributed to. It's a MediaTek based Chromebook. And we've done quite a support for various IP blocks that are found in ARM devices. Is it one of the latest a companion or one of these two sets? I think so, yeah. One of these latest MediaTek? Yes, probably. Just a regular Chromebook and you can run... This is a regular Chromebook. What kind of... What is Linux? You can do the Linux. It's ChromeOS, right? Just ChromeOS. This is ChromeOS. This is off the shelf. We just went to a store and bought it. But we've been working on it. Because the ones we work on obviously we cannot bring. Can you explain what is a part of that that you work on? So we've done work on, for example, the codec hardware but also other bits of hardware that we've upstreamed to the Linux kernel. And also we work on with Google on having these devices in the kernel CI project, which is a project to do continuous integration on the Linux kernel itself. So upstream running whenever someone submits a patch to the upstream Linux kernel, we build it. We run it on hundreds of devices. And we report to the maintainers what works and what doesn't. Especially what doesn't. And all this work is useful for when people want to run Linux in these two, right? Yeah, yeah. Not just the ChromeOS. Not just ChromeOS, right? Because a lot of what we do is we bring the drivers to upstream Linux, so you don't need a special kernel. You can just use a standard Linux kernel, which is useful for someone like Google, because they have 8-year support on these things, right? So they want to upgrade the kernel. They just don't want to ship an 8-year-old kernel. What's happening behind you there on the screen? So this is our lava lab. So what we're showing is that we're building different... We're showing all these boards. Yes. So we have continuous integration across all these boards. So the developer writes some code. And then our system will build it for all of these boards, create a system image, boot the board with the image, and bring all of the results of all the tests that it runs into a UI where the developer can see. So it's easy. If you have a complex system that might run on many generations of boards, many different types of hardware, also the developer probably doesn't have 10 boards on his desk to actually have more confidence in what we built. The lava has been developed for a few years, right? And it's open and everybody can implement it the way they want, and you have a special way to get it to the next level? So we actually contribute to the open source projects directly. So it's not like we have our own special flavor. We're actually part of the community. But one of the services we have is that we host a lava lab where we can take people's devices, put them there, and then have both kernel CI and Mesa CI. Yeah. All right. What else can we film here to boost? So the next demo, I will... My coworker here, Marcus, will show it. Yes, it's an AI demo. What's the latest happening with panfrost? What are you doing here? What we show here is basically a machine learning demo that runs on a complete open source deck, like operation system. But the important part here is that we use like an open source graphics driver. And what this demo is about is basically a video compression focused on web video conferencing. So basically, at the beginning of the call, we take like one image of the sender and send it over to the receiving side. And then following up, we just extract key points for every image. And we use these key points with the first image that we extracted to reconstruct the face. And that way, we are actually able to reduce the bandwidth 10 times in comparison to HT64. Which makes sense because we just transfer the key points for every image instead of the whole video stream. Is this working or just a prototype? This is basically more like a prototype than a full-featured product. But it would be easy to just take it and integrate it into an existing video conferencing software. Because the last three years, I don't know what happened. There was something weird where people had to stay at home and everything and they were doing a good chance. Exactly. So there's been a lot of talk. I mean, there's been trillions of petabytes of video conferencing bandwidth out there. You can save 9% of it. This complete demo is already open source. You can just download it, flash it to your embedded port or you can see and just use it. And the next step would be to actually integrate it into an existing video conferencing software like Zoom or Microsoft Teams. You could potentially have rock chip powered 4K webcam and if the other guy also has it, or even if they don't or I don't know, the software has it, then suddenly they can see you 4K image with 10 times less bit rate. Yes, up to 10 times lower bandwidth. So you could do 4K video conferencing with like 1 megabit. Yes, you could. You probably need like a little bit more capable hardware because like this one is like very resource constrained but it would be possible to run like in 4K. I did a video with Nvidia once. We talked about some, all kinds of optimizations there. They were thinking of doing, I forgot what it's called, but I guess you like to do everything in the open source so everybody can use it. So the question is how soon are people going to use this in products or? We hope pretty soon, as I said, like the next step would be to actually integrate it but since this is open source, we can actually optimize every single bit. As I said, it runs like an open source graphics driver which actually allows us to optimize the whole machine learning pipeline to run in real time and also like support like 4K for example which is often not possible even on like Nvidia hardware like the Nvidia Jets. How good is the panfrost? Is it fully working? Everything is perfect? It's really, really good and works like out of the box. Really good out of the box? All the Mali GPUs? Not all the Mali GPUs, there are like more support than others but... The newest ones, they get great panfrost or is the oldest one that has problems? It's more like the older Jets are a little bit more supported and are more optimized, I would say. It's always like a combination of software and hardware, right? The latest one, I think they call them Immortalis, the GPU and it will take a while before panfrost is ready for that. Yes, it will take a while before panfrost is ready for this. So ARM hasn't open sourced the GPU driver yet? No. They don't want? I don't know if I can share too much about this. There are definitely like discussions in this direction. Awesome. What's happening behind you? What else we can talk about here? Maybe we can switch to them because... Do you want to give like a name? I really want to say that we have a lot of advice to offer. Alright. Let's see what's up in there. Oh, at first. I can ask them to try to go through the contracts and sort of take a closer look. Alright. Hi. Hello. Introduce yourself. My name is Laurent Panchard. I'm the CEO of 5Dison Board. We are a Linux consulting company specialized in multimedia and camera support. So what does the ideas do on board? On which board? Lots of boards actually. So our goal is to enable camera support for all of our customers. We make your cameras work for you. So anyone who needs to capture images to create a product, we enable that. We don't care about the end user applications. So we support customers in the automotive market, industrial markets, medical markets, anywhere where you need to support cameras, we can be there. So we work with SoC vendors. We are supporting both from an SoC from Rockchip, from NXP, from ST, TI, Exilings, all across the board. And we also work with OEMs who integrate those products and SoC in their own products to enable them to use them as efficiently as possible. So when there's an SoC and it has support for cameras and stuff, isn't it just there? Or what else do you need to do to make it great? Unfortunately, many things. Because usually vendors will try to BSP with a custom solution. They will have a camera stack that is vendor-specific. So for OEMs, switching from vendor to another one is very difficult. Or integrating just support for another camera sensor on the same SoC is very difficult. You need to go through a whole tuning process that involves closed source tools that involves usually working with the selected key partners from the vendors that can be a long and costly process. So we want to create and we are creating a completely open ecosystem with an open source camera stack that solves those problems. We focus on interoperability between different SoCs and different camera sensors. So we can make any camera sensor work on any platform. And we also work on creating tuning tools to enable our customers to do the tuning themselves at a cheaper cost and leveraging all the knowledge that they have in the house. Just behind, I saw Libre Computer. Are you working with these guys? Is that your partners and all that? So we are partners of Conabara. They are hosting us on their booth. And Libre Computer is indeed a company that Conabara is working with. We have a working camera solution on some of the Libre Computer boards. And we partner with lots of SBC vendors and it's nice when I see the AI just makes the video cortex amazing. Are you also a part of doing that? That's a Conabara project. That's a really cool one. That's actually a demo that you can see. We saw it right there. Yes, it's detecting the face. That's the same one. So you do have face detection and reconstructing the face and sending just the face over the network lowering the bandwidth. So they have created this AI demo and we have worked with them to replace the USB webcam that they were using with a row sensor. So now you can use the same demo on any kind of SOC with any kind of row sensor. Can you describe one of the really cool SOCs? They are great. But then the camera, the imaging chip is kind of like separate all the time, right? And you have to talk with the SOC and what's the language that's happening? There are SOCs that lack cameras you bought inside. So you have to use an external ISP that's usually a chip that will be between your row camera sensor and your SOC. That makes it more costly to integrate the full solution. But we do support some of them. We have worked, for instance, with OnSummit with some of the external SOCs and developing drivers for that. So those are solutions that we support as well. So depending on the type of SOC you use, whether it integrates an ISP or uses an external ISP, that is something that we can work with. How many of the cool ones have it integrated directly? I wouldn't say necessarily the cool ones. I mean, there are lots of interesting SOCs that definitely focus on cameras you bought. They may have lots of other interesting features but without an integrated ISP. But otherwise, yes, we're more interested in the ones that do integrate an ISP because that gives usually a more integrated user experience. Is this also stuff here? Yes. So those are samples of both the system that we support, just a small subset. What are you looking at here? So we do have the Debix board. It's based on an NXP iMix 8M Plus. That's the first SOC actually, the NXP product line that integrates an ISP. There was a bit of a game changer for them. But we have camera modules from Audi, Camokadoe, Raspberry Pi. This is a cool device actually. So this is a Raspberry Pi Zero, behind the hood, with a camera sensor from Raspberry Pi as well, the latest camera module. It's actually used as a USB webcam. So it runs Linux inside, full open-source support. In there. Yes, in there. That's the same one you have here. So that's the Raspberry Pi Zero inside. You run Linux on that, you run LiP camera on that, and you connect it through USB to your computer and it's recognized as a webcam. What's the best platform? Do you want to do a 4K video conferencing? What should I use? So if you go for higher resolutions and higher performance, we are looking at a moment at new SOCs, such as the ROC5 board, for instance, distributed by OKDU. That's based on a new ROC chip SOC that's very interesting from a kind of point of view. We're just starting developing on that, so that's something we fully support yet. But for the existing supported platform, the iMix 8M Plus is actually an interesting one. It's very versatile. I wouldn't say that's necessarily the best one for video conferencing as such, but for lots of industrial use cases, including higher resolutions, we can super a lot of applications on that. What's cool about the embedded world, when you support a platform very well, it's going to stay for decades. It's going to stay for years and years. It's going to be useful for potentially millions of products. Yes, it does. That's also something where I think we have an added value. If you look over here, this is a tablet running, or developed by Google, that's a Chrome-wise product. They offer 7 years support and software updates on all the Chrome-wise devices. This one is reaching end of line very soon. And because we have an open-source camera, it means that users can even extend their lifetime beyond that. So you can use the main line in iMix kernel, you can use a lead camera, you can use open-source software, and you have support even beyond what the vendor would provide. 7 years is already quite long in that kind of market, so that's quite too far from Chrome-wise. But you can even go longer than that because of the open-source ecosystem. Nice, that means these Chromebooks get a bunch of updates somehow? Yes, yes. 7 years from Google, and after that, depending on whether you use this one. Cool, that's awesome. That's a good book here, lots of action happening. Where are you based? I'm based in Finland personally, originally from Belgium, based in Helsinki. I'm here that collaborates a lot of French guys, they have a big office in Montreal, so certainly French speaking, but not just... In 90 years on board we have 7 people we're distributing over already 5 different countries. Caraba is much bigger, around 140 developers, so they definitely span the whole globe. And working in the open-source, is possible to make money? Absolutely, yes. We're selling not the art of work, we have lots of requests from customers. We do not sell our software indeed, but we sell consulting services, we work with SOC vendors to enable their products in this whole ecosystem, and we also work with OEMs who want to integrate them later in the product. Any kind of support they need, that's where we come. And all the camera stack, all the kind of drivers, that's indeed released as open-source, so everybody can benefit from that, but the other way around, we do benefit from the work of other developers, so that's also how we can be competitive. Is it possible that somebody will make really cool devices? I don't know what kind of sensor size is here, but could it have like a 1-inch sensor, micro-forters, ABSC, they just have amazing video conference or something like that? Yes, absolutely. You could go for really, really large sensors. Obviously, if we're looking at extremes there, sensors that have resolutions, I don't know, 400 megapixels, for instance, you will need custom hardware to go with that, because it's such a niche use case that you want to find out of the shell of SFC that will offer all the features that you need. But apart from that, you can go high-resolution, you can go, this one is actually large, but you can certainly go larger than that. And you support it? Yes, we can definitely support it. You can work with anything people want? Yes. The biggest problem we have, actually, when it comes to standardization, is connecting the sensor to the hardware. There's no standardization when it comes to this connection. All those ribbon cables that you see, they have different pinouts, different number of pins, and so that limits our customers when they want to connect a camera that's not meant to be connected to a phone. So there's always some adaptation that you have to do. So having standardization, that feel would be great. That's hindering development in a little bit. But from a software point of view, we can really support anything. As long as we can establish a relationship with the vendors and get documentation to be able to develop drivers, we can certainly support them. All right. Thanks a lot. You're welcome. Thank you. So what's the latest? The latest. What's happening? Let's start again, because I didn't hear that last bit. I don't think so. Maybe we can have Tyler. No, Tyler's busy. We have to show the demo. What about... What if Tyler... No, Tyler's busy. No, no, no. Tyler's busy. Tyler is busy. He's got two major customers waiting for him. No choice. OK, let's try. Hello. Shellbox, how are you doing today? Hi. So what's the latest you're talking about here, the embedded world? Embedded world this year has really been about the growth of Linux and security. We have seen a huge change in the last two or three years from where security was just talked about to where it's starting to be implemented in real products. That's been driven by OEM's need to produce secure devices that can be updated for their customers. This is mainly in the industrial, robotics, automation and infrastructure space, but at some stage it will meet the consumer space as well. We're seeing huge demand of MCU customers moving from their current application layer into an MPU space where they'll need a heavier operating system like Linux. And so what Foundry delivers is the easy button for Yocto, easy button for security which allows them to go and productize their devices six to nine months quicker with much more security. And you've been executing and this thing has been happening the last couple of years? We started the company five or six years ago. We were a spin-out from Leonardo. We saw the need for a horizontal-based Linux in IoT. The same way as we had a standard platform in the PC industry, the same way as we had a standard platform in cloud. Even when Android came out, that standardized where the mobile phone was going. Now we have something equivalent in IoT, but it's not just usable on one or two devices, but a whole wave of major SOC companies that work with us to have a standard Linux across an industry platform through ARM and X86. And behind you, you're illustrating some of the stuff here that you're talking about? Yes, so if you look at the illustration and we start over this side, most OEMs start here with selecting their hardware and they'll go into their EVK and BSP through their development cycle. They'll then go, OK, is that right? They could come back again and work through the BSP of the hardware vendor. That's quite simple to do. You can do tooling with that. You can do some proof of concepts with that. The problem comes down when you want to go into your own hardware and you have security in there and you have to build in the tooling and you have to build in the testing and you start going around this cycle. Security is one of the last things to be thought about and one of the hardest things to put in. What we do at Foundry is enable this whole cycle to be shortened because you're using the same code base from day one. You can choose multiple different chipset vendors. You have automatically built in the security. You have all of the Yocto layers that you can then customize to what you want. And as you go through the development and then into prototyping it's the same software layer. And then when you go into production it's the same code base again. You've used automatically what you start off with, what you edit, what you develop, what you produce, it's the same code base. You don't have to go through multiple iterations and proof of concept. And then when you go into maintenance you have a full CI, CD, DevOps cloud giving you daily CVE updates with the full software stack from Bootloader to Docker container on top that allows your device to be over the air updated with the most secure over the air update possible for IoT all the way through the device lifecycle for 10, 15, 20 years. It seems like all encompassing covers everything and you manage to do that. You manage to write a platform that just writes everything. We are probably the only platform that goes from Bootloader through to Docker container through to OTI that allows OEMs to be very flexible with their SOC to be very flexible with their cloud. We don't tell people where to put their data. We are cloud agnostic. We are SOC independent but we add all the latest security from all the latest vendors. So I've been going to the embedded world since 2011 and so all the embedded world has been having troubles in the security and you're bringing that to the next level. We're making Yocto and security easier. Security is never going to be easy. It has to be multi-layered. It has to be improved every day, every week. That's why we do maintenance and OTA to make sure all the latest CVs fully tested on your hardware every time you need it so you can update your product in the field. This is almost unique in the industry. Potentially IoT devices will be updated forever. Potentially. When they sell really cool products, thanks to your technology they'll be just forever. You can have updates as long as we keep the DevOps platform going, as long as you buy us we will keep updating it because it's automatically generating for you. Nice. Here you have some demos at the booth. Let me introduce my partner, Raul. Raul, do you want to do a couple of demos? Sure. This year at the MetaDeword we are not bringing devices, but we are using KeyMood because we have a large number of platforms we support. We decided to use KeyMood and that's enough to illustrate what we have. So what do we see here? This is some of the platforms we support. I will play the video so it will switch to another view where we have more platforms. I see NXP I see Arduino NVIDIA We have a bunch of devices we already support out of the box and the customer can always add their own device to the factory. Just pick one and it's fine. You can get started with us. We work with the Toradex all the different boards. We work with Toradex ST, TI, NXP a bunch of NXP as well. Our demo this year is basically we have three stations at the booth. Each station has their own KeyMood and we can see the factory which is a very good view on the CI CD that we provided to the customer. We can see all the builds we have done for that factory. The platform and the container builds here in that view and then we can also see the device is connected to that factory. You have this three station with three devices all the device following the same tag in the latest version and also we can check which application is running on each device. Well if we go back here to the device one of the cool thing we are demonstrating here is this is our device it's running five different containers and it's cool that we have a container running Flutter and another container running QT this is to show the beauty of container and the ability to isolate one from each other so this is the Flutter application running one in one container and then as we I will switch to another application this is our cog browser showing our website switching again I have QT so this is all running the same device it could be done on iMac on TI or any device and you're switching I'm switching containers in the same device just to show the ability of having multiple containers running the same device Is that something many people want to do as brand new idea or Well we have customers asking for how they can have for example a browser and a Flutter in the same device and things like that and this is where container comes up and help a lot and also most of our customers using containers for developing as well because it gives you the ability to isolate from the base operating system and then you can keep updating your base operating system moving forward and get the latest version Nice, what else are you talking about here at the booth? Well, let me see The video I was showing before show all the 30 minutes getting starter you have so if you come by by the booth we can give you in 4 minutes how is the first half an hour playing with factory in 4 minutes like an accelerated video so basically we are showing this and all the problems we have that Ian already covered today Alright, and the 3 stations basically the same so it's more station for so we can help more customers and kind of discussions you have here the embedded world, people come and are very interested to get people who are working with Yocto and they always need like over there updates and security they come by and say hi and try to understand what we offer and basically most of the conversations around Yocto, around security around container as well and that's more our focus this is why Cool, alright, thanks a lot Thank you very much Toradex booth, my name is Daniel Lang and I'm the Chief Marketing Officer for Toradex You have lots of stuff Lots of stuff, yeah so he's tracking your face so it's watching me our system on module with the Verdi, IMX8 and plus which has a neural network processor and we integrated that with a very nice kit including a camera either from econ system or vision component we have a very modular backplane where you can have power over ethernet RS485 and so on and the whole thing will be, it's in a case so you can actually deploy it in a real field so you can put it on your factory floor on a tractor and so on and that's there in the head of our friend here Do you want to have networking with this kind of stuff? Yes, so really nowadays everything is connected you want to get data for the processing all the AI is on the device but then maybe you want to you want to know which products are good if there is a criminal around or something like that so the metadata you get off but the processing is on the device and it's actually bored and even the droid is from our partner Ozone we have a software feature which lets you optimize your neural networks which you have maybe in TensorFlow or PyTorch it lets you optimize for our hardware so it runs very efficiently on this module and depending on the application we can do up to 300 frames per second so yeah, maybe we can have a look a little bit closer to the module itself so that's just the card so let's walk here on the way we have some demos from our customer so that's from emphasis and that's a pollen analyzer they use micro fluid technology to analyze pollen and with that you can have more productive trees and fields so it helps growing more food without using a lot of chemicals and so on and they're still a startup and they're very innovative and they use our hardware but they also use our Toryzen platform which I can show you a little bit later but Toryzen basically allows you to have a very fast innovation loop so you can develop something, you have tools for development and then you can deploy it in a very secure field out to the fields to these devices and they're connected and they're found every couple months they found new ways how they can use their technology and they can update their devices make them more capable over time so very similar and a lot of Tesla is doing with their cars, they're doing that here on a scientific instrument so you're all about enabling very innovative stuff yes, we really enable this company we have a lot of PhD in that technology to build a product because you still need embedded computing you need Linux, you need to do security update you need a UI and all of that and we really help with that so emphasis can focus on their core competition and then here also a device we do a lot automation, it's a small robot we do a lot of medical things and here you can also actually see what's in there so there's our module, an APALIS module, a display and then of course it also controls the system and then you have a touchscreen a very nice here, it's running so I don't know, I probably shouldn't interrupt it maybe I can try and see what happens pause, you know, continue, so you have a UI and there's also my also very innovative and also even this one you need to do updates they use more offline updates at the moment but also here we have a very secure solution so you're not corrupting the device, you don't break the device because that's a no-go in online then let's go to our over here here you see an overview of our modules and let me try to pull in our hardware product manager so that's Simon, our hardware product manager and he can tell you a little bit more about our modules so this is our portfolio of system modules you see three product families here you see our mid-range and high-performance APALIS system modules then moving a little bit to the right side you also see the Colibris, these are traditional system modules from the entry level to a medium performance and on the right side there is the Verdin family our latest addition with our brand new Verdin AM62 module at the bottom this is something we are introducing right here, launching it an embedded word and it offers a very competitive entry point into the Verdin family of sorts competitive entry point does that mean good price, what does it mean? it means basically the lowest total cost of ownership so it means if you purchase our hardware, integrate into your product, deliver it and maintain it over the lifetime you are going to end up with the lowest total cost, so it's not specifically the lowest price of hardware that's an important differentiation there's also, when I look here it's an A72, it's a big chip it's powerful, you do everything from powerful to smaller yeah, so scalability is an important concept for us we try to lay out the portfolio in a way that depending on your needs you can go from the entry level and scale up to the high performance and we provide a nice coverage for all of your needs there basically if people want just one, they can go ahead it's possible to all kinds of quantities how does it work to start working with you? yeah, so we are trying to facilitate it from the beginning we provide extensive technical support with all the materials we have products focusing on facilitating evaluation so you can order evaluation products that make it easy for you to wire up your hardware experiment with the hardware if you feel you are more working on a different level, then pick a different product which is better suited for that in terms of availability all of our products are long-term available, 10 years 15 years, we have that commitment publicly documented on our websites and this is one of the key values customers are keeping coming back to Toravex so if people want this, what would be the package maybe they would get? so this is a song on its own, the package the purpose to take would be one development intended carrier board like our traditional bird in development board which is visible elsewhere we would offer some accessories there we would offer a starting package, maybe we would select one configuration out of the many depending on the concrete needs of the customers we would be discussing before ordering with our team cool, thanks a lot so maybe I can also pick my favorite here, so that's Mellow, that's one of the carrier boards for the bird in module, you can plug it in and that board is good for prototypes but it's also great for low volume manufacturing so most of our customers they design their own carrier board but if you only have a few hundreds you maybe just want to use it directly something like that and you can see a lot of common interfaces extension connectors it's very compact so it's kind of and it's brand new so I like it a lot and you can get all the resources so you can also use that as a starting point and then modify it for your needs, if you want to have other connector or remove some of the connectors so basically if you want to go from there and just do something very specific they can start there but then they can ask you how do I be more specific and optimize for that exactly, you can see here we have a lot of different boards and all of them are fully open so you can just download the Altium design files and go from there and you can do that yourself or you also have a partner network which can help you so if you said oh I need a smaller or a bigger board or I need different interfaces I need two ethernet and so on it's very very easy to do that and yeah so maybe have a look at some of our other demos here this is cool, this is a pretty cool one this is for visually impaired people and that's actually a sweetie, it's an intel real sense and then our product is in here and I can already feel it it vibrates so it actually detects the environment and then it gives you here on this microphone, it's not a real microphone it's going on your cheekbones and it vibrates and it gives the visually impaired a sense of the environment so you don't run into stuff what's the resolution? they can see walls, they can see obstacles yes, so I by sound, by vibration exactly it vibrates and if you try it out it's very weird so you need a little bit training and you know I can see so for me it's hard to use but you can do that and it's from our customer here is this like your mass production? yes, you can go on their website people use it and it's changing lives yes, it really helps visually impaired to navigate better it would be awesome if eventually blind people can see yeah, I think here is more like not running into things but I'm not a specialist on this one yet another nice demo here is from Practica it's an orphan and it's super fast so you can make cross sauce pizza, I don't think anything is running and I'm also probably not the right guy to demo it but you can have recipes and do that microwave and heat and everything and it's also fully connected and this is used in some really big sandwich chains and coffee chains and some really big customers worldwide and this is our Verizon platform with our modules so here you can actually see their control unit where you can set that it uses Verizon so it's connected so they can do over-the-air updates they can improve it they can monitor it what's going on how many times it's used and they need maintenance so they can do predictive maintenance on it and so on and then here you see our module integrated and that's a good example of a carrier board nice it looks awesome so there's special open technology and they want to do cutting edge UI for it and stuff not just the UI, the control of the whole orphan we also have coffee machines and things like that here you can see our display ecosystem you want to talk short it's on video it's another hardware product manager and maybe you can really talk a little bit about our display ecosystem so you can get a picture with me as well so we're working on making the SI displays easier for our industrial partners so we're partnering with multiple display providers today we're showing off a Reverdy displays where we are able to guarantee availability of these displays for 5 plus years our ecosystem is made out of the standardized spin out that you can see here with the display adapter that you can add on the mezzanine or that we integrated directly in our newest volume carrier board for Verden the Mallow so we work on getting that ecosystem open running and available for our customers during this year with an availability of more than 5 years so that they can work with DSI as a native interface and as the primary display interface of the Verden family because when you have displays you have to run them you have to run the touch maybe all the different things and it's a way to integrate it so what we do really we accelerate your time to market to partner up with this display so they work basically plug and play so they're easy to integrate and then of course you also work with crank queue, slint, flutter and all these UI frameworks so you can do the UI very very simple so maybe have a look a little bit on our software one of the main differentiators from Toradex to other summands is our software here you see how we try to emulate a typical developer desk maybe a home office and here you can see how we integrate our software how we make it fast so we are integrated with official studio code so it's a very popular very productive modern environment but I think through to finish it up can talk a little bit more about the highlights so through what can we see here all right great yeah so what we're looking at here this is the developer desk this is kind of the central view of what a software developer using the Torizon system would look like so this is just my Ubuntu laptop with all the tools that are needed to develop applications to develop the custom operating system and just generally to work with the Torizon system as a software developer so the first thing that we have most developers is we're going to create a new project so I'll just bring this window up these are all the project template types that we support out of the box so if you need to create a python app or a QT app we have templates in place to allow you to quickly get up and running with any of these project types so you don't have to be an expert on containers or on Linux or on Yachto to be able to start deploying your application quickly onto the device so let's jump over here I've got a basic python hello world app that I can bring up and so as a python developer all I care about is my high quality python source code I don't need to know the complexity like I say of containers or anything I come in here and I start writing my code I'm ready to debug I hit the F5 button it's going to copy the device over it's going to hide all that complexity from me it's going to package it as a container it's going to deploy it to the board allow me to do source level debug just as if it were code running here natively on my desktop once I'm then happy with it I commit it to my local Git repo I push it to the cloud and GitHub we have reference implementations with our CICD system or excuse me with the GitHub CICD system GitHub workflows so that when I push new changes it can automatically be configured to a cloud based build to check out and deploy say to a QA group of devices QA does their work and then they push the changes up one level further the changes then get checked out and pushed up to the horizon platform for deployment to the entire production fleet you think most developers will use the visual studio or they can use other? there are probably I don't know if it's 50-50 there are some people that are used to Yachto used to a traditional command line maybe they're coming in with some code and container setups already that might not use this directly they might use it for inspiration but anybody coming in that's say coming from Windows CE or some other environment that doesn't know this environment that those are the people we are really targeting with this visual studio code extension and you support a whole bunch of platforms like people prefer that or this they can work yeah in fact those templates that we showed at the very beginning the idea is that they're easy to generate so if you have need for a specific application type that's not supported by our templates it's easy to create a new template for your specific framework or whatever it might be cool thanks a lot now we talked a lot about the developer experience so let's get a little bit of bigger picture overview and for that let's look at our wall here I take this here so here you can basically see the overview of our horizon platform so before we were actually at the developer location here and there you could see the visual studio integration and you can also see that we have our horizon core offering so that's the basic Linux, it's the Octobase, Linux fully open source and you can use that, pull that in and then for example you need a Qt or a Codesys or a Crank or AI and that's all provided as a container so it's very easy to integrate and then Astro showed here on the visual studio place that's integrated so you develop like you would develop on a desktop and why did we do that it's really to make it easy for people coming from Windows, Windows CE web applications and so on so they can easily move to an embedded device and can be very productive then we have a very secure way to deliver your software to your end devices and we use something, a framework a whole thing we call the horizon but the framework below is Optane and Optane is very secure, it's actually built for state level attack and it's commonly used in the automotive industry and then also we have a way back to get information like device monitoring, remote access and so on, we see that here later and then we have fleets, so most of the demos you see here the brewery and the agriculture and the beer and the coffee machine, they're all connected and maybe I can show you that here I can give you a very short overview, so that's our control panel, so for example before we were at the open so you can here see what the open is doing, how many recipes they processed, like the current, you know the CPU I think you can also see the pre-heating thing if the door is open and so on so you can use that for predictive maintenance, you can do remote objects, you can push OTAs here and so on That's all about security? Yeah, so there's a lot of security and we have our security expert here and maybe let me trust, try to steal it So John lead our Verizon platform so the the stuff in the cloud team and he's also our security expert, so maybe you can talk a little bit about the security philosophy behind the Verizon platform Alright, so I would say if there's something that defines the philosophy we take we try to do the hard thing first and then work to make it easier and especially when it comes to security around software updates, it is a topic that can sound simple when you first look at it but gets just devilishly complex and has a lot of implementation and if there's one thing that I would like people to know about software updates I think people usually start with an assumption that, hey, if I sign my updates, I'm fine if I worry about some transport security and I sign some updates I'm gonna be okay and that is an attitude that tends to work well until it really doesn't. The core of any software update system is actually about securely communicating repository state from your remote repository that can attest the validity of software you have to install to the device in a secure way that considers the timeliness of updates that considers the pertinence and compatibility if you have a multi ECU update scenario and actually dealing with all of those things and working with all of those edge cases is genuinely a devilishly difficult prospect so my background, I worked in the automotive industry for quite some time, working with safety critical and usually cars you can just click, you log into them, you can hack them in 5 seconds that used to be like that, right? I would say that that is something that very much woke up the automotive industry. The famous G-PAC in I want to say 2014 really marked the turning point and in fact that was a big part of the impetus for the Uptane project which was when the automotive industry essentially got together to figure out what should we do to worry about update security in our vehicles and they contacted folks at the Secure Systems Lab at NYU Tandon the SSRI Institute in Michigan and developed an extension to the update framework which is a cloud-dative computing foundation and Linux foundation project that is kind of the gold standard for software update repositories and how you would extend that automotive case where you need to have central direction and control of updates need to update secondaries and so the development of that standard started in 2017 I worked pretty closely on it Toradex ended up approving me because they were interested in that system and our update system at Toradex is now based on the Uptane standard so we deliver updates signed with two independent chains of metadata one that attests software validity and timeliness and consistency and all of these guarantees and another independent metadata chain that attests the actual installation instructions and you can probably tell I'm pretty passionate about software update security so the Torizon has the full security the secure updates forever when you have all these cool devices potentially they could be updated forever stay safe and it just works yeah so this is the thing that I mentioned we try to do the hard thing first and then make it easy so when you get to the really hard questions about repository state it turns out that practical research on flaws in software update systems often what it comes down to is flaws in key management and another common mistake that you see in update systems in the wild is thinking that okay I can I have a PKI management system I can manage signing software updates with my existing X509 asserts I can do it with my centralized key management system there are some special needs of software update systems that demand some difference in how you deal with them so one of the things that's built in from the start is the ability to rotate keys and sign changes in the trust relationship of the repository which is a long way of saying when you start up and you sign up for a Torizon platform account you connect your devices to it you can see here we have some actual devices connected those devices the root of trust is actually managed by us it's online you can click to update a package you can click upload a new software binary use our tools whatever however you don't want to have a single point of failure in a software update system what if you have a developer who left credentials to this system on the sticky note what you can do with these devices in production is take all of your signing keys offline so that when you are ready to do production releases of software you can have a prod repository that has an extremely limited set of software that there's only these specific versions which get secure attestations from the device exactly what software is running only these ones are authorized and it doesn't matter even our developers even if you had complete access to our systems as long as you have those key keys those important signing keys offline well you are fundamentally safe from the pitfall that people get into with signing offline and I think that a lot of people who start rolling their own update system with RAUG or SW update end up with troubles with the signing keys it gets compromised somebody loses it what is your strategy for rotating it effectively on the device and our software update client handles that automatically transparent which is what lets you start with an online key and then roll it offline without pain that sounds awesome so hopefully maybe you have 100% security support eventually it's difficult you can see we are really thinking a lot about it we are spending a lot of time we choose the hard way a little bit to make it easy for our customer and we really a lot of people would say is it overkill for what you do but we are not doing consumer grade IOT we really do critical device we do medical devices we do industrial so if something happens the damage is huge that's the reason we go that way we really care about the thing is we know all about that and all about the thing to make it easy for our customers so they do not need to have all that pain for that we have people here who love the pain to go through that it keeps improving this is the thing a lot of our customers choose to update their system but also on our side we need to keep an eye on everything and update it maybe one last we talk a lot about security which is extremely important but the other thing our customer really care about is reliability and for that really brave we are actually very confident about reliability and we just switch the demo here but you can actually come here so if you for the next show we have here a demo we say like try to break it so you can actually try to break our over the air update system you can pull the power you can have corrupted files you can play around and you can win a usb power bank if you do that so now we are actually just switching our demo to our partner from lightberry QNX so we are also working on offering on our system so I don't know if you want to short talk a little bit about what you did with Toradex and how you support our models so today we have planned a short demo for QNX running on the Toradex Apollo spot and we will just show how you can get QNX up and running on the Toradex spot and show a few of our demos and we are also planning options with our IDE have anybody been able to break it? so that's the order that's still about Verizon so Verizon is breaking Bruno did anybody break it? no not yet hopefully someone can break it so we find a problem we can fix but no one yet are there good hackers at the embedded world? maybe walking around yeah good so I hope you could learn a little bit about our products and our presence at Verizon about our customer and you know we hope we can work with you very soon and you know see you again here I think you probably have a very fun job working with all these amazing interesting projects people have with the talk about PhDs they have all these science all this knowledge and they want to turn it into a product and you are there to help them that's really one of the coolest part of the job and you know we also look for new people so you can really work with very innovative companies there you know for example we have 9T Labs they do continuous carbon fiber 3D printing we are in satellite hyper cars super cars PhD doing new micro thing using ultrasound to split DNA you know a lot of wheelchairs we really helping people that's also very nice we do a lot of medical devices which really help people a lot of people getting older so we have a lot of connected devices helping them we help generate more food with less chemicals so it's really really cool to see what all these customers do we have 1000 customers worldwide so it's a very exciting field to work and the order quantity can be from very few to yeah so our products are used from a couple hundred to a couple ten thousand so our very high volume customer maybe have 70,000 devices a year or very low end customer maybe have 100 devices a year and we do everything between if you go very very high volume to do a design in so maybe you only have 10,000 pieces at the beginning and then you ramp up we can help you and of course also on the low end we have very good support we are very famous for our online resources so even your small companies very easy to get started so not only the universities in Switzerland but anywhere in the world if they have some amazing idea thinking how to get it done just start contacting them so we love to work with university a lot of them you know creating startups have smart people or they go into companies also for recruiting so if you are in a university and you need any hardware contact us you have special you can get free hardware discounted hardware we have programs we collaborate on studies you know master's thesis and so on so we love to work with universities and how's it going in Lucille and your headquarters or the different offices it's more and more yes we are yes we hire we check our website we just went to a bigger office in Brazil in Campinas we have a very nice new office we are just going to move in Seattle to a bigger office the office in Switzerland is growing so it's going very well so very happy about that thanks a lot introduce yourself Hi my name is Robert Day I'm with the arms division like to talk a little bit about today is an initiative called Sophie which is really aimed at enabling the software defined vehicle revolution that's coming so what is a software defined vehicle so a software defined vehicle is essentially the features and the functions of the vehicle and often the differentiated features and functions will now be driven by software so it gives the opportunity to kind of design how a car is going to feel and how it's going to operate by software versus traditional hardware modules so it allows the automakers to basically decide how to keep their car updated through software what is Sophie so Sophie stands for scalable open architecture for the embedded edge so it's an initiative that we kicked off about 18 months ago and it's really looking at how you can develop efficiently developed software and then deploy that in a vehicle so what we did with Sophie is we looked at what other industries do agile development and deployment of large amounts of software and so we found that the cloud and cloud development was a really good starting place for us because people do that every day so you develop in the cloud and you basically deploy to a cloud instance that's what cloud development is all about in this case instead of deploying to a cloud instance we're going to deploy to a vehicle so we're using these cloud standards, cloud tooling but what we do is we actually look at how is a car different software and a car going to be different a couple of things are one it's real-time we need to worry about real-time stuff we need to worry about safety which you don't have to worry about in a cloud instance so what do we need to do to these standards or to these tools to make them work for a car so you'll often hear the term data center on wheels for the next generation of vehicles it's not a data center on wheels a compute might be similar to a data center but the software that's going to be running and what that software needs to do is not going to be the same as a cloud instance so that's why we came up with this initiative when we kicked it off we wanted to make it an industry initiative so we got a group of companies from both the automotive world but also from the software development world because this is a software development thing so the governing body members of SOFI are ourselves Caread VW Woven Planet Toyota Bosch, Continental from the automotive industry these are the companies that are going to be essentially adopting things that we put together as a SOFI architecture but we also brought in people who understood cloud and software so we also have Red Hat Suze and AWS as governing body members that are really kind of looking at what does this need to look like to make it work for software-defined vehicles so it's quite a long answer when you talk about AWS Red Hat, they do a lot of stuff in the cloud and now they all their skills they can put it towards cars they understand development in the cloud, they understand software development all we're saying is we need to look at it going into a vehicle than going into a cloud instance and so they have those skills for the large amount of code software code that's going to go into a vehicle that sort of development, that agile development there's a term we're using cloud CICD which is continuous integration continuous delivery because that's what they do they keep integrating and then they deliver our vehicle is going to be the same when you buy a car a software-defined vehicle that vehicle will get updated and when it gets updated it's going to get better so no longer when you drive your car off the dealership to get kind of worse over time it's going to get better, it's like our cell phones that we get new versions of operating systems new tools, new applications updates, upgrades and it's not just the UI of the infotainment system it's actually all the way down to the wheels it's everything the market is changing to things like electrification where people are just redesigning the vehicles to be EVs so you've got a clean slate you can really look at all of the you know, the functions the electronics in the vehicle you're almost starting from scratch which is great and so this initiative is really there to try and enable this and accelerate this and behind you are you showing a demo yeah so let me explain this so one of the things with Sophie is we wanted to have applications within Sophie to show the automotive developers how you would use Sophie to develop a particular application so this is running something called AutoWare which is an autonomous software stack we're actually running it on this system here and this is a system that we had a partner company adlink develop and it's called the Ava developer platform so it's basically got a powerful ARM processor in it it's a 32 core ARM processor because we didn't know what people would want to develop using this so we kind of over provisioned it so that you could run any software defined function including an autonomous stack that you're running here so what this is really trying to do is to show how you could do your development which CPU is it so it's an Ampere Ultra 32 which is 32 core and it's got a nice cooling it's got connected with GPUs yes so this is running GPUs as well so a lot of the functions in these high end automotive systems like autonomy they don't just use CPUs they use accelerators so this system is actually expandable it's got three PCIe so you can plug in graphics cards whatever you want networking so it's really quite a nice system to get started because let's say we don't know what people are going to develop on it we're hoping the whole car the different parts of the car and you know part of the SOFI architecture is also looking at how you can have different applications running on the same system so there's no reason why you couldn't have your ADAS system and your IVI system running on the same compute system if it was powerful enough what I'm looking at here is to do with self-driving and stuff like that so auto wear is an autonomous self-driving stack so this is just sort of showing a view of what auto wear sees obviously we're not connected to a car so we're showing an example of one of the parts of auto wear showing the point cloud and what it's seeing from the sensors but this is a kind of pre-recorded sensor stuff because we don't have a car here but at least it gives you an example of you know what we can run using the SOFI architecture and using this development platform it's quite amazing that the ARM processors people are developing for self-driving cars and electric cars and everything and the power they have is amazing it keeps getting better you can even swap boards out in your car to keep it updated so it's really interesting when we're going into this software defined world there are a couple of things so firstly what SOFI is trying to do is really take away the real dependence on a particular hardware system so you're right you could move from one SOC to another SOC whatever or if there's a newer version of the SOC you could do that the other interesting part is you could over like we did with this thing the automakers could over provision it's not what normally happens because normally you build it to what the car is doing but if you're going well I think I'm going to have extra functions I'm developing those as an automaker and I want to maybe upsell them to you know the consumers of my products put a bit more compute power in there and that would be an interesting shift in how automotive works yesterday I did video with they were talking about 128 core in the car like you could have the bigger it's cool so this one you could actually put 128 core in it if you wanted to so it's you know it's it's a com HPC module that we put in it so it's it's upgradeable even in this development system and here it's showing the this is the architecture of SOFI so there's a cloud part this is where you kind of do your development using the SOFI framework in the cloud but obviously this needs to get deployed into a vehicle so there's a sort of similar view in the vehicle so up here you basically start developing your code you're developing in the clouds you can even do your testing in the cloud you can do things like digital twins to kind of like a good representation even in the cloud of what your vehicle system looks like and then what we're doing is we're using some cloud technologies like orchestration which allows you to then deploy your different workloads which are shown in the kind of green boxes here down into the computer on the edge which is what this is sort of simulating right now and so another part that another technology or methodology is used in cloud is containers so containers again allow you to develop these workloads these applications in the green boxes there and then sort of seamlessly deploy them from the cloud down into the edge I guess it's working very fast this ecosystem people are asking for for this right now there's so many electric cars coming to the market right now many of them want to be self driving yeah so the different applications it's all going to be driven by software so you're right there's an appetite for it so the Sophie members we're now over 70 announced members so there's a lot of people in the automotive ecosystem that are really getting involved with this initiative and it's got some really good momentum and it's because there's kind of a need for it out there and there's a desire to start developing software now even if the vehicle is not going to be produced until whenever I want to get my software development going early and now I want to be able to continuously integrate and continuously deliver and here at the arm booth you have more different demos talking about this if we go around the other side this is an example this is not actually running Sophie this is one of our SOC partners so Telechips, I think this was announced today the Dolphin 5 SOC and so this is kind of cool because it's running different applications that you would see in the cockpit on a single SOC so we've got an example of running a movie we've got an example of surround view and surround view camera so there's a camera here you can see my hand so they've got cameras all around this to do the surround view and then you've got your cabin functionality over here so it's basically showing the power and the performance of you know this SOC which we can see here with our Cortex A76 high performance high performance, 8 nanometer this thing's rocking so you can see this display here this is running on this SOC it also has our Mali G78 which is a GPU and it also has our NPU in there as well so it's a really good example of the different types of IP that arm producers that can be used by our silicon partners so this again this is showing how you can run different applications on a single GPU platform a bit like we were talking about around the other demonstration this is just showing it in a very graphical way for a while I've been seeing such amazing arm processors but it's hard to see all the potential actually used in devices and cars a great way to do that you have so much stuff going on and you process all of it at the same time and if you think about a vehicle there are so many different systems now running in a vehicle I mean this is just showing what you sort of see showing good cameras around your car you know it's showing potential for sort of ADAS functionality display functionality and so yeah it's a really good way to see quite how performant these things are right now and also the sort of software applications that run on them Thanks a lot You're welcome Products from PyTec and from Pyramid You say Pyramid Is it not Pyramid? Yeah it's Pyramid if I pronounce it in English then it's Pyramid but we are a German located company and it is called Pyramid And right here is a big touch screen What are we looking at? We're looking at our first auto kiosk full auto kiosk with the touch display from PyTec our company you have a display which has 6mm glass so that it's really robust and you have a solution for ticketing for example It's a beautiful, it's very good display high quality the best touch manufacturing In Asia exactly we have a display with 2500 nits and as I told you a big glass so that it didn't get broken Even if young people are around it's okay to get beer and throw it again against it wouldn't break And then you implement a whole kind of smart stuff behind Exactly we have a scanner inside we have a printer inside we could also use a A4 printer for ticketing for railroad for example we have a payment system but we can do what the customer needs so we have for example a payment inside we could also use a different version Is this the VIP machine is it a bit expensive or that depends on what you think what is expensive we have a solution for outside so it's full outside, it's a little bit more expensive than the inside version because you have sunlight you have rain outside and we have to do it IP65 Alright, what are we seeing here behind you on the left side we see our new stripes this is for digital signage we have a special design we named it stripes 48.5 inches where you can put different content of advertising for example we had a customer from ice cream and he told us that he could do all his ice swords on it strawberry, pistachio hazelnut and so on this is also manufacturing by FITEC but here you have no touch display this is only for digital signage if you want to see a touch display a little bit bigger one, we have it on the left side this is our 55 inches touch display 4K 4K and you see in here our solutions the pyramid world here you can see where we are located in which business we are located for example burger stores we have the possibility to do digital signage at the beginning of the store with our A-frames also from FITEC and we have the solutions like our kiosks where you can order or where you can buy the burgers and so on this is a very nice 4K touch the highest quality viewing angles it's always optical bonded so it's a little bit more brightness and you don't have an air gap inside so it cannot be getting rubbish inside so what else do you show here at the booth? at the booth we also show our industrial PCs on known order this side we have the 19 soil racks like this one is the KAI version we have our 19 soil racks with front IO on this side so what people use this for? for servers for example safety relevant technology we are using this often for storage for example and machine vision industrial machines like CNC machines also on this side we have the back IO version and here we can see that we have also possibilities to make individualization for customers for example a logo for the customer or different kind of colors and what's happening here? this one is another FITEC display it's a full outdoor A-frame we call it this is also digital signage it's for customers who can put their content like commercial and get the customers into their shop so this is also in projects we have it with an acu system with battery and you can put it in walls for example and say go to this burger store and buy our burgers at night and you put it back exactly this is the idea of it you can roll it on a couple wheels exactly you can you can take it and put it outside at the beginning of the at the opening of the store and the last one you can see is here we have our small touch display from FITEC and in this time we have a 7 inch touch display with cover glass with 7 inch so it's robust and very small and the bigger one you can see here for example the 12 inch 0.1 version RME open frame where we can for example use the Siemens SPS with our panel PC what is this? Siemens SPS Steuerung in Germany so this is used for different kind of machines to use or to control the machines and our panel controls the Siemens SPS so people can put that in their factory or something? Exactly you can put it in your factory in your machines and you can control like we do here light on, light off only a small solution for what we can do nice and down there on this side we have our box systems different kind of industrial PCs and on this side we see our rubber frame so this is a rubber frame capacitive touch display for 10 inch and 50.6 inch our range in the displays go from 7 inch to 55 inch in normal way and we could do a bigger or smaller customer needs nice and this one looks very bright it is very bright we have here again optical bonded and the big thing here is you have a very small frame very very small touch and it's touch again we have always touch you can see here our solutions for example in the in the center in the data center where we have our touch systems and the last one that I would show to you is our rugged size our rugged display this one is with a 6 millimeter glass it's rugged so you can throw something again against and it's also if you see it it's also so closed that it is IP67 so there is no possibility for water or for dust to come inside you can go swimming with it and you can throw it oh no on the ground and nothing that did happen alright so this is a big next step of the future the Phytek Pyramid joint company right? yeah we are a joint company and we we grow together at the moment and it is very great because our industrial PCs and our displays grow together to our self-order kiosks for example so we can use both technologies for our self-order kiosks and you based in Germany? we are based in Germany in Freiburg and have different factories worldwide and Phytek is in in China? yes Phytek the factories in China but we have storage here in Germany and our sales guys and our headquarters in Germany so anybody who has big ideas with touch and everything you are ready? of course of course as I told you we are a manufacturer for what the customer needs and if someone came to us and say hey we need the transparent display somebody say to us we need the transparent display it is also touch you can see it here it is a touch display where you can see also behind the people and everything what is going on if somebody wants this we also can develop this nice that is called OLED touch monitor transparent 55 inch exactly 55 inch is only one solution we could do it smaller if the customer needs it you can do 100 inch I think at 84-85 it would be the ending and not for transparent for normal touch displays it would be the limit I think at the moment thanks a lot you are welcome my name is Anton Kuzmin and I am responsible for all the development efforts here at RIS embedded company and one of our most recent development in partnership is with the new renaissance chips so this is the RZ what do you call it the VL V2L V2L is not this particular chip but we have it on the other design those are RZ5 so it is a microcontroller chip embedded controller built around the RISC-5 architecture based on under semiconductor RISC-5 core and recently we introduced this sget standard OSM model built around this chip and a new development platform for this model which is a 5-berry it has this chip mounted on the Raspberry Pi form factor board so you can easily prototype all your designs with this chip and use it for development with the yes of course so it is actually running there this demo is running on the other renaissance chip a bigger one RZ V2L and it is a nice one yes it is dual core ARM CPU with their video processing unit and also artificial neural network accelerator chips and you see here the demo running the YOLO neural network with their image recognition identification so it runs the streaming video from the USB camera through their accelerator chip and image recognition neural network and then shows it on the display with bounding boxes and everything what is this this one is the same standard familiar as a OSM in the recent sget family but these two boards are for the different CPU it is also ARM architecture but a little bit bigger so with sget OSM standard we have a small size medium size model and also a bigger size models are available those are ARM it is from STMicro STM32MP1 but it is running on Linux for example yes and this one is running YOKTO distribution YOKTO Linux we do support different embedded Linux distribution on all our products and that is YOKTO which is an industry standard but for the easier start for somewhat lightweight distributions we also support a build route what is happening over there you have a kind of list all the stuff you do above you can see almost all the keywords or the services we are providing but here on the display we have a nice product line we have starting from different system models based on the different CPU architecture those are small ARM CPUs then we have NXP iMix 8M Mini then we have NXP iMix 6UL then all the product those are even a legacy but they are here to stress that we support for our industrial customers we support longevity lifetime 15 plus 15 years for many products for some of them we can see later on we have even 20 years of lifetime which one is 20 year old not 20 year old but with a guaranteed lifetime for 20 years those microchip smartfuge sorry not smartfusion polar fire system on chip FPGA and that's interesting that it has a combination of the FPGA matrix and hard CPU core quad core RISC5 architecture and all those models are running also RISC5 running on the FPGA no, RISC5 running on the hard processing core on the same chip with the FPGA alright that's fully deployed yes, absolutely supported and support for those models is available both from our meta layer for their and also upstream to the mainline Linux kernel and to their microchip so they are not only fully running the linux distribution but also all the changes upstream so we contribute our hardware support for our hardware to mainstream linux distribution your company is around the world we are located in Munich area but customers we have a lot quite a lot of customers in the States and you support everywhere all around the Europe Japan, some China and Indian customers recently Israel yeah, so all around the world I would say maybe South America is not that visible on our map but generally working everywhere and people who have projects work with you and you work together to get it done and we provide not only selling our models we are providing services to help our customers to design in our models and also if some customers have specific requirements or if they need some custom development we can develop a system for them based on their requirements and the technologies we own and also we are trying to participate in a partnership with both semiconductor renders as a renaissance we are a design partner of Intel, design partner microchip and we are trying to participate in promoting those products or in MIV summit it's microchip event we have webinars and trainings together with Arrow semiconductors and Leitone and FPGA Congress presents so quite active on promoting those projects and products as well just getting because many people are still somewhat scared of their FPGA technologies and especially in system on chip FPGAs because such projects involve their expertise or require expertise both from the hardware design side from the FPGA development and at the same time from the embedded software development and what is the strong point of our company is that we are capable to provide an integration of all those technologies we know all of them and we can help customers and willing to help customers and educate people FPGAs have a lot of potential you can do anything you want with an FPGA and they see a lot of development recent years and yes we are very enthusiastic about it especially when you want to have long term support you can potentially make things work longer and longer because you can update effectively you can update your hardware without updating your hardware you can provide a new functionality extended functionality and change it but what is good about them both now Intel and microchip supports that they do guarantee a lifetime of 15 or even 20 years for their projects and we believe that microchip or microCMI have not absoluted any single family so some of their chips are selling for more than 10-20 years already on market and that serves as a good promoting point especially for the industrial customers for the customers in the embedded area because the design cycles are relatively long it takes another 2 to 5 years to design in something and then you are not going to change your hardware platform for another 10-15 years you just have to use it and have to have a supply and have to manufacture your devices thanks a lot you are welcome if it is something suitable or you can make anything out of it we are going to start here but if you want to start with this stuff over there maybe we just start here to introduce and you say welcome maybe if you just do a little bit there just shake it a little bit my video is going to be in 4K so I will do it like this and I will try to ask you questions so is this like a live broadcast so right now it is live but I will record so I will publish it in 4K after let's start again it just can be cool when I do like this you can say hi my name is Frank Chen I am the CEO and CTO of Exocen so we want to welcome you to Embedded World 2023 and what do you do at the Embedded World we are a storage provider so we do full service custom storage for industrial aerospace, automotive and cinema so we are here to showcase our storage products and it is like hybrid reliability when you are in industrial products flash storage is a great way to do industrial so compared to traditional HEDs where they are kind of susceptible to vibration shock different air pressure SSDs flash they are made of completely made of ICs so they are not susceptible to any shock and vibration any thermal issues any altitude issues so this is where we build hybrid reliability storage so what are we looking at here so what we are looking at here is an industrial grade solid state drive this is our own custom design where we design all the hardware the firmware we do all the testing internally we do all of our manufacturing it supports up to 8 terabytes and very low power so compared to a traditional HED where you see like 3.5 inch HDD that's about 4 to 5 times the size of 2.5 inch we can store much more capacity it's got a much higher storage density compared to HDDs are there algorithms you can work on to make things more reliable and the data to be safe no matter what basically we own all of the firmware so we develop our hardware and we write all of our firmware for SSDs so that includes several important critical components like flash management data integrity and data security so we also implement RAID functionality on our SSDs so that we can recover data in a backbone transparently in case there are data corruption or something happens to the data what is the whole process of getting data out of a broken device or device that has issues to recover so like most of the vendors will tell you that you cannot recover data from a broken SSD which is quite incorrect because if you own all of the base technology from the hardware to the firmware then you should be able to recover the data so in an event something like this the data needs to be retrieved from this maybe something hardware there is some hardware issue with this drive we take apart the case and we'll actually connect the driver with our debugging tool from the debugging tool we'll be able to re-initialize the drive and initiate the drive recovery process and you sometimes include that as part of the service? yes we include data recovery as part of our differentiated service because we're able to go much more beyond file system level data recovery going to hardware level data recovery and firmware level and you have data recovery around the world? yes we support this by either remote recovery or you can we can have FBE outside to do the recovery or you can send in a device to us we can do a factory level data recovery and there's many different kinds of storage that you do huh? yes so for industrial we have various role factors this is a 2.5 inch which is optimized for thermal so that means you can operate without any thermal solutions fans or air flows to cool off the drive it will be able to run at full speed at 85C so these are M.2 20-80, 20-42 and 20-30 so they don't have a casing so additional thermal solution needs to be applied and we work with our customers to come up with the most optimal solution for thermal mitigation in addition we also have a thermal ATC algorithm which automatically adjusts performance to ensure that the drive operates at an optimum performance in any thermal condition because these kind of storage products can get pretty hot it's one of the things sometimes I touch the SSDs and stuff it's pretty hot and it's a challenge design things correctly so really it starts off from the component we don't sort commercial flash and build them as industrial grade so these all of our components are industrial grade so we buy industrial controllers from our SSD controller vendor we buy industrial grade flash and we also package our industrial grade flash because sometimes the amazing magic factories that kind of make SSD and all this stuff they also have patents in certain ways of doing things or how can use those or do differently customize the way the algorithms are designed so basically we have our own patents on how we manage our flash how we optimize our performance so we found over 70 patents worldwide on our SSD technology nice and you're tightly collaborating with the R&D of the next gen I guess because the way you design the memory you have to think about firmware early about how you want to design it right so we work very closely with our controller vendor which is Marvell on their next generation on their future roadmap controllers and we work very closely with the main flash vendors like Kyosha WD and Micron, Samsung and Hynix and do they sometimes collaborate with you on the end product like even they might use themselves as separately or you only keep your expertise only your brand we're open to collaboration but mostly we do our own and we collaborate with a lot of partners they're not necessarily flash vendors or controller vendors but we work with partners like NVIDIA, Mercedes we work with partners like R&D and R&D to help them build and help ourselves build ecosystems for storage people buy a self-driving car or EV or some cool products I'm not going to ask exactly what but they might be your storage in some of the cool products in the world yes some of the our storage is in some of the most unexpected places for high performance and high capacity applications so it could be somewhere flying in the air in space low orbit satellites airplanes underwater exploration oil and gas automotive, industrial, medical you name it we're pretty much in there so all these guys developing cutting edge things they know how to contact you and say I need this performance and you might be able to provide it yes and not only that but not only do we do cutting edge we also do trading edge because our goal is to serve our customers no matter what their needs are so as long as we're able to provide differentiated service whether it be performance, capacity, power or security or any other features that our customer want we're happy to implement for them and support them so the other storage products you have here this is a unique product that we have it's the PI4 series E1.S so it's the first of its kind it's an enterprise form factor storage but optimized for industrial industrial and automotive applications so for industrial it's minus 40 to 85 with encryption, TCGO and for automotive we actually added conformal coding and we added BJ underfill so that it's shock and vibration optimized and it's protected against elements like salt humidity and things like that so this could be in some amazing cars or airplanes or boats all kinds nice what next do you want to show so next we probably want to show you this is still it's durable this is our automotive grade PCIe 2.5 inch UDOT 2 so again like we said we have BJ underfill we have conformal coding this is up to 8TB and this runs at about half or one third of the typical data center SSD so the whole drive runs at 7.5W so it runs very cool it's optimized you know minus 40 to 85 wide range of temperature let's say by the best 2024 affordable self-driving EV why does it need to have so much storage what is an electric car for example doing with all this storage actually most production cars don't need that much storage they probably only need maybe up to 512 gigabytes of storage these are actually by now used by the car makers to log a lot of autonomous data 8S data logging because they cannot just upload all this it's too much for upload yes it's too much for upload to the cloud so they need high performance high capacity local cash to store that data that they can later take that and upload to the data center alright so it's special development for cars maybe it's not necessarily mass production for mass production we actually have another line of products they're called BG SSDs so they're BG PCI SSDs built for production level cars so they're smaller in capacity but also very high performance nice it could be like Formula 1 car that could be tested and while they're testing it and optimizing it they want to store sensor data on the car while they race around it'll be stored on these drives potentially well these are automotive grade SATA drives so they're a little bit slower than PCI but they also come with the same shock and vibration, environmental protection at a SATA 8TB so they're still very high performance but maybe not so high compared to PCI but they're also very low power so they're about one third of the power consumption of PCI so for those applications that doesn't require very high data rates but very high capacity maybe they want to store large amounts of data over a long period of time that would be an ideal solution because flash 3D now all this stuff sounds amazing, the question is how long is it going to last I have so many hard drives I don't know if they still work it's hard to can you give a number? we believe that with the proper firmware if you store at room temperature these drives will last you at least once you store once you put the data on there these drives will last you at least five years with the proper algorithms to to make sure that to ensure that the retention hold up over time do you think if somebody finds this in a hundred years there's a chance it still works? that would be kind of iffy actually unless you're running this in a lower bit mode like PSLC mode then it's probably more likely as long as you don't put too much too much data if you don't wear out the drive too much like in terms of writing too much data to it and if you put it in PSLC mode I think it may be able to last that long I did that video once with some researchers they were talking about 3D NAND and it sounds so fascinating but it sounds like potentially you could one day maybe get much than hard drives for cheaper was that not going to happen we're seeing that happening we're seeing that happening for 1TB at 1TB below we're seeing the SSD cost is actually getting lower than HDD so I think as we continue to scale scale down scale up on capacity maybe within 2-3 years there'll be 2TB and keep going so we keep going and we do see there's going to be scaling limit on HDDs in terms of how much capacity can actually carry is there a chance that the capacity might exceed hard drive could you suddenly overtake and be like 80TB by the end of this year we'll have a 2.5 inch that's 32TB end of this year wow that's pretty much overtaking it's kind of like overtaking yes but it gets kind of expensive so not on the cost wise but definitely on the capacity on the form factor because most HDDs are 3.5 inch so this is 2.5 per volume density the storage density of SSD is already way past HDD there's no chance you can make some 32TB that are affordable that'd be so cool I'll throw all my hard drives away well actually in my household I don't have any hard drives left so I recently upgraded my NAS with our 8TB SATA and it was a very enjoyable experience because now you can rebuild RAID at least 100 times faster than running an HDD but you need at least 2 you need to mirror everything so I bought 5 I got a discount because we build drives alright I must be really fast so what are you trying here so these are industrial SATA I think they're minus 40 to 85 kind of like the automotive except without the coating and the and the BG and the film so these are very small factors of SATA and you actually do show rockets yes well airplanes yes so our connectors they have to be mill 8110D certified we also build custom storage where we should build ethanol our drives I should build with ethanol or samtech connectors that's shock and thermal shock and vibration resistant nice it sounds awesome these kind of technology these cables sometimes I've seen like in the video world I've seen sometimes they have an HDMI with the screws you kind of screw them so it doesn't fall off yes yes yes you have the similar kind of stuff for SATA connectors and everything actually for we're building one that's not just for SATA and PCIe but we're building one USB storage that has automotive grade USB so it's protected against shock and vibration nice cool what snakes do you want to show here well since you're using camera we can talk about cinema SD cards, micro SD cards your market is all the professional video makers also yes we have a big market on cinema so you can buy our products on Amazon on on the e-commerce platforms and so we want to showcase our industrial grade CF express and CFast cards so they can go up to 2 terabytes for CF express and 1 terabyte for CFast so like I said we are leading the edge we are leading edge on CF express but we also continue to support our trading edge technology like CFast now this is really high bit rate video supported on these yes it says max sequential that's CFast you want to look at CF express so we can do 1500 megabytes 1.5 gigabytes sustain with the proper cooling that's a lot of a lot of pictures a big raw 8k video 8k 120 8k is probably 2 gigabytes so it's getting close to it's getting close to 8k so you're going to support it when all these cameras come available soon there's already some but when more and more cameras become available you'll have the storage for it yes so we actually support Canon R5, R3 we support the Nikon Z9 we're the fastest card for my Nikon Z9 we also support the Panasonic GH6 we also have a certified card with red one of the premier Hollywood cameras for filmmaking red guys they do a lot of 8k stuff oh yes right now they're not 8k yet but the CF express will be able to support 8k in the future how is it kind of like on these forums with rumors when people chat about the next generation cameras I wondered if they could do rate recording like not just copying but speeding up when you have two cards it's 1500 plus 1500 and at the same time you might even have multiple copies so that would be a little challenging for the camera vendors because that's a lot of power consumption but with the CF express parts what's different about CF express versus micro SD and SD is the controller is much more powerful so they can do a lot more for the user in terms of data protection in terms of data integrity so with CF express we're pretty confident that you don't even need to do backups anymore just one card would be enough and if something goes bad we still have our data recovery service that could help users recover their data because with my SD cards I like to do recording but that's the SD card it's a different controller it's a much more cost conscious controller compared to PCIe so the capability is also limited especially air correction and you also have the small one type A yes we also have the type A cards right here so for the type A card we took a different route compared to Sony we actually did all of our own PCB PCB design and we did all of our own casing design so that's where we we're able to push the limits of CF type A limit capacity so right now we're offering up to 480GB on the type A but very soon we'll offer 1TB wow is that going to be the best type A 1TB we believe it will be a very good cost performance for the type A market it's also very low power but in the Sony for example the A7 S3 they have dual slots yes so you back up record to those type A so it's actually for Sony camera there's actually limits on what you can do with the backup recording because on the SD card side the most you can do is V90 so you cannot record very high resolution at V90 rates so even if you want to back up on your SD card there's a certain resolution that will not be doable but they have two slots for the type A I think if they have two slots then they can do backup but I'm not sure if they can support two slots at the same time it's a lot of power it's a lot of power and you need to design with two pieces with two pieces in length which I'm not sure if Sony supports that or not so Canon has been doing some AK cameras like the R5 I think it's called and you totally supported those you have great storage for them and you have a lot of customers in the US and Europe or Asia we have a lot of customers across these three big regions and you can go on the Amazon for these regions and they will be there so Amazon is our main outlet for MPNH Amazon is our main outlet for media card type of products and also it's more for our B2B products how about the camera store that wants to resell those they just contact you and place a big order normally we have our distributor in the US our distributor in the US is One Source Video so that's where we supply all of our cards to the US and they can distribute for us and here I see a rugged SSD this is a portable SSD that we can provide for the cinema world and the content creators definitely need to get on board with these kind of rugged SSDs so basically it's IP67 water and dust proof so you can take this anywhere anywhere where there's water, dust humidity, salt it's well protected for that kind of use environment and we build a very high capacity drive storage for this USB it's up to 8TB so imagine the amount of content can put out to this drive and because it's such a high capacity we really design this to be high performance so it's about we can sustain 2GB per second at 50 to 60 degrees celsius 2GB per second what is the connection because I have one of these N1 pro MacBooks and they say they support 40GB that's a thunderbolt but I don't know if there's any external hard drives that support the 20 or something 20 for sure from those MacBooks on the new MacBooks probably so because they were talking about USB 3.2 Gen2 by 2 this is USB 3.2 Gen2 by 2 it is supporting that we totally got the 20 because some of these SanDisk products SanDisk was saying it's 20 but actually it's only supported 10GB on the MacBooks so what's so different about this storage is we don't use SLC cache to show like a burst performance like you can do 2GB for 2 seconds and the rest is 1GB or even less we try to make sure that you can record across the whole drive at a sustained 2GB so what that means what that means is really you can offload the whole drive in one hour 8TB of data in one hour that's amazing I know a lot of cinematographers complain to us about how they're using HDD it takes them forever to transfer media with this you get at least 20x acceleration offload that means you have the controller in there that supports that latest USB-C and stuff we optimize the whole performance power thermal to be used for USB Type-C what's amazing with that solution is that people should get the MacBook with not so much internal storage just get one of those and then they can do all their video editing 4K60 off the external same experience like if it's internal maybe even faster maybe even faster we're actually pretty confident that it should be faster because NAND Flash's performance is limited by capacity so at 8TB you have no performance limit it can saturate the USB 3.2 Gen2 by 2 because the M1, M2, M2 chips that Apple is selling pretty amazing but it's nice for them to keep the storage external the heat keep it outside your MacBook and save some money and put the money in external storage you probably get more per dollar yeah not only that they charge quite a bit for an additional 256GB and 512GB think about it for the same money you can get 8TB and it runs faster I always choose the lowest one and I want to go for the external and get much more storage per dollar so people use this for direct video editing and they're in the front you have a demonstration you're showing it to booth oh yes this is pretty much this is what we're doing with conformal coating so the drive is conformal coated and it will operate as if it's not in water is it something like what they were talking about P2I they're putting it in some kind of oven and then it gets coated by something yes so we actually have an automated manufacturing process that coats the PCB with a hemi seal and it's cured in an oven for about 2 hours and once that's done then protective coating forms on the PCB that protects against water but that's not all of your storage right some of them it's an optional manufacturing process and then that's the ultimate liquid cooling well we don't really need the liquid cooling because the drives are actually fairly low power no need to liquid cool you're showing that it works and you're showing that it just keeps working even though it's in water yes so a busy embedded world lots of potential embedded industry people we're excited to be back here last year the crowd wasn't as much as this year so we're looking forward to meeting more customers and sharing with them more of our products services if it's relevant once again hi my name is Thomas Lorenzo I work at ARM in the IoT line of business and welcome to the ARM booth so here the embedded world is pretty much the ARM world everybody is doing new ARM projects here there's a lot of new announcements also yes super excited about the partner announcement this year we've seen several partner announcements in the MCU world such as the ST microcontrollers which were announced also renaissance is showing off the Cortex M85 which is one of the highest performance M-class CPU we have to offer so really pushing the boundaries of performance for machine learning for signal processing but also for regular scale workloads there's a lot of demand from some parts of the embedded world to get more performance on all these microcontrollers we're definitely seeing several trends driving more computer capabilities at the edge there is certainly machine learning as I said we find demos around voice activated door locks for example which is quite interesting we find demos around connected motor control integrating machine learning into that for predictive maintenance but also at the high end of the spectrum definitely more applications around machine vision integrated also machine learning capabilities into that so basically running a multi core Cortex A based system I saw some booths that were talking about Cortex M and GPUs and stuff like that yeah absolutely now definitely see vision and graphics becoming more interesting also in the MCU world so doing things like smart displays in appliances for example so that's certainly where GPUs are now extending into the microcontroller world as well very excited about that and Cortex M with AI Cortex M with AI absolutely as well yes so there is as I mentioned Cortex M85 from Renesus that's integrating now the Helium technology so Helium is a vector processing capabilities on Cortex M for signal processing and machine learning so that's really pushing the capabilities of those microcontrollers but also we see in partners as we have over there if we want to walk over there is that some of our partners like Himex they integrated Cortex M55 which was the previous recently announced and integrating our EPU-U55 which is a neural network processor so you can see here the demo with the camera so what is the ARM processor it's a Cortex M55 combined with the EPU-U55 so really targeting machine learning in tiny applications at the endpoint essentially and it all goes on one SOC really targeting battery power efficient solutions with usually battery operated applications and they can do a lot of stuff right there all the stuff that usually is in the Cortex A with all kinds of stuff everything is running at the edge at the end of the endpoint you can do all the inferencing at the end for enabling applications where you don't even need any cloud connectivity that's cool what else do you want to show we can show over there if we go back there so we have something we started last year is initiated it's a project called ARM Virtual Hardware the idea here is really that you can start your software development before the hardware, the physical hardware is actually available so virtual hardware essentially is a virtual representation of a physical port so if you think about in the past you always had software development and hardware development were kind of sequential so it always software development to some extent started after the hardware was available the concept of a virtual hardware now really enables developers to start with their software development before the physical port or the hardware is actually available so very excited about that because it enables faster time to market is it simulation? Yes, it's basically using some ARM technology we've had in-house for decades which we're using actually for our modeling of SOCs and we have another solution where we partner with another company where we're providing basically a modeling of full physical so you could actually start to develop on a virtual Raspberry Pi for example Is it FPGA stuff or it could be different? No What it needs to run on? It basically runs in the cloud so you log it into AWS essentially it's running on AWS instances in the cloud so there's no physical port there's nothing involved in that Cool So if we go around there's something else I wanted to show you as we talk about the edge right there so many different use cases right here so many different use cases it really is sometimes mind-blowing right, you have all the way from Cortex-M smallest microcontrollers all the way up to as I mentioned before really machine learning or the machine vision running on Cortex-A based devices and what we see especially in the A-class world is that often you have platform development for each individual platform which often is we look at that we see that that there are things which our partners or the developers have to do for each individual platform but there's opportunity really for reuse right so we don't have to reinvent the wheel often so what we're working here on this specific program called System Ready is really about what can we standardize in the software development floor in the software stack so that our partners or the ecosystem can focus their efforts on differentiation so what we're showing here is there are five different boards so Raspberry Pi and some Pi-4 Arduino, AD-Link and Contra what you can do here is you can use various different operating systems which basically can put on any of those devices so System Ready essentially really works around the interoperability between operating systems at the layer between firmware and the operating system so standardizing that layer essentially so that any operating system can land on any System Ready certified hardware nice so inter-pillability yes exactly so what that means is really you can run any operating system on any board so you can the demo here is that you can use an USB stick with for example Fedora or SUSE you plug it into that board it basically puts the system you can use the same USB stick to put up your Raspberry Pi for example so this is really where we see a lot of value in this program because putting the system is basically much much easier you don't need to spend time to develop your code and to just put essentially your Linux image into the ARM is to be interoperable and once you have software on one you can do the other but now this is the next level to get it even more put a USB stick a little bit like what some people do in the X86 and they just have a Linux and they somehow boot it I think it's about the developer experience and this is what we really want to enable with standards that's why standards are so important for us because it enables the software reuse there are aspects of the software flow or the software stack which are often non-differentiating so that's where we developed and collaborated with the ecosystem and developed together to get a system-ready standard to really ensure that any operating system can land on any ARM system-ready platform all right cool so that's a busy show and lots of demos there's lots of other stuff but we cannot cover everything it's very very busy it's great to see the industry being together again and it's great to see so much stuff on ARM so what's the latest here at the Embedded World 2023 hi there well today we're showing our many new things this is one of them this one is called our Jasper Commix Press Carrier Board what you see here is the top side of the board which has all the IO and all the expandable sockets things like that and on the back side of the board is the Commix Press sockets you can see here the standard Commix Press connectors this supports a Type 6 Commix Press both compact and basic form factor and this metal plate around it is a mounting plate and what happens is the Commix Press and its own heat spreader fit in the middle of it so this thing is a general purpose platform and you can work with any Commix Press Type 6 module and then the idea of this is that it's meant to be put inside of a rugged system so it performs a scalable platform for a rugged computer system and what we have here is the whole board is designed for rugged applications all the connectors are rugged and latching the board PCB is thicker if you can see that it's a thicker PCB it's a .093 or 2.3mm thick PCB and then we have latching connectors for everything we have M.2 and PCIe mini card for IO expansion we also have PC104 Express socket here for IO expansion as well and then we have PCIe by 1 and by 16 so if you have a Xeon Comm or later generation Commix Press then you get the PCIe Express by 16 lanes and so this can allow access to those lanes for high performance rugged IO like video capture or graphics or other 10 gigahertz or so on when you talk about the Commix Press Type 6 basic and compact there's a lot of stuff that use that it's pretty much the sweet spot of Commix Press the two sizes the compact is more popular because it's smaller 95 by 95mm the larger size, basic size usually will have higher Xeon processors or more memory capacity because it has more room for memory sockets but this work is both of them so it was designed to be a general purpose so that it can fit either the mainstream or the high end Commix Press so it's like a Comport with much more speed well it's whatever the Commix Press can offer so Commix Press comes in many different levels right the whole value of Commix Press there's two values the first one is that you can design one carrier board like this and you can plug in any Commix Press modules so that you get scalable performance so if the customer wants to improve the performance you can pick a different Comm that has a higher performance and put it on there or if they want lower performance or lower cost you can pick a different one so it fits the application precisely the other thing is that because one Commix Press module goes on to life or becomes obsolete or hard to buy you can switch to a different module so it gives you the ability to maintain a product lifecycle much longer than normal this is really critical for anybody in a long life application such as military or railway or medical those are the typical applications where they need very long lifetime and so what Commix Press or any con based solution does is it gives you access to a solid platform where the platform itself remains the same but the performance can increase or decrease by time and it can last longer so the product can stay life longer this is Intel? this is all Intel X86 technology where does Intel go? and how do you design around it thermally everything is smooth so it's so the idea that we do is it's all conduction cooling so the module goes on here unfortunately I don't have one to show you oh here's the module but I'll show you this this is the module it doesn't have the heat spreader so here you can see the Commix Press module it's mounted on top or underneath rather in this case the carrier board the same Jasper carrier board that you see here is actually here now it's upside down and this plastic cover would normally would be a metal heat sink cover so that it's for heat dissipation so all the heat is going through the heat spreader and into this case upper surface with heat sink fins on it as normal that's how the heat dissipation works so then you can just throttle it a little max speed depending on the thermal characteristics depending on the thermal characteristics of the enclosure you can max out performance so it all depends on what the enclosure can handle and what are the applications like 2 or 3 what does that look for this is a rugged system that's really targeting military applications because of the size and ruggedness this is a typical box PC you would see or a mission PC mission computer you would see in a military vehicle where it's got the rugged connectors that are all sealed and super rugged you can step on them and kick them and so on they've been around for decades they're very solid, very popular this is intended to be a rugged computer system that offers scalable performance so it has comic stress inside of it so that you can pick the performance level you need pick whatever processor the customer wants and it's going to take a long life cycle because again in military programs they want the product to go for a very long time they don't want a risk that the product has to be redesigned if the CPU goes out of life so with the comic stress based solution then the product can last longer that's the whole purpose of this platform and sometimes when something is military spec military spec it's not necessarily military it could be all kinds sure this could also be used in any rugged application like for example mining these vehicles also require computers and they also require extreme ruggedness high shock and vibration tolerance so mining is another very common industry where products like this are used and how about these oil rigs out in the sea those usually have a lot of demand for wide temperature and ceiling for rain and water and that kind of thing maybe not as much shock and vibration although we have a new program now on oil rig where they're talking about high shock and vibration because they are drilling and the pipes are very heavy and sometimes they drop the pipes on the platform and they said when the pipe drops it's super high shock so there's another application where the shock and vibration protection is needed are you in rockets? we have a few projects and satellites yes and we are actually in some satellite networks that are going up our use is one of our products in them but it's different product family needs to work there's a whole different set of issues the launch for sure is a high shock and vibration environment once you're up in space the issues become temperature extremes as well as radiation tolerance those are the other issues and also zero pressure so there are other issues with components when you're talking about high pressure or low pressure environments and this is cutting edge in terms of what people need and what people want in the industry what's special about Jasper is that it has a lot of I.O. and a single board and a lot of expandability so that if you just use this as a platform for your application then you're pretty much guaranteed not to run out of space and not to run out of time you can add whatever I.O. you want to expand it in many different form factors either mini-card M.2 or PCI-E104 as well as you can have your choice of common express modules from dozens of manufacturers and so you have long lifetime as well and you can pick whatever processor you want this is going to be a common available for it so that's what this gives you so we're pushing it as a platform for high performance applications that need a lot of I.O. and this metallic plate is required this is the mounting plate that is used to mount it into the enclosure it's not required you don't have to have it it's just a convenient mounting plate because the actual cooling is from the common express heat spreader which is inside the middle here so this is only a structural member and it can be done your own way the real thermal dissipation is coming from the common express heat spreader which is in the middle here and on the piece of paper right here you talk about some of the other specs these are just the general I.O. features that it has so you can get a quick view there another nice thing about this product is that we come from the analog I.O. world so Diamond actually started making analog I.O. data acquisition boards and so to this day we still sell a lot of analog I.O. and we embed the analog onto the SBC so we have a whole line of SBCs where we have integrated analog I.O. on them so it saves the customer an extra board instead of having to buy an analog I.O. card and plug it on the board it's all built into the main board so this board has it also but this board has the same circuit right here this particular model is the low cost model so it doesn't have it but it's available with a complete data acquisition system here FPGA based data acquisition with analog and digital and it has auto calibration to maintain high accuracy over wide temperature range and to show the whole range of the other products that are similar so here we have another product this product is called Saturn this is more of a mainstream typical single board computer everything is on one board here it's got PCIe 104 expandability it's got M.2 and mini card expansion as well and the CPU is on the back side you can't see it because it's under the heat spreader but on the other side of these thermal pads is the processor this is the single board computer and a more traditional form factor again with a heat spreader mounting plate for efficient cooling and then expansion and I.O. are on the top side so here again we have the same analog I.O. circuit as you saw on the other board and again this board comes both with and without not everybody needs it so we make it in two versions this is the model without the analog I.O. and the people who want the analog I.O. they want it for real world it's easy to have or ethernet everybody has that but a lot of times you have things like you have to monitor pressure or you're monitoring water flow or you're measuring light intensity or you're measuring gases in a system because you have some kind of like I.C. manufacturing you need to measure the gases the gas flow and the gas purity that kind of thing so our boards are very common where air quality is a big application for us people are measuring here alright the boards pretty much have analog I.O. on them even this little tiny one called Zeta here this one I'll bring closer so you can get a better view Zeta also has analog I.O. right over in this area that's a pretty nice lens you have there and over here you can see again it's ComExpress it's a two-layer two board solution where you have the carrier board with the I.O. and the expansion and you have the Com with has the CPU on it so that's how we're able to put analog I.O. on a board this small this is called our Zeta board so are you famous for supporting the analog I.O.? I think we're pretty well known if you ask people about Diamond they would say oh Diamond's the P2104 I.O. company we're trying to divorce ourselves from being seen as a P2104 company because we do so much more now we actually have four complete product lines analog I.O. is part of the I.O. product which is again part of the Singapore computer product line which is here then we have a very successful line so here we have single board switches expandable switches with daughter boards with more ports on them here we have a switch module which is very much like a Com but the switch module is basically a switch instead of a computer and the same thing where you plug it onto a carrier board and the carrier board completes the circuit and has the rest of the I.O. so what these are are carrier boards that break out the I.O. on the switch module so this one is this is the actual switch everything it's built in application software and this one is the I.O. so it's got the magnetics and the connectors the SFP and so on so it just breaks it out to a full solution typically what customers will do is they will buy the switch from us which is the hard part because it's the whole high engineering and software development and they'll make their own carrier board this is the easy part so it divides the solution into hard and easy so they can get to market much faster and they can make a custom solution without doing all the work I.O. and you have a history in doing that you're right when did you start doing that Diamond was started in 1989 so I've been with the company since the beginning since I'm the founder and so we've been making analog I.O. digital I.O. Ethernet serial boards for 33 years now what are you doing in 89 what was the product in 89 we were making slot boards actually PCY slot boards I support and then we moved to PCY104 and the first came out PCY104 was just the isobus and so we took all of our existing circuits and moved them into PCY104 factor and that became very popular so from there we went to make X86 single work computers then Ethernet switches and so on and it's fun to see all the markets that adopt your technology you know we have the craziest applications applications I would never even imagine people do all kinds of crazy things one interesting application was chip packaging they had to measure the thickness of the material being made the aluminized plastic film they had to measure it at high speed when it's being manufactured and so they would zigzag a radioactive sensor across the material and the amount of scattering dictates the thickness of the material and they would use our analog I.O. board to measure the scattering so that was an interesting application that was a very big one for a long time I hope the chips aren't radioactive just the packaging then it goes on to a roll and it goes off to wherever but you get all kinds of when you make I.O. you see all crazy applications around the world and things you never even knew about existed but it's just amazing so a little slice of life we have something like 500 active customers which is pretty amazing for a company our size we have 500 customers and they all do different things we have UAVs, we have submarines we have military vehicles we have potato chip manufacturing we have medical one of our big customers also makes a a lung simulator so all the respiratory equipment in America has to be calibrated so that you don't blow out someone's lungs so it has to be calibrated for different body types and you know baby adult woman man that kind of thing different conditions and so they have a lung simulator that they use to calibrate their equipment and that uses our computer board with integrated analog I.O. as well so it's a really interesting application for analog and here they embed a world, lots of meetings with all these non-stop meetings so we have a lot of sales partners that come to visit us here so typically we have sales meetings going on all day this is a small lucky break in our meeting schedule so I think the guy's off to lunch right now because I'm having more meetings soon so yeah we have meetings here and we have some customers come to see our newest products and we're going to do a second video right now let's do a second one let me show you one more thing though we mentioned this product here this is called Geo I mentioned the fact that it uses the common scratch module for performance scalability in long life but what's also interesting if you look inside here if you can get in there you can see that there's a board to board connection so the I.O. board doesn't have cables all the I.O. connectors are mounted on a circuit board and that plugs directly onto the main board so it illuminates the cables in the box and what that does is number one it makes the box more compact because you don't need the room for the cabling and number two it allows you to get higher shock and vibration because you don't worry about the cables and even number three it makes it easier to manufacture because quicker and easier and more reliable to build a circuit board than to build cables when you build rugged systems like this the number one sticking point is always the cable manufacturing to get the parts to build it to get them to build right have them come out right it's much easier to build a circuit board so better you've solved cables we try to solve the problem but also we have built-in expansion connectors so all of this is for the standard X86 board all of the I.O. on this board comes out these three connectors and if you want to add more I.O. remember that on here we have two mini-cars if you want to add more I.O. all of these could get cabled out to these connectors on the backside of the panel board so the case and the connector design remains exactly the same so there's no custom design required so we eliminate custom design which eliminates NRE charges, reduces time to market and provides a more rugged compact product so it basically solves every single problem that you could think of and you provide this amazing solution and maybe there's a big team that then goes on customizing some software or whatever they need so the specific application the customer will basically buy a platform like this they will write their own application and they may add additional I.O. like a radio or some other sensor they have or some military bus interface or whatever it's going to be maybe they have some other kind of I.O. they need and they can put that inside the box and they don't have to do anything with customization it's all ready to go just plug the board in so it's great for quick turn easy custom customized solutions with no NRE charge cool and then we're going to do another video okay, okay, sure okay, well that's ready right now okay did that come out okay? not too many uns or anything no, it's perfect, I was thinking we could do one long but I think it's good to start with the second one two is people with a lot of different things people can write exactly everything else that's more related over there maybe then we go to this one second and then it looks good let me start it right there so I'll do it like this and then you can say hi hi, I'm Jonathan Miller at Diamond Systems and we're here at Embedded World and today I'm showing you our newest NVIDIA Jetson Carrier Board this is called Osborne and if you know Diamond's products you know they're rock and roll stars or rock and roll characters so here we have Osborne so there's a couple people whose last name is Osborne maybe you can think of them anyway Osborne is our newest product and this is focusing on the AGX Oran which is NVIDIA's highest performance carrier a GPU module the Oran module mounts right here on this socket in this area on the board and then we have it's covered up on this side let me see if I can pull this off to show you there we go, okay now we have expandability so we have two mini cards we have M.2 socket here for Modem and we have a camera adapter board here in addition we have here a PCIe by 8 slot connector so if you have a high performance graphics card you can or camera adapter interface board you can plug it in here so it's got a lot of expandability as well as a lot of value on the board already now the interesting thing about Osborne is not just that the AGX Oran support but also it has a concentrator connector over here so all the I.O. on Osborne comes out on a single I.O. connector what this allows us to do is make us a group of interchangeable I.O. adapter boards it's just like you might plug an I.O. board on top of the board here you plug an I.O. adapter board on the connector so you bring out the I.O. in exactly the way you want it so in this particular case we have CASE commercial style connectors like DB9, RJ45, USB HDMI and so on and these connectors this is for CAN right here RJ11 for CAN and then POWER so this is the commercial style connector board that can be used to go inside of a typical box PC kind of enclosure but if you wanted to use Osborne or use the AGX Oran inter rugged application then we have a second connector board that looks like this right here with the military style connectors so here we have the exact same boards board connector but now we have different connectors on it which are designed more for a rugged military application so the same carrier board can fit in any different application and of course CASE can design their own carrier board as well I.O. board as well so that if you want to do something different or a different kind of connector or even different military connectors whatever then we give you the design files and the specs and everything and you can make your own adapter board to adapt Osborne to your application try a little bit what do you call this connector here oh so this is basically just a module connector I don't have an example of the Oran module here but the Oran module is like a big metal cube kind of thing and it plugs directly onto here and it has all the input output power and everything everything you need to drive the Oran is on this connector 699 pins that's a lot of pins has a lot of money and then that's everything that's the whole interface then it mounts on these four on these four threaded spacers here standouts and then you were talking about this right here so this is just a standard desktop slot connector that you plug in a typical slot board in it has PCIe by 8 on it the Oran module has PCIe by 8 available so we bring it onto that connector so some customers want to use those desktop style adapter boards for whatever reason graphics or ethernet or something like that so we give them the ability to do it with this connector here and there's 10 gig 10 gig ethernet is available on this board yes there's one 10 gig port available on the Oran we bring it out here so it's actually on this connector I think the bottom one is the 10 gig ethernet using a 10 gig fi right here this little part so this is high performance super high performance the idea is high speed video end high performance GPU for analysis and inference whatever it is going to do with the energy coming in and then 10 gig high speed ethernet out so you can get the data out really fast to the network so it's a real high speed edge processor the whole chain is high speed you have many years working with Nvidia we've been working with them for I guess four years now and the first time was 2020 just before the pandemic we had our first Nvidia products so it's about four years now and you have a bunch more? yeah this is our latest one but we actually we have one more development coming up we're showing it here this one is called Jackson so Jackson is for the other Oran products Oran is actually a family of the latest Jetson modules the AGX Oran is the high performance one we just looked at this one is for the Oran Nano and Oran NX those are the lower cost lower price they call it entry level and they're like a dim style so they plug into this socket here very low cost socket so the module will plug into here and then on top of it you have your cooling solution and the IO comes out here so this is more for the lower lower end applications it's something like five to ten times the performance of the previous generation so you get incredible value incredible performance at a very low price with the new Oran Nano NX what's the main CPU? it's all Nvidia GPU the Nvidia GPU is actually with an arm also yes they have what they call it I can't think of the word right now heterogeneous design so they have an ARM complex which is either quad core or hex core a CPU complex to run basic Linux OS, basic applications, user interface things like that and then they have a secondary GPU which is for all the fun is which is for all the video and analytics and inference AI, robotics and so on it's all being done inside the GPU complex so it's kind of a dual processor the kind of thing they can do is they can bond a bunch of networks together they can encode HT65 and do all these advanced things people want to do the most common thing I think is video analytics so take a picture and analyze what's in that frame do you recognize the person can you detect the motion of the person and what they're trying to do things like that we have one application for example where they use the Jetson to analyze the video frame to watch people going into a building so they can tell if it's one person or two people or if you're trying to pull equipment in to authorize this kind of thing so they use the AI in real time to detect whether the action at the door is authorized and so that's an interesting application the entrance control and there's other boards around here you want to talk about so this is the previous generation this is called Floyd very similar to the previous generation of our carrier board here you can see the module is actually right here this is the Jetson module there's a solution on top of it it's a heatsink with a fan on it so it's very similar there's a larger board with a little bit more IO on it this is called Floyd a lot of people have been using it Floyd is pretty popular Floyd is being used to electric vehicle charging and pill dispensing and all kinds of applications right so that's it these smaller systems are designed for this kind of enclosure so that's why they have those connectors on them so we put them inside a box so you buy a complete solution you can buy the box, make your own enclosure or you can buy a complete box system from us ready to go so the modules in there, the Linux OS is installed everything's ready to go, just power it up what's the minimum order quantities to work with you just one if you want to customize, there's a story but we sell one to $9.99 all the time alright, cool thanks a lot for the show and always busy in the embedded world no matter what I think of, someone else thought of it already a lot of innovation here it's really a drink it from a fire hose to come here and see what's happening it's so much fun, very educational use yourself this is Dimitri Sanastasiu on behalf of Verida Labs I'm leading the product growth team Verida Labs is an embedded Vision AI software provider and we have developed our solution which is Perceive AI a software and services platform that can help you develop and deploy at scale Vision AI solutions that are targeting the edge so I'd like to discuss a little bit about the platform actually it's the platform itself so this means that you need to have a very good understanding of all the steps involved when you want to create a new Vision AI solution so in this case for example when you want to develop a traffic flow monitoring sensor, you can just directly first of all create it's really equivalent can we do this again because it will take some time to load and we will waste time sorry it was live streaming but it's okay I'll record I'm recording in 4K I didn't know it will be live streamed it's also live let's start again sorry I didn't know about that hi ok so this is Dimitris Anastasiu hello from Merida Labs I'm leading the product growth team we're doing embedded Vision AI software and we have developed a platform to help enable the development and deployment of Vision AI at scale this platform is called Perceive.ai and I'd like to show you more here so in order to develop a good Vision AI solution at the edge you need to have a good understanding of all of the steps involved first step being the Vision Twin part where you have the modeling of exactly what goes on in a scene in 3D so you can select the right module from the camera and lens and optics so this example is the traffic flow monitoring sensor that we have created for the smart city context together with Renaissance more on that in a little while the next step will be to annotate and actually choose the right classes for your data so through our platform on the cloud you can directly manage all of your data sets as soon as you're done with that and you know what you're doing the next step is to start creating models so first of all you can choose and create a new model here by choosing the name, the description, the task type you can select your data sets and make sure that you use the right data sets to create a model and then you choose the deployment platform in this case we have selected the Renaissance MPU DRZ Bitwell as soon as we have created a model we can directly see how well it performs right on the cloud after it has been trained you can select a different model or you can deploy a different model on the edge device that way you can easily understand if something is going well or not going well and of course the final step is to get metadata directly from the sensors so you have deployed sensors on the field but you don't need to see the images you just need to get metadata in this example we have a traffic flow monitoring application so we are counting cars per minute I'd like to show you a little more on how that works so here is the real time live demo of exactly what goes on we have developed this solution this Vision AI sensor together with Renaissance it has a tiny ML engine and what it does is in real time it counts the number of vehicles that are crossing a very specific lane this is the sensor level dashboard this is why we are seeing here also the imagery with the metadata and of course as shown previously this is directly sent to the cloud this platform actually this solution is meant to run at very low VAT consumption and it's meant to be used in a smart city context and it runs our tiny ML engine inside per CVAI compliant it interfaces with MQTT so it can be directly integrated into a third party platform or maybe digest the metadata directly to another software software providers infrastructure that's pretty much the summary and what's there on the wall what are you talking about here actually discussing about different use cases that we have deployed over the years some of them now we have hit the part where it just discusses about the platform but the idea is that you have a lot of different markets where Vision AI can be relevant so this example that we show previously had to do with smart cities but at the same time we're also let's say market agnostic we have developed a lot of solutions for industrial players for example in the industry for context for warehouse monitoring for inventory management for track monitoring or also understanding what goes on in retail in terms of product recognition self interaction inside specific stores perhaps some consumer analytics so all of these are enabled through our Per CVAI platform and this is the let's say the thing that glues everything together from conception all the way to deployment down to the edge cool alright are you like a new company you've been working a long time on this stuff we have been working for quite some time so we have been in the embedded vision field for I would say more than 10 years now so we have a very good expertise of what's going on in embedded this is why we're able to deliver solutions through our platform in a very timely manner that said we have been working in AI more than 5 or 6 years now and the latest offering is of course our platform which enables also other developers aside from our company or together with us with engineering services to provide very good products and what's the collaboration with renaissance we're working tightly together to create the vision AI sensor so we're working for bringing this new product to the market and actually through the business cases that we're working together we'll be able to scale in thousands this specific solution some of these newest ARM processors even the the Cortex-M stuff and microcontrollers are getting extra performance to do all this kind of stuff yes some of that can be offloaded there it can be done in very in many applications sometimes you need an accelerator like the RZ-V2L but indeed some of the applications can be offloaded to ARM as well as architectures it's depending on the use case and if you have the right expertise and we're happy that we do you can go very very low and do very nice stuff with vision where are you based? in Greece we're 30 people strong growing quite fast in the past few years in where? in Greece so the headquarters are due westwards from the capital which is Athens the city is called Patras it's right next to the university so it's coming directly from university we've had quite a few young talents joining us now we're 30 strong so it's 30 people most of them engineers but we have also our senior people working for quite some time on this group now a lot of people talk about AI machine learning computer vision and you are at the forefront of this we consider ourselves a leader we have been in that for quite some years and we care about bringing good quality products to the market so we have deployed mass marketing in China we have very good collaborations in products that plan to scale in the many thousands there is this buzz word but there is a huge gap we have what we call the CV tech depth this means that it's very easy to deploy an open source solution actually to develop an open source solution by using some of the models that are already available but productizing an actual creating a product that does solve something that's an entirely different discussion you need to understand how to deploy in production and for that you need to have an understanding like the one presented previously you need to understand end-to-end what's going on you cannot just assume that you have the best images and assume that these images automatically reach you in most of the cases you have to deal with diverse scenarios with different environments being able to adapt having the infrastructure to get the right data and at the same time adapt to the environment and also meet the below materials and the performance requirements so there's a lot of discussion to be done there because the buzz words are nice but we have to have products hit the market and no dancing and nothing else and thanks for being here you are at our Renaissance booth you are now together with me my name is Frank Orber I am leading the SST team this is our system and solution team in EMEA I am together with my colleague Brad Rex Brad is leading the SST team in America and Brad he will show us a huge new demonstration a huge new tool we call Quick Connect Studio and Quick Connect IoT and this is unique in the industry and please be excited please see what we are doing and Brad if you can give us some highlight information on Quick Connect Studio would be really appreciated Thanks Frank Quick Connect Studio and Quick Connect IoT are kind of the hardware and software for our brothers of each other so Quick Connect IoT allows us to take our modules plug them together in different formats and variations that way you can make your application exactly what you want and now what we are doing is we are making a software version to allow us to auto create the software and that's Quick Connect Studio which we are showing so we will basically have a drag and drop type of solution where we can take our MCU board we can take a connectivity board and we can take all from scratch we are doing it graphically so in this case we would have an MCU that is reading temperature and humidity data and sending it out to the elite we would then compile this Brad just to interrupt you is this now local on this machine or where is it so this is all done cloud based so it's on a cloud instance so it's all done online in a browser we are compiling here we are going to use the cloud's compute power so it's much faster you don't have to have a very powerful laptop to do the compilation took 2 seconds to do the full compile and then we are going to get the project file we are all done so what you see here is we get all the source code for the actual project and then when we compile we get the debug file so what we will do is we will download the debug file and then we can actually do the programming so that's there and what we will do is switch over to our other tool this will allow us we will erase it we will select the file we just we are going to program it so we will do the programming so the device is programmed and since this is BLE we have a BLE app so what we are going to see is if I get a smart phone app here select our device and you will see all the temperature and humidity data coming out if we had a different sensor like a flow sensor you would see flow data so the app and the data that is being sent out from Quick Connect Studio is automatically configured based on the actual devices that you are using what do we need do we really need a high sophisticated high end machine to run on locally or what is requested from our customers so you don't need much other than the hardware and a browser that is pretty much all you need and a way to flash the device locally to make Quick Connect Studio work so really low barrier entry you are not downloading any tools just a real simple browser and some hardware you can do it with a Chromebook you could also you need something else maybe you can actually load it on your smart phone and do all the configuration here obviously you need a way to download the file onto a PC but you could use a smart phone with a browser to do all the actual development of the solution can you talk a little bit I see here two renaissance controllers what is the next step so right now good point Frank we support two different MCUs Bluetooth, Wi-Fi and a couple different sensors our plan over the next few months and then for 2023 we are going to get all of our MCUs supported in here all of our sensors we are also looking at partners so other technology that renaissance does not have like gyro accelerometer things like that we are going to get an ability to have our partners actually add in support here as well it is a lot of fun and exciting things coming in 2023 this is a brand new way of doing things it is the first way in the industry to actually do it from this level there are many online IDs that already exist today but that is really source code level programming and debug and development this allows you to start one level above that at a solution level then take that solution make the code and then from there you can then go on and do the next part of your actual development process with the wig something what do you call it drag and drop what you see is what you get same style so basically if we look at it it is exactly mirrored to what we have so that is the whole point and then again it comes with example codes and you can modify and improve yep so the example code will intelligently look at which devices you have and auto create that example so for instance if it is BLE it expects that you would send it out to a BLE app like on a mobile phone if it is wifi then you would send it out to a web browser like an address page so that you can take a look at it so it intelligently knows which devices you have and it will automatically reconfigure the question comes from the way of having what about a price do we charge anything to customers what is our business model to work with customers so right now we are in our beta stage for the tool and development so right now the tool is free of charge to use it is renaissance.com slash qc studio users just need to sign up for my renaissance account and from there they are able to then use the tool free of charge please submit feedback on how you are using it we want to get the features that users are most implementing on added so right now it is completely free to use really and then hardware for hardware availability renaissance provides samples or you can get them through distribution digikey, mauser and avenue future all of our major distributors affordable most of the boards less than $100 the little pmod boards are probably 10 to $25 range last question we are now here at day 2 of embedded worlds you have met a lot of customers can you share high level what is the feedback of customers what is your impression of being here in Germany yeah everybody is very excited first thing people ask is like oh it looks very modern, it looks like visual studio code people are used to that they like the new look and feel it is very dark, it is very black it is very coder like we can change the colors people are very excited for this it is an option, people want options in how they do development so this gives them more flexibility is there a potential way to collaborate like a team could be working together on stuff in there it is possible, the tool doesn't support it today but definitely the framework here is very flexible to allow that we can add extensions things like github integration so if you have a collaborative source control management we can do that I joke with people you can get chat gpt in here as well probably at some point and you can just tell chat gptd to make it work again these are all plugins and it is fully adaptable I think we are really excited to rent a sauce that we are able to produce this tool for people and I think in 2023 we are going to have a lot of fast updates coming in that are going to make this tool really awesome thank you very much Fred what we are doing here it is the next step of evolution the next step in our development with our customers I hope you will be here in one year again we will see much more Pmods we will see much more kits airball kits, boards starter kits from our site and you will see also some of our partners having here modules so please come again next year stay interested being interested in our solutions on our website www.rendersize.com and you will figure out we connect studio so thank you very much for being here yeah welcome to the rendersize electronic news at the embedded worldshare 2023 in nuremberg germany my name is ben vestoff I am taking care for the ramcu product marketing and I have the pleasure to introduce you to the latest development if it comes to ai embedded mcu so what we have is brand new of course it is brand new the latest you can get the latest and greatest you can get right now it is basically bringing you the coltex m85 core which had been introduced by arm somehow march last year and we have been ensuring first implementation on the actual silicon of that coltex m85 core by june embedded world last year whereby now today we are introducing the latest implementation having ai people detection implemented on these parts right is it the first time that ai is on a arm cortex m so let's re-find that it is the first time that anybody on earth had ever launched a device with coltex m85 implementation on silicon level and secondly it is the first one being on earth again and it is not on your worst side but on earth have implemented ai on this course there is no other vendor yet out there offering you coltex m85 implement on an actual silicon right so this is the running demo and I want to just for the viewers out there just to imagine and experience this instant let's say I mean we are very close to that now but the instant ai experience so let me press the reset button there we go wow that is MCU that's half a second I'm not sure very quick quick at least and that allowing you to have applications or to address ai to vision ai people detection ai to application where you not been thinking about yet before because usually ai were implemented on mpu's where you have a kind of or you have the audio loss ok good so where you had an mpu used which had a kind of boot up time 10 seconds for to get the linux running or so this is an MCU an arches and the instant ai experience is the enabler for applications like air conditioning right if you have an air conditioning control unit or a wall outlet at home and you are entering the room as soon as you enter the room this guy is able to detect you where you are how many people are and where you are located so the ventilation can be put into the right direction to get you feel cool and fresh right another application maybe video doorbells where you may not able to detect people approaching your house you see people coming along and you can detect that there is a person secondly you detect the person faces and with the faces recognized you can may ring different doorbells signals right etc there are a lot of different applications out there which can be used for these Cortex M85 with ai and important to recognize here as well compared to these Cortex M7 cores out this is giving you a general performance uplift a significant performance uplift a roughly 4 times compared to a Cortex M7 at the same frequency 4 times 4 times that is a important point you need to take away as well so therefore this M85 implementation is now giving you a chance to make AI vision AI people detection implemented on an MCU quite efficiently and enabling to address applications you never been thought about before is it the most powerful arm Cortex M ever yes the most powerful arm microcontroller up to now let's say up to now so next gen maybe it's also going on smaller nanometers than some of the previous microcontrollers or you don't necessarily talk about the manufacturing size I think we can talk about the manufacturing size and we are using of course latest technologies for implementing our MCU's this guy going to be part of the RA family and really is known as using your latest process technologies as 40 nanometer for example or even more advanced technologies in future products so it's power efficient even though it's very powerful yeah of course because that's one of the benefits of MCU's having power efficient solution with good power profile and everything is embedded flash and better RAM embedded accelerator called helium right so everything is there what is this accelerator that's a helium vector extension part of the Cortex M85 which can be used for I mean it can be used for different purposes but we are using it for accelerating the AI interference right so you can use it also for DSP applications or DSP calculation applications is it very special to add more stuff on the SOC and now add the AI stuff it's not an SOC it's an MCU from our point of view we differentiate SOC and MCU sorry so it's an MCU you have the AI stuff so right now with this solution as you can see we offer a fantastic instant on AI experience you press the reset button you have immediately performance out there so there's literally nothing you need to add to make it a perfect fit device for various applications I mean this is just a picture viewer right I mean it feels like a Cortex A yes it giving you the Cortex A feeling but it's the actual Cortex M Cortex M85 CPU MCU experience what's down here that's a M85 base motor control solution with predicted maintenance right again using the Cortex M85 core with predicted maintenance algorithm in there to identify the motor turns and there's an unbalanced load or a dead bearing or whatever in that kind of motor you can recognize that and then based on that information you can call maintenance crews come along to replace that unit in the field of course I mean to be honest this is not the typical motor you would expecting for predictive maintenance right this is definitely not the motor because it's a too small motor but that's the way it is for demonstration at Embedded World Shadow Boost you don't have the space to fit a large industrial motor in here right so would you say this is the the star of the show what do you mean the M85 yeah I mean come on this is the only one on earth Embedded World 2023 the winner of new kit launch and if customers want to be a winner as well they have always the chance to come along to our booth check out what ramsus look like and check out what ramsus can do for them right so we are there to help you this device is going to be a market this year cool thanks a lot welcome is that on pause yeah