 Welcome to Toradex office here in Seattle. Today we will talk about Verdin, our new system on module. My name is Daniel Lang and I'm the CMO of Toradex. Today here with me I have Samuel Imgris. Hey Nikola, thanks for coming by. Yeah, my name is Samuel. I'm the CEO of Toradex and today we really talk a little bit about on the hardware side what we have new for you guys. So we've actually been working for the last new months on a new and third module family. So most of you guys are actually familiar with our colibri form factor. So if we quickly look back here we have this is a 15-year-old form factor that we're still maintaining today. Our most successful product family, the colibri family and then we introduced about six years ago the Apollis family. This is really a high performance arm system a module family and today we're going to talk a little bit about our new Verdin system a module family. So like all other product families from Toradex we're developing system of modules that are pin compatible within the family. So today we're introducing the first some of the Verdin system a module family and this one is based on the IMX 8M mini processor. We also going to be offering a custom variant based on the 8M nano which is a pin compatible processor. So typically on the SOM you'll have a processor, system memory, flash, all the power controls, wireless connectivity and typical some of the high-speed peripherals like a gigabit ethernet 5. But then what you usually need is a carrier board to really develop the custom part of your embedded system. But I'd like to let Brandon talk a little about really what we had in mind when we came up with our new Verdin family. Yeah so I'm Brandon Shibley CTO of Toradex and so when we designed Verdin you know what we were doing is bringing out some higher-speed interfaces than some of our previous products you know Clebris a 15 year old family. With Verdin we're bringing out interfaces that like such as USB 3.0, gigabit ethernet, PCI express, MIPI DSI and MIPI CSI and we're doing that on a higher density connector here 260 pins on on this connector and so this is the development board that we created. This is the board you'll work with when you're developing a product with our Verdin system on module. You can see it connects in just like this. This is allows you to access most of the pins on the connector through these headers as well as any the other interfaces through various connectors and transceivers. So we have things like you know the dual ethernet that we talked about. This is one ethernet provided by the SOM where there's a five provided on board. The second provided over RGMI with another five here so we have two ethernet that are possible here. We have PCI express here we're bringing out the MIPI DSI and many of the the new SoCs coming out are supporting MIPI DSI well whereas other display interfaces aren't so universally supported across all those SoC so we find that you know it's going to be easier to support many of those interfaces over MIPI DSI with various adapters and so that's what we're showing here is that provide adapters such as this which adapt from MIPI DSI to HDMI. We also have like CAN RS-45 or RS-232. Those are provided over the UARTs of the connector. We have native HDMI, MIPI CSI here, USB on the go and a variety of other interfaces. We have many type specific interfaces also where we've dedicated some of the pins of the module for interfaces which are really SOM specific so things that are not universally supported across all of the products and those are accessible over the connector here on the development board. So that's about everything we've done here with Verden. I'll pass it on to Daniel. Yeah so we talked a lot about hardware which is of course very important very important part to make your hardware and design easier but Toradex is not just about hardware we do a lot about software so if you get the Verden module you tap into the whole Toradex ecosystem so means you get that famous support you're used to and you get all the software so you get the Octo BSPs you also get our new Torizen easy to use Linux distribution for industrial application and that's supported on Verden from day one and with Torizen we have a whole ecosystem of partners such as Codesys, Crank, Qt and many more providing containers so it's very easy to get started very easy to install that it's designed for easy updateability today everybody's connected you need to be able to update your software easy on the system and so we do all of that. We also have an ecosystem of system integrators so if you can design your own carrier board we have a proven partner who specializes in doing own carrier boards do a design for you or we have other partners having off-the-shelf carrier boards ready when that comes to market where you can just plug in the module same thing on the software side we have system integrator you need cloud connectivity you need anything special they can help you get started fast and lower your your time to market. So there's a lot of customer feedback you've gotten to implement basically it's like the next gen of Toradex right? Yeah so before we do that we did really extensive market research we looked at a lot of different things we looked at different connector we took a very detailed look at board to board connector we look at soldable solution and and we really decide that's the best for our target market our customer need vibration proof they need cost optimized they need easy assembly so they don't have sometimes the technology to do a deal with extremely fine pitch all of that we included all of that we did a lot of research about which interfaces they need and that's basically the outcome of that and we did the same thing on the software side with Torizon so we are very confident that that's a good solution and we are also very welcome you know for input you know to add additional software features and and so on. So it's it's could would you say future proof also in terms of what the trends are? Yes it's very future proof was really designed with outlook as Colibri for example is around from 2005 and we're still supporting that we still bringing out a new products there and same for here so we already we will launch with with two SOCs as Samuel mentioned IMX 8M Mini and IMX 8M Nano and we already have a roadmap with with more chips coming out in the same form factor so you will get the same scalability you are used to from our other families. And with a new over-the-air updates and the system and the whole Torizon that's it's a new way of doing providing support. Exactly so it's future proof on the hardware side but it's definitely future proof on the software side with Torizon with a stronger focus on connected devices on security. We of course also will have a real-time patch as you're used to be you know you have heterogeneous multi-core here with M4 cores which be integrated we have tooling to make it easy to get started we have plugins for Visual Studio we support Visual Studio code we invest a lot in also bringing Python on board which we see some people are beginning to use in embedded industries and so on. So this is also gonna enable even more in the labs more stuff. Yes of course on the labs we continue to innovate we continue to try out new things we have tools like flash tool which gives you an estimate about your flash lifetime and a lot of other additional features we develop in labs. I can also we talked about our ecosystem which of course it's provider of additional feature value added but we also have a very close connection with our suppliers like NXP and we have here at today Jerome from NXP and you can introduce yourself and he will talk a little bit about what kind of problems we like to solve together and why Toradex works so close with NXP. Hi thanks for having me today I'm Jerome Shang with NXP I have responsibility for cloud partnership programs and so what here we're trying to solve our real customer problems in places where we are decentralizing typical cloud machine learning workloads down to the edge. So if you look at what has been done here with Toradex NXP is working with AWS to enable their operating system and runtime and NXP is working with Toradex to enable computer vision at the edge as we blend the three of them together we essentially have a way to manage the lifecycle and onboarding of the device you also have a way to train model deploy model and reiterate and so that typically is a task that was cloud-based in the past now moving to the edge so what we're solving here is that typical machine learning model in the cloud is as efficient as how much power you can throw at it machine learning model at the edge is as efficient as self-contained you can remain and how much can be executed on the device so SageMaker Neo which is in our case the ML framework we're using is especially efficient in the auto tuning feature that allows the Toradex model to run efficiently at the edge and this is where you see the strength and power of ARM based computing compared to non-cost effective x86 based platform to take advantage of the power cost and performance areas that are typical of ARM based computing right now. ARM based computing has a lot to do with efficiency right and you have all this power out there it's nice to use it to to find new ways to use all this the CPU power and do all this machine learning with it. So you will see in that specific context that the work done around the ARM compute library is bearing fruits in places where the typical execution provider or the actual runtime environment choose which function is best accelerated and or executed either on the CPU or the GPU and so that's the specialty of SageMaker Neo with this auto tuning feature that allows us to choose which is the best function run on which accelerator. Many people in this industry choose NXP for the open source the availability of lots of open source stuff around it right it's it's one of the important things. Correct so if you look at the reason why people would go to NXP it's first and foremost because of the breadth of the portfolio because you can find the right devices that meets the performance and cost of your specific application. Then when you look at what we upstream to the open source community you'll find all the plumbing to actually accelerate edge-based machine learning and the different partners we have in the ecosystem to actually accelerate the development for end-users. It's here again about easing the customer journey in their development and the development with Toradex is definitely speeding up time to market. And you even have sometimes some of the GPU in the open source support and stuff like that when people want to do that or it's something that's in development a lot right. Correct so we're currently trying to balance GPU workloads, Norant Accelerator workloads and so the work we're doing with our partners is enabling open VX APIs for that matter and so you will see a whole fleet of product coming with Norant Accelerators that we introduced at CES this year. Nice and this enables so many new applications so many new startups should just be watching and making stuff making new new ideas. Our specialty is bringing workloads at the edge so you will hear NXP being very efficient at the edge in places where we're trying to get you to more cost-effective solution than typical x86 architecture in order for you to start balancing workloads in order to actually distribute those end nodes, manage your life cycle in a secure fashion. Correct. Nice so you are definitely getting more and more busy with all these solutions right? Yeah it's going very well we see you know a lot developing as I said and he points out you know we're moving into edge compute machine learning is a big topic so it's open like embedded vision so there's a lot of new opportunities coming of course the whole IoT the whole connectivity a lot of new challenges Vitoradex can help you solve and bring you faster to market.