 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 right 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 the ST microcontrollers which were announced also renaissance is showing off the Cortex M85 which is one of the the highest performance M-class CPU we have to offer so really pushing the boundaries of performance for for machine learning for signal processing but also for for regular scale of 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 right driving more compute capabilities at the edge right there is certainly machine learning as I said we find demos around voice activated door locks for example which is quite quite interesting we find demos around connected motor control integrating machine learning into that for predictive maintenance but also 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 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 in in appliances for example so that's certainly where GPUs are now extending into the microcontroller world as well very excited about that. Cortex M with AI. Cortex M with AI absolutely as well yes so 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 partners as we have over there if we want to walk over there is that some of our partners like Hayamex they integrate the Cortex M55 which was the previous recently announced and integrating our EPOS U55 which is a neural network processor processing unit so you can see here the demo with the camera. So what is the ARM processor? It's a Cortex M55 with the EPOS combined with the EPOS U55 so really targeting machine learning in a tiny application that the endpoint essentially. And it all goes in one SSE? It all goes in one SSE 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 it's in a Cortex A with all kinds of stuff now it's also in embedded. Everything is running at the edge and the end point you can do all the inferencing at the end point with 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 to so we have something we started last year is initiated is a project called ARM Virtual Hardware and the idea 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 board so if you think about in the past right 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 board or the hardware section 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 the in-house for for decades which we're using actually for our modeling of the of SOCs and we have another solution where we partner with another company where we're providing basically a modeling of full physical board so you could actually start develop on a virtual Raspberry Pi for example. Is it FPGA stuff or it could be different? What it needs to run on? It basically runs in the cloud so you have you log into AWS essentially it's running on AWS instances in the cloud so there's no physical board there's nothing involved in that. So if you go around there's something else I wanted to show you it's really about as we talk about the edge right there so many different use cases 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 a really machine learning or the machine vision running on Cortex-A based devices and what we see especially in the A-class word is that often you have platform or software development for each individual platform which often as we look at that we see that that they are things which which our partners or the developers have to do for each individual platform but there's opportunity really for reuse right 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 how can we what can we standardize in the software development floor or in the software stack so that 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 Pine 4 Arduino AD-Link and Contra and 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 the layer between firm and the operating system so standardizing that that layer essentially so that any operating system can land on any any system ready certified hardware nice so interoperability yes exactly so that what that means right is really you can run any operating system on any board so you can the demo here shows that you could use an USB stick with for example Fedora or SUSE you plug it into that that board it basically puts the system can use the same USB stick to to put up your Raspberry Pi for example so this is really where we see a lot of value in this 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 your essentially your Linux image on it's one of the arguments of the arm right is to be interoperable and once you have software on one you can do the other but now this is the next level 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 yeah absolutely I think it's about the developer experience right 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 is the 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 the system ready standard to really ensure that any operating system can land on any arm system ready so at the 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 right yeah now it's super exciting to be here it's very very busy it's great to see the industry being together again and great to see so much stuff on arm