 So here at the Lunara Connect, you have the new Ultra 96 development board and who are you? Hi, I'm Glenn Steiner. I'm with Silencs. And yes, I'd like to introduce the new Ultra 96 development platform. We have integrated ARM Cortex A53 processors, RFI processors, and then our magic sauce, which is the ProgrammaLogic. ProgrammaLogic allows you to take your C code and move it over to hardware and dramatically accelerate that code. And we have a couple demonstrations for you today. The first of which is our tiny yellow demonstration. As you can see in this particular demonstration, we are doing person recognition as well as other object type of recognition. And this again was taken as C code, moved over to hardware using our CDHDL tools, SDSOC, and demonstrates dramatic software acceleration. In this case, hundreds to thousands of times faster. Hundreds to thousands of times faster than running on the A53. Than running on an ARM. Yes. So up to thousands of times faster. Yes. And what is this hybrid Q&N? Is it one way of doing machine learning? We're actually demonstrating here different techniques for machine learning algorithms. The first which is the tiny yellow, which means you only look once. The second demonstration that we have for machine learning is a CNN neural network. And in this particular case, neural networks, you can go and you can have floating point, 32-bit, 16-bit, 8-bit. We've actually, in our research labs, have moved from down to 4-bits, 2-bits. And in this particular implementation, we're doing a 1-bit neural network. What is a 1-bit? Which means that rather than using, let's say, 8-bits of precision, you're down to 1-bit of precision. And you would think you would lose accuracy. And that is true. You do lose accuracy when you're talking about the same number of neurons. But with a modest increase in the number of neurons, you're able to get the same amount of accuracy, but a dramatic reduction in the amount of logic and the amount of power that is consumed. In this particular demonstration, we're showing a recognition of traffic signs. And as you can see, we went out in the field and had 300 different signs collected. The database was trained on a German traffic sign database. And if you look, you will see that we are actually looking at images. And here, you're seeing 13,000 images that are being recognized per second compared to if you're doing it in software, it would be 2.17 images per second, which means that our overall acceleration is 6,200 times faster than if you were doing it in software on an A53. So there's going to be lots of FPGA in the AI of the future? We certainly believe that. And that's one of the ways you are releasing your A6 boards. Do you want to get the open source community to embrace more FPGAs with the Linux and all that stuff? Absolutely. We developed the Ultra96 board. It was actually an internal development. We fielded about 1,000 of these boards to our engineers within the organization. People got excited about it, said, you know what? Let's go ahead and roll this out publicly, make it available to everyone so that they can start doing their own development. And then finally, I'd like to introduce our last demonstration, which is the package that you can receive. Initially, when you do the purchase, you do receive just the board, as you saw earlier. And under this black part here, is it for cooling of the FPGA? Yes. Under the black part, we do have a small fan. It blows air directly onto the MPSOC device. Because it runs very fast? It runs very fast, but it only consumes a few watts, depends on what you're doing. And for our demonstration, we have added the Grove Starter Kit, which plugs directly on top of the board. And then in order to provide an exciting out-of-box experience, we actually boot up to Linux. We have the Enlightenment Desktop. We bring up a HTML-based window where you can go... Kind of like a Chrome browser? Or many Firefox? Yeah. You honestly don't switch browser. But probably, you're correct, probably something like Firefox. And from the browser window, you're able to select particular applications that you might want to go and work with. And then you can see the design. You can quickly click Run Project. Which will then run the code. In this particular case, you can see it's just cycling through very quickly with different colors and descriptions on the board. And then you're actually able to then open up an HTML editor, see the code, modify the code, and create your own software applications. So you have a very nice out-of-box experience. Get you quickly started with the Zinc Ultra Scale Plus MPSOC and the Ultra 96 development platform. So with the Ultra 96, hopefully, you hope that there's going to be a lot more developers working on your FPGAs? That is the objective. And they will learn how to develop on FPGAs, how to optimize code and all that. And you can use your tools. Yes. With the kit, you have access to our Vavado design tools. So initially, you'll want to get out, get started with the quick introduction we have to our developing code. That code is actually run natively and compiled natively on the development platform. As you move on, you then progress into our Vavado development tools using our software development kit, which includes the Eclipse environment. And from there, once you identify some code that might not be running fast enough because you've profiled it, you then use our SDSOC tools, which is included as a coupon, free coupon upgrade. And using the SDSOC tools, you can take the code, convert it over to a hardware description language, automatically replaces your software function with a call to the hardware, and you can accelerate your software. So you can just program in C or C++ or something like that. That is correct. And then it just translated into FPGA language. That is correct. You develop your code in C or C++, and then the SDSOC tool will translate it over to HDL. And the board is $249, available very soon. The cost for the board is $249, available about a month from now. So that's a great deal, right, to have access to your FPGA, which is very cutting-edge technology, right, which is stuff that goes into, maybe, I don't know, rockets or something. Yes, you actually are right. We have been on a variety of different technologies. Our products go into all market segments. We help power the internet infrastructure. We go into automotive applications, medical applications, industrial applications. You're talking about rockets. We were on the Mars lander. Nice. And it's been running on Mars for, like, seven years now or something. Yes. So that means you have high, low-temperature support, all that kind of stuff. Well, we have parts that are designed to go into space applications. Nice. And there's a new announcement with the 7nm, so there's a big future with silings with all these very advanced FPGAs that are coming soon. So if you start working on this, you should be ready for your next stuff, too. That's true. We're fully committed to the ARM ecosystem and the Lonaro ecosystem. We've moved through our ARM Cortex A9 generation. We're now at the A53 generation, and then with our 7nm announcement, we will be moving to another generation of ARM technology.