 Hi, my name is Richard Riegrach from the Qualcomm Data Technology Center Engineering, and this is the reference evaluation platform. It has two CRB systems that you just saw earlier, mounted in a 1U form factor. Right there in the chipsets. These are the two chipsets mounted here. A whole bunch of RAM. That's right. Six channels on each side, a dual channel that you are, and network interface here, we're at Melanox booth. So we're here with Melanox. Right, and here's the CX-5 card. And uh, Melanox, right? So who are you? So I'm Scott Schultz. I manage a lot of the HPC and artificial intelligence activities here at Melanox. And you work together with Qualcomm to do this solution right here, right? Absolutely. So, you know, the industry is really excited about ARM and it's actually becoming a potential solution for a lot of HPC applications. And of course, it's no different than any other platform that would require a CPU offload architecture from a network perspective. So right here you have your network card to make super fast networking? Exactly. And so this is the card, you know, kind of outside the box. It definitely is, it's a ConnectX 5. It's a dual port card. This is 100 gigabits a second, right? But it'll support Gen3 as well as Gen4. It's a by-16 card. It has all of the collectives, tag matching, and MPI offload capabilities to remove latency from rendezvous type operations. We have support for in-network computing and all of the storage for NVMe offloads and whatnot. So this is the de facto standard for HPC and machine learning. This is the same card that goes into the DOE systems that were just recently announced that we're all excited and waiting for. But again, you know, there's a lot of interest across the industry for very high performance ARM solutions. And we feel that the ConnectX 5 paired up with Qualcomm is a very exciting solution. So how's the performance for the solution? It's huge. It's a rock star. What are you talking about? It's a rock star. So because you're 10 nanometers, right? People are very excited about this. It's very high performance. Absolutely. The performance is really good. We've got performance parity and we've got performance advantage. But our story is a TCO story. Performance per watt is extremely good. And Qualcomm has been working for many generations of custom for smart phones, right? Absolutely. So this is knowledge of doing custom that helps you with this one too. That's correct. It's a fifth generation micro architecture from Qualcomm. Integrated and all the experience that we've learned about designing for low power is integrated into this chip. And that's why a lot of our differentiation for the performance per watt is very significant. And it's been a smooth ride so thus far to work on software. Everything to work together in the ARM ecosystem. Absolutely. We've actually at Melanox we've been working with ARM technology for quite some time. And we've been enabled and ready with all of the core offload capabilities for several years actually. All the previous generations of different providers? Yeah, right, exactly. So ARM technology has come a long way in the last several years, right? So I feel personally, right, that this is going to be a very compelling solution, especially when you couple it with the latest and greatest interconnect technology.