 Hello everyone, welcome to theCUBE's live here in Dallas, Texas. I'm John Furrier, host of theCUBE, three days of wall-to-wall coverage. Of course, we've got the three fabulous guests here, myself, Savannah Peterson. Savannah, you look wonderful. Thank you, John. I feel lucky to play the part here with my 10 gallon hat. Dave Nicholson, who's the analyst uncovering all the Dell, super computing, HPE, all the technology has changed in the game. Dave, you look great. Thanks for coming on. Thanks, John, I appreciate it. All right, so you look good. So we're in Dallas, Texas. It's a trade show conference. I don't know what you'd call this these days, but thousands of booths are here. What's to take here? Why super computing 22? What's the big deal? Well, the big deal is dramatic incremental progress in terms of super computing capability. So what this conference represents is the leading edge in what IT can deliver to the world. We're talking about scale that is impossible to comprehend with the human brain, but you can toss out facts and figures like performance measured in exoflops, millions of CPU cores working together, thousands of kilowatts of power required to power these systems. And I think what makes this show unique is that it's not just a bunch of vendors, but it's academia. It's PhD candidates coming and looking for companies that they might work with. So it's a very, very different vibe here. Savannah, we were talking last night, but before we were setting up our agenda for what to drill down on this week. And you were, by the way, that looks great. I wish I had one. We'll get you one by the end of the show, John. Texas is always big in Texas. And that's the thing here. But super computing seems like that had a lull for a while. It seems like it's going to explode. And you had a chance to review the papers, take a look at it. I won't say closet hardware nerd, but that's your roots. Yeah, yeah, very openly hardware nerd. And I'm excited because we saw a lot of hype around quantum and around AI five, 10 years ago, but we weren't seeing the application at scale. And we also weren't seeing, quite frankly, the hardware wasn't ready to power these types of endeavors at scale. Whereas now, we've got air cooling, we've got liquid cooling, we've got multiple GPUs. Dell was just showing me all eight of theirs that they put in their beautiful million dollar piece of equipment, which is extremely impressive for folks to run complex calculations. And what I'm excited about with all the, I love when we fuse business and academia together. I think that that doesn't happen very often. I've been impressed. I mean, when I walked in today right away, I'm sure you all can't see this at home just yet, but we'll try and give you a feel over the course of the next two days. This conference is huge. This is way bigger than I was expecting. You know, a lot larger than where we just were in Detroit. And I love it because we've got the people that are literally inventing the calculations that will determine a lot of our future from sequencing our genome to powering our weather forecasting, as well as all of the companies that create the hardware and the software that's going to actually support that, those algorithms and those. And the science and the engineering involved has just been going on since 1988. This conference, this trade show, is going on since 1988, which is, it passed the test of time and now the future with all the new use cases emerging from the compute and supercomputing architectures out there. It's from cradle to grave. If you're in this business, you're in school all the way through the industry. It doesn't seem to stop. That university student side of it. We had a whole student section here. So you don't see that very often in some of these tech shows. It's like from students to boardroom. I actually brought the supercomputer from 1988 with me in my pocket. And I'm not sure that I'm even joking. This may have as much processing power, certainly as much storage with one terabyte on board. I sprung for the one terabyte folks. But it is mind-boggling the amount of compute power we're talking about. When you dig below the surface, which we'll be doing in the coming days, you see things like leaping from PCI-E gen four to gen five. And the increase that that gives us in terms of capabilities for plugging into the motherboard and accessing the CPU complex and on and on and on. But something Savannah alluded to. We're talking about the leading edge of what is possible from a humanity perspective. And so I'd like to get into, as we're talking to some of the experts that we'll get a chance to talk to you, I'd like to get their view on what the future holds and whether we can simply grow through quantitative increases in compute power or if the real promise is out there in the land of quantum computing. Are we all sort of hanging our hats, our large 10 gallon hats? Yes, our hats. If we're hanging our hats on that, that that's when truly we'll be able to tease insight out of chaos. I'd like to hear from some of the real experts on that subject. I'm glad you brought that up because I'm personally pretty pumped about quantum computing, but I've seen it sit in this hype stage for quite a while and I'm ready for the application. So I'm curious to hear what our experts are. That's an awesome, I think that would be an awesome bumper sticker, frankly Savannah. I'm pumped about quantum computing. Who is this person? Who is this person? I want to see it first. Did someone show me it? Yeah. 400 qubits I think was the latest IBM announcement, which means something I'll pretend like I completely understand what it means. So that's what that means, David. Well, so Savannah, let me mansplain it to you. Yeah, let's hear it. So basically it's, when conventional computing, you can either be on or off, zero or one. In quantum computing, you can be both, neither or all of the above. That's the depth to which I can go. I like that. It actually is distinct as humanly possible, really. Sounds like a Ponzi scheme to me. I'm not sure if I... Well, let's get into some of the thoughts that you guys have on some of the papers we saw, Savannah, and Dave, your perspective on this whole next level kind of expansion of what super computing and super cloud and super apps will do for this next gen. What use cases are kind of shining out of this? Because, you know, it used to be, you were limited by how much gear you had stacked up, how big the server could be, the supercomputer. Now you got large scale cloud computing. You got the ability to have different subsystems like advances in networking. So you're seeing a new architectural, almost bigger super computing. Isn't it just a machine? It's a collection of machines. It's a collection of other stuff. What's your thoughts on these, this architecture and then the use cases that are going to emerge that were not gettable before? So in the past, you talk about, you know, 1988 and, you know, I'd say a decade ago, the race was to assemble enough compute power to be able to do things quickly enough to be practical. So we knew that if we applied software to hardware, we could get an answer to a problem because we were asking very, very specific questions and how quickly we got the answer would determine whether it was practical to pursue it or not. So if something took a day instead of a month, okay, fantastic. But now we've reached this critical mass. You could argue when that happened, but definitely I think we're there, where things like artificial intelligence and machine learning are the core of what we're doing. We're not just simply asking systems to deliver defined answers. We're asking them to learn from their experiences. Starts getting a little spooky. And we're asking them to tease insights out in a way that we haven't figured out. So we're saying, give us the insight. We're not telling the system specifically how to give us that insight. So I think that's the fundamental difference. That's the frontier is, you're gonna hear a lot about AI and ML. And then if you retreat back a bit from supercomputing, you're in the realm of high performance computing, which is sort of junior version of supercomputing. It's instead of the billion dollar system, it's the system that, you know, schlubs like Facebook or AWS might be able to afford. You know, maybe a hundred million dollars for a system. Casual. Just sort of casual kind of thing next to the coffee table in the living room. But I think that's really gonna be the talk. So that's a huge tent when you talk about AI and ML. Yeah, I totally agree. We're having some of the conversations that we've had for a long time about AI and bias. I saw a lot of the papers were looking at that. I think that's what's gonna be really interesting. To me, what's most exciting about this is how are we pulling together all of this on a global scale? So I'm excited to see how supercomputing impacts climate change. Our ability to monitor environmental conditions around the globe and different governments and bodies can all combine and all of this information can be going into a central brain and learning from it and figuring out how we can make the world a better place. We're learning about the body. There's a lot of people doing molecular biology and sequencing of the genome here. We've got, it's just, it's very, I don't think a lot of people realize that supercomputing pretty much touches every aspect of our lives. I mean, we've had it for a while. I think cloud computing took a lot of the attention given that that brought in massive capabilities, a lot of agility. And I think what's interesting here at this show, if you look at what's going on from the, like I said, from the dorm room to the board room, everyone's here. But you look at what's actually going on above the hardware. CNCF is here, they have a booth, the whole cloud native software business. It's going to be interesting to see how the software business takes advantage of how these architectures, because let's face it, I've never heard a developer say, I want to run on slower hardware. So no one wants that. So now, if you abstract away the hardware, as we know with cloud computing and DevOps, cloud on premises and edge, David, this is like, this is again, nirvana for the industry because you want the fastest possible compute system for the software. At the end of the day, that's what we're talking about. So I asked the gift question to my Wharton students this morning on a call. And I asked specifically if I could give you something that was the result of supercomputing's amazing nature. What would it be? Would it be personalized therapeutics in healthcare? Would it be something related to climate? Being able to figure out exactly what we can do. There's a whole range of possibilities and what's interesting is... What were some of the answers? So a lot of the answers came down to two categories and it was really, it was healthcare and climate. A lot of understanding. And of course, a lot of jokes about how eventually supercomputers will determine that the problem is people. Is people, yeah, I knew you were headed there. Don't people just fly custom jeans? Yeah, or, well, so one of the good ones, so it was also that while we're here. A person from a company who shall not be named said, oh, advertising. It was the, what if you could predict with a high degree of certainty that when you sent someone an email saying, hey, do you want to buy this? They would say, well, yeah, I do. Dramatically lowering the cost of acquisition for an individual customer as an example. Those are the kinds of breakthroughs that will transform how we live because all of a sudden industries are completely disrupted. Not necessarily directly related to supercomputing, but you think about automating the entire fleet of trucks in North America. What does that do to people who currently drive those trucks? So there are societal questions at hand that I don't necessarily know the academics are considering when they're thinking what's possible. Well, I think the point about the ad thing brings up the whole cultural shift that's going on from the old generation of, hey, let's use our best minds in the industry to figure out how to place an ad at the right place in the right pixel at the right time versus solving real problems like climate change, our culture and society and us getting along as a country and world. Water, sustainability, fires in California. Yeah. I mean, come on. So I got to say, I was curious when you were playing with your pocket computer there and talking about the terabyte that you have inside. So back in 1988, when supercomputing started, the first show was in Orlando. It was actually the same four days that we're here right now. I was born in 1988 if we're just talking about how great 1988 is and so I guess I was. So were we Savannah, so were we. I think I was in third grade at that time. We won't say what you told me earlier about 1988 for you. But that said, so 1988 was when Steve Jobs released the next computer. He was out of Apple at that time. Yeah, that's right. Eight megabytes of RAM. It's called theCUBE, I think. It's respectable. That's all. It was, it was theCUBE, which is pretty exciting. But when we're looking at, yeah, on the supercomputing side, your phone would have been about as capable. So where will we be in 20 years? It's amazing. Will our holograms be here instead of us physically sitting at the table? I don't know. Well, it's going to be very interesting to see how the global ecosystem evolves. It used to be very nationalistic culture with computing. I think we're going to see global flattening of culture relative to computing. I think space will be a massive, massive discussion. I think software and automation will be at levels we don't even see. So I think software, to me, I'm looking at that's the enablement of this supercomputing show in terms of the next five years. What are they going to do to enable more faster, intelligent horsepower? And what does that look like? Is it, it used to be simple. Processor, more processors, more threads, multi-cores, and then stuff around it. I think this is where I think it's going to shift to more network computing, network processing, edge, latency, physics that's involved. I mean, everything you can squeeze out of the physics will be interesting to watch. Well, when we peel back the cover on the actual pieces of hardware that are driving this revolution, parallelizing of workloads is critical to this. It's what supercomputing consists of. There's no such thing as a supercomputer sitting by itself on a table, even the million dollar system from Dell, which is crazy when you hear Dell and a million dollar system. And it's chilling there too. Right, just hanging out. But it's all about the interconnect. When you want to take advantage of parallel processing, you have to have software that can leverage all of the resources, and connectivity becomes increasingly important. I think that's going to be a thread that we're going to see throughout the next few days, with the motherboards, for lack of a better term, allowing faster access to memory, faster access to CPU, GPU, DPU, networking, storage devices, plugging in, those all work together. But increasingly, it's that connectivity layer that's critically important. Questions of Infiniband versus Ethernet. RDMA overconverged Ethernet as an example. A lot of these architectural decisions are going to be based on power, cooling, density. So a lot of details behind the scenes to make the magic happen. I think the power is going to be thinking 20 years out. Hopefully everything here is powered sustainably 20 years from now. Because power pull, the more exciting things going on in your supercomputer, the power suck is massive. That when we were talking to Dell, they were saying that's one of the biggest problems. Biggest concerns for their customers. And that's going to play into sustainability. So we've got a lot of great guests. We've got folks from Dell in the industry, a lot of the manufacturers, a lot of the hardware, software. Experts going to come on and share what's going on. We did a post, Why Hardware Matters, a few months ago, Dave. Everyone's like, well it does, now more than ever. So we're going to get into it here at Supercomputing 22, where the hardware matters. Faster power as we say for the applications. Mr. Cube, we'll be back with more live coverage. Stay with us.