>>from around the globe. It's the Cube with digital coverage of AWS >>reinvent 2020 sponsored >>by Intel, AWS and our community partners. Welcome back to the cubes. Virtual coverage of AWS reinvent 2020. It's virtual this year. We're not in person, so we're doing remote interviews. Part of the three weeks we'll be covering wall to wall a lot of great conversations. News to cover and joining me today Off Fresh off the news off Andy Jackson's keynote, We have two great guests here. Jason Kapoor, senior product manager for Accelerated Computing at A. W S and eight time Medina Chief business officer, Havana Labs, which was recently acquired by Intel Folks. Thanks for coming on, gentlemen. Thank you for spending the time for coming on the key. Appreciate it. >>Thanks for having us. >>J Town. So talk about the news, actually. Uh, computers changing. It's being reinvented. That's the theme from Andy's keynote. What did Andy announced? Could you take a minute to explain the announcement? What services? What ap What's gonna be supported? What's this about? Take a minute to explain. >>Yeah, absolutely. Yeah. So today >>we >>announced our plans to launch and easy to instance based on hardware accelerators from Havana labs. We expect these businesses to be available in the first time from next year. And these air custom designed for accelerating training off deep learning models, a zoo we all know like training of deep learning models is a really competition. Aly extensive task. Oftentimes it takes too long and cost too much. And we're really excited about getting these instances out of the market as we expect for them to provide up to 40% better price performance. Thani on top of the line GPU instances, >>a lot of improvements. Why did anybody do this? Why heaven or what's the what the working backwards document tell you? What is it customers looking for here is or specific use case? >>Yeah, absolutely. So, you know, over the years, uh, the use of machine learning and deep learning has, like, really skyrocketed, right? So we're seeing companies from all the way like 14, 500 to like start ups just reinventing their business models and use using deep learning more pervasively. Right. So we have companies like Pinterest, you know, you'd use deep learning for content recommendations and object detection to Toyota Research Institute that are advancing the science behind autonomous vehicles. And there's a consistent cream from a lot of these customers that are, you know, innovating in the deep learning space that you know the cost it takes to experiment, train and optimize the deep learning models. It's too high. And, you know, they're looking at us as one of their partners to help them optimize their costs, you know, bring them as well as possible while giving them really performing products and enable them to actually bring their markets, their innovations to market as soon as possible. Right? S o. Do you answer your questions straight on your wants? The working backwards. It's a feedback from customers that they want choice on. They want our help Thio lower. Uh, the amount of compute resources and the cost it takes to train the new planning models. >>Hey, Tom, why don't you weigh in here on Havana and now part of intel? What trends are driving this? What's the motivation? Were you guys fit in? What's your view on this? >>Yeah, So Havana was founded in 2016 to deliver a I processors for the data center and cloud for training and inference deep learning models. So while building chips is hard, building, the software and ecosystem is even harder. So joining forces with intel simply helps us connect the dots. Ever since the acquisition last year, we were able to significantly boost our armed. The resource is, and now we're leveraging inter scale in number of customers and ecosystem and partner support. >>So what's the name of the product? Is there a chip name got? Was it Gowdy is the name? >>Yes, the product is man angle. >>Okay. And so it's gonna be hardware. So it's the hardware software. What's involved? Take us through the product. >>Yes. So Gandhi was designed from the ground up to do one task which is training deep learning models. To do that well, we focus the architectural to aspect efficiency and scalability. The computer architectures is a combination of fully programmable TPC tensile process, of course, and a central g M n G. These DPC course are programmable Villa W seen the machines that we designed with custom instruction, set architecture, er and special functions that will developed specifically for a I. The Gandhi cheap integrates also 32 gigabyte off H B M to memory which makes it easy to port to. For GPU developers, Gandhi is unique in integrating 10 parts of 100 gigabit Internet rocky on cheap. And this is opposed to other architectural, which use proprietary interfaces. So overall, improving the cost performance is achieved through efficiency, namely higher utilization off the computer and memory resource is on cheap and the native integration off the rocky interfaces >>J Town. This is actually interesting, as this is the theme for reinvent. We're seeing it right on stage today. Play out again another command performance by Andy Jassy. Slew of announcements. How does Gowdy fit into the AI portfolio or Amazon strategy? Because what a town saying is it sounds like he's doing the heavy lifting on all this training stuff when people want to just get to the outcome. I mean, the theme has been, just let the product do what they do kind of put stuff under the covers and just let it scale. Is that the theme here is this. >>What does this >>all fit in? Take us through how this fits into the A, I strategy for Amazon and also what what what is Havana Intel bring to the table? >>Absolutely. Yeah. So with respect to our overall strategy and portfolio units, it's relatively straightforward, right? So we're laser focused on making sure we have the broadest and deepest portfolio off services for machine learning, right? So these range from infrastructure services specifically compute networking and storage all the way up to, like, managed and all services, which come with pre trained models and customers can simply invoke them using an A P. I call right eso. So from a strategy perspective, you want to make sure that we provide a customer to a choice, uh, enable them to pick the right platform for the right use case, help them get to the Khan structure they actually want, right eso with Havana. And you know, their acquisition with Intel, we finally have access to hardware software and the ability to kind of build out a ecosystem beyond what you know judicially is being used. Which is was a GP used right eso. So the engagement with with Havana, you know, allows us to take their products and capabilities, wrap it around, and easy to instance, which is what customers will be able to launch right on doing so. We're enabling them to tap into the innovation that Teton the rest of the Havana team are working on while having a solution that is integrated with the full AWS stack. Right? So you don't you don't have to rack in stock hard. Bring your data center thes. They're gonna be available standard. Easy to instances. You can just click and launch them. Get access to software that's already pre integrated and big den and ready to go right. Eso so it actually comes down to taking their innovations, coupling it with an AWS solution and making it too easy for customers together. I've been running with the respective training the deepened models. >>Well, here is the question that I want to get to. I think everyone's on everyone's mind is how is it Gowdy different or similar than other GPU? Specifically, you mentioned the software stack on the AWS What you get the software stack inside the chip. How is this different or similar? Two other GP use. And what's the difference between the software stack versus a traditional libraries? >>So from day one, we were focused on the software experience and we were mindful in the need to make it easy for developers to use the innovations we have in the hardware. Most developers, if not all of them, are using deep learning frameworks such as tensile flowing pytorch for building their deep learning models. So God is synapse AI software suite comes integrated and optimized for tensorflow and pilotage, so we expect most developers to be able to take their existing models and with minor changes to the training strips to be able to run them on Gowdy based instances. In addition, expert developers that are familiar with writing their own kernels will be provided with food too sweet for writing their own TPC kernels that can augment the Havana provided library. >>So that's the user experience for the developers, right? That's what you're saying >>exactly, exactly, and we will provide detailed guides for developers. In doing that, Havana will provide open access to documentation library software models and left toe Havana's kita and bi directional communication with the Havana developer community. All these resources will be available concurrently with the AWS Instances launch. >>Okay, so I'm a developer. How did I get involved? It's software on git hub I use the hardware is on Amazon, obviously, in their instances. It's a new instance. Take me through the workflow develop. I'm into this. I wanna I wanna get involved. What I what am I doing? Take me >>through? Yes, I think it s so If the developer is accustomed to using GPS for training the deep learning models three experience is gonna be practically the same, right? So they'll have multiple options to get started. One of them would be, for example, to take our deep learning, Um, it's or Amazon machine images that will come integrated with software from Havana labs. Right. So customers will take the deep Learning Army and launch it on an easy to instance, featuring the gaudy accelerators. Right? So when with that, they'll have, you know, the baseline construct off software and hardware available to get up and running with right, we'll support, you know, all different types of work flows. So if customers want to use containerized solutions, thes instances will be supported R E C s and E s services. Eso using containerized kubernetes you know, thes the solution will just work on. Lastly, we also intend to support these instances through sage maker eso. Just a quick recap on stage maker. That's a manage service that does end to end that provides end to end capabilities for training, debugging, building and deploying machine learning applications. Eso these instances will also be supporting sage maker. So if you're fiddling with sage maker, you can get up and running with this. This is fairly quickly. >>It sounds like it's gonna enable a lot of action and sage maker level. Then can that layer on the use cases? I gotta ask you guys quickly, What's the low hanging fruit use case applications for this product thing? This partnership, Because you know that's gonna be the first Traction said, What are some of these applications gonna be used for? What can we expect to see? >>So typical applications would be image classifications, object detection, natural language processing, the recommendation systems. You'll find reference models in our get up for that and will be growing at least a Z you can imagine. >>Okay, where can people find more info? Give us the data. Take him in to explain. Put a plug in for how What's all the coordinates? U r l sites support how people create, Um, how people get involved. The community. >>Yeah, so customers will be able to access information on AWS websites and also on Havana Labs website. So you will be kicking off a preview early next year. Eso I would highly recommend for customers to find our product pages and signed up for already access and previous information. Utah. >>Yes, and you'll find more information on Havana. A swell a Savannah's get up over time. >>Great announcement. Congratulations. Thanks for sharing the news and some commentary on it. This is really the big theme. You know what Cove in 19 and this pandemic has shown is massive acceleration of digital transformation and having the software and hardware out there that accelerates the heavy lifting and creates value around the data. Super valuable. Thanks for for doing that. Appreciate taking the time. Thank >>you so much. >>Yeah. Thanks for having >>us. Okay, this is the cubes coverage at 80. Best reinvent next three weeks. We're here on the ground. Will remote. We're live inside the studio. We wish we could be there in person, but it's remote this year. But stay tuned. Check out silicon angle dot com. Exclusive interviews with Andy Jassy and Amazon executives and the big news covering. They're all there in one spot. Check it out. We'll be back with more coverage after this break. Thanks for watching. Yeah.