Reimagining Cloud Performance using AMD on AWS with Mike Thompson at the NY Summit

August 8, 2023

Episode Summary

#awscloud #cloudcost #costoptimization #amd #finops

In this episode of the Jon Myer Podcast, the host, Jon Myer, engages in a conversation with Mike Thompson, a senior product manager at AMD. The episode centers around AMD's presence at the New York Summit and its success in the cloud computing realm. Mike's participation at the summit is motivated by his desire to connect with the public cloud user community and elucidate the factors behind AMD's achievements. These achievements are primarily attributed to the company's ability to offer faster performance at a lower cost and increase computing capacity without inflating expenses.

The discussion also delves into the strategic partnership between AMD and AWS. Mike underscores the close collaboration between the two entities and highlights AMD's fourth-generation Epic processor, known as Genoa, which has been embraced by AWS across diverse instances worldwide. The collaboration has led to significant enhancements in performance, exemplified by the Elastic MapReduce (EMR) team's utilization of AMD Epic processors to enhance EMR Spark price performance by 50%, essentially doubling productivity within the same cost constraints.

The conversation shifts to the challenge of AMD's visibility relative to other semiconductor companies. Mike attributes this lower recognition to the fact that cloud users often overlook the processor's role in their buying decisions. In the context of cloud computing, users often prioritize metrics such as core count and performance, inadvertently overlooking the pivotal role that processor selection plays in overall efficiency and cost-effectiveness.

The advantages of running workloads on AMD processors are also explored. Mike highlights three key reasons: the potential to gain more computing capacity at a lower cost, the ability to expand capacity without substantial additional funding, and the improved security facilitated by AMD's Infinity Guard feature, which safeguards data in use. These benefits underscore the value proposition that AMD offers to customers seeking performance and security in their cloud computing endeavors.

The episode concludes with Mike offering insights into AMD's future roadmap. He mentions the forthcoming release of the fourth-generation Epic processor, Genoa, and its preview by AWS. Additionally, Mike introduces AMD's MI 300 GPUs, emphasizing their potential to revolutionize generative AI applications. These GPUs boast integrated memory that minimizes data movement and subsequently enhances the efficiency of data-intensive applications.

Throughout the conversation, Mike's anticipation for sharing more about AMD's offerings at the AWS New York Summit is palpable. The episode provides listeners with valuable insights into AMD's role in cloud computing, its strategic collaborations, and its commitment to delivering performance and security to users.

Are you looking to elevate your brand through compelling and highly professional video content? Myer Media has the solution for you, from start to finish we offer complete video creation services that include Customer Case Studies, Creative Content, Podcasting, and more.

To learn more CLICK HERE:

Headshot

About the Guest

Mike Thompson

Mike Thompson loves board sports, driving the next best things in tech, and mentoring kids to develop strong character through sports and iterative design. Now he’s Sr Product Manager in AMD’s Cloud Business Group, showing the world the glory of AMD-powered products in Amazon EC2 and AWS Services. He’s worked deep in Engineering, Sales, and Marketing across many markets. Feel free to reach out to him on LinkedIn

#aws #awscloud #finops #cloudcomputing #costoptimization

Episode Show Notes & Transcript

Host: Jon

Hi everybody, my name's Jon Myer. Joining me today is Mike Thompson, who's a senior product manager at AMD. Mike, thanks so much for joining me,

Guest: Mike

Jon. Thanks for having me. Appreciate you taking the time.

Host: Jon

Yeah, of course. So Mike, what brings AMD and you to the New York Summit this week?

Guest: Mike

Well, there are a couple of things. On a personal level, I've spent quite a while on the West Coast, and so I wanted to come to New York for the bagels and the pizza.

Host: Jon

Great. All right, let's go grab some.

Guest: Mike

Yeah, but a close second, maybe a one B if my boss is listening, is to come to talk to the public cloud user community about why AMD has had such success in the cloud over the last five or so years. And that boils down to two things. Faster performance at a lower cost, or more computing capacity at the same cost. So I've got two sessions that delve into that. One is a general purpose,  AMD in the public cloud, particularly at AWS, and the other one focuses more on the applications that the financial sector needs out of the public cloud. So two sessions. Well,

Host: Jon

Mike, can you tell us a little bit more about the partnership between AMD and AWS?

Guest: Mike

Yeah, it's a pretty tight collaboration. We're working together daily. Some of the latest and greatest are we've got worldwide coverage across a broad variety of instances. I think we have 15 different instance types deployed at AWS with broad global availability. AWS is a leading-edge adopter of AMD's fourth-generation epic processor. We call it Genoa, the general purposed M seven A instances previewed last month in June, and I'm optimistic that more instances will be adopting Genoa soon. Something that the AWS EMR team, the Elastic MapReduce team has done recently with AMD Epic processors was to use Epic to improve EMR Spark price performance by 50%. Same job, the cost or same cost, double the work.

Host: Jon

Nice. I like it. Yep. Mike, can you help me understand why do you think AMD is probably the lesser-known out of some of the other semiconductor companies?

Guest: Mike

So why would that be? I think a lot of it comes down to people just being habitual creatures and a lot of cloud users don't think about the processor that they're running on. That's not a choice that's made in their buying decision, which can be a big mistake with AMD processors, AMD Epic processors. You get higher performance and lower retail costs, and so that can end up being a costly lack of choice in the customer's buying decision.

Host: Jon

Mike, I can tell you my experience going way back in time, and I won't date myself, is that we always thought about the processors when we were building computers, when we were building servers in the data center. But now when you're in the cloud, I don't think processors thought about it in the sense that, I'm going to put this in there and my performance is going to increase and go x. I usually look at the number of cores or the performance in some instances and aspects. What are your thoughts on that?

Guest: Mike

Well, I think you hit the nail on the head there, Jon. I think that's the way it goes. A lot of users have other things to worry about, some deadlines to hit or whatever, but some significant benefits come from actually choosing the right processor from price performance to security, particularly with AMD Epic processes where we have a security processor built in that can make a real difference.

Host: Jon

Mike, let me ask you, why should customers or consumers run their workloads on AMD processors though?

Guest: Mike

Well, there are three main reasons. There's getting more computing or capacity for lower cost, or increasing capacity at the same cost. Many customers have constrained budgets, which are either reducing or maybe they don't have the room to fund some new kind of initiative or applications. If they migrate their applications to AMD, they can free up some budget to go invest in capacity expansion or new applications. And then on top of that, like we just discussed, there's the aspect of security with AMD Infinity Guard, which provides the ability to protect data in use. There are some real differentiators there. Some customers need it, everybody should use it because of the global security situation that we're in today and other customers require it. So those are three of the main reasons.

Host: Jon

All right, Mike, as I wrap this up, what's on AMD's roadmap that you want to share with the listeners?

Guest: Mike

Yep. First of all, our next generation fourth generation epic processor, we call this Genoa, has recently been previewed by AWS. This is the general purpose instance. It's called M7A, and that was previewed in June. We're looking forward to that being generally available in addition to the CPU. There are also GPUs from AMD MI 300, which I believe will be a real game changer, particularly for generative ai, which is generating all the buzz these days. It's got a ton of integrated three, which brings the data very close to the accelerator itself, and that helps solve one of the key challenges of those very data-intensive types of applications where data movement is the application killer. So by bringing all of that memory integrated into the chip immediately adjacent to the G P U itself, it should make a real difference for those applications.

Host: Jon

Mike. Nice. I'm looking forward to what's next for AMD in learning more information at your sessions that are happening here at the AWS New York Summit. Yep. All right, everybody. Mike Thompson, senior product manager at AMD here at the AWS New York Summit. My name's Jon Myer. Mike, thank you so much for joining me. Nice to chat

Guest: Mike

With you, Jon. All right, see you

Host: Jon

Later. See you later.