Advanced Micro Devices, Inc. (AMD) Earnings Call Transcript & Summary

[Presentation] Welcome, everyone, and thank you all for joining us. This is a really exciting day. We're here today to talk about the data center and showcase our next-generation server processors and all that they can do. High-performance computing has never been more important than it is now. From how we work to how our kids learn, to how we connect and do business, computing has really enabled every aspect of our lives, and this trend has become more clear over the past year. We're actually now in a high-performance computing mega cycle, and it's driven by cloud computing, digital transformation, 5G and AI. And the fact is everyone needs and wants more compute. Now at AMD, we focus on pushing the envelope in high-performance computing, and that's across PCs, supercomputers, servers, consoles or edge devices. AMD is all about high performance. Over the last several years, we've invested heavily to deliver a leadership portfolio of enterprise-class computing products. Only AMD offers a complete set of commercial-class CPUs, GPUs and differentiated solutions. And that's across enterprise PCs, professional workstations and data center servers as well as the world's fastest supercomputers. You'll see AMD Ryzen, Radeon Instinct and EPYC processors, and these are all designed to meet the needs of the most demanding commercial users. Now today, we're here to talk about our server processors. We've made a tremendous amount of progress with EPYC in the last 3 years. AMD EPYC powers the world's most important workloads. And as the demand for more compute increases, our customers always want more. They want more performance, greater scalability and new layers of security. We thought about all of this when creating our next-generation server processors. And now I'm very excited to show you, for the very first time, our third-generation AMD EPYC processor. Our new EPYC processors extend our leadership in server CPUs and not by a small amount. 3rd Gen EPYC is simply the best server processor available. We deliver more performance, best-in-class scalability and differentiated security. It actually increases our lead in overall performance per socket, enabling maximum compute density, and it also now delivers per-core performance leadership. At the heart of 3rd Gen EPYC is our brand news Zen 3 CPU core. Zen 3 is the latest example of our commitment to building the best technology, and I'm so proud of what the team delivered. With Zen 3, we redesigned the core to deliver our largest generation-over-generation performance increase since we first launched Zen in 2017. Compared to our previous generation Zen 2 cores, Zen 3 delivers 19% more instructions per clock cycle. And as a result, 3rd Gen EPYC CPUs are the fastest server processors in the world. So now let me show you the performance of EPYC in some key workloads. Let's start first with HPC. In HPC, floating point performance is so important, and today, our 2nd Gen EPYC processors in a dual-socket configuration are 76% faster than the competition's best dual-socket processors when running the industry standard SPECfp rate benchmark. And now with 3rd Gen EPYC, that performance increases and we deliver more than double the performance of the competition at 106% faster. What that means for researchers is they can use EPYC processors to reduce the time needed to find the next scientific breakthrough, either by running more simulations in a given period of time or using bigger data sets and more complex models. Now let's turn to look at the cloud applications. When you look at cloud, integer performance is key. And so when you look at the SPECint RATE as benchmark, today, our dual socket 2nd Gen EPYC systems are 81% faster than the competition. And with 3rd Gen AMD EPYC, we extend that lead again, and we're now double the performance or 106% faster than the competition. What that means for cloud providers is that they can take advantage of that additional performance to significantly increase their compute density and provide more performance per socket or rack than previously possible. When it comes to enterprise performance, let's look at how EPYC compares when running Java virtual machines as measured by SPECjbb. Once again, we see that 2nd Gen AMD EPYC is 79% faster than the highest performing offering from our competitor today. And we extend our lead even further with our 3rd Gen EPYC processor, which is 117% better than the competition's best. So what this means for CIOs is that AMD EPYC gives you the best performance, which helps drive better value for your critical business applications. So now let's take a look at 3rd Gen EPYC in action and how this performance translates into real-world experiences. Virtual desktop infrastructure is a critical enabler of today's rapidly changing workforce. VDI provides remote access to key business applications and tools while maintaining a company's data security, and as the number of remote workers increases, businesses need to have the ability to scale their VDI infrastructure to support a larger number of users while maintaining low latency to ensure a good user experience. Let's take a look at how our top-of-the-stack 3rd Gen EPYC processors perform in a VDI workload. Login VSI is the industry-standard virtual desktop load testing tool that simulates the typical workload of a business user. The benchmark adds active VDI sessions to the server, and as the number of users go up, the average response time or latency also increases. And the server continues to add sessions until it's fully saturated and can no longer add more users and still deliver a smooth user experience. So as you can see, the competition has stopped while the 3rd Gen EPYC system can service many more VDI sessions thanks to its higher performance and compute density. So the results are a dual-socket 3rd Gen AMD EPYC system is supporting more than double the VDI sessions compared to the competition and that comes from the combination of high-performance cores, industry-leading core density and high memory bandwidth that makes EPYC-powered servers the best choice for the enterprise. So with all of these examples that we just walked through, they really highlight the power of our flagship 64-core EPYC processors, which deliver the highest performance per socket in the industry. Now in addition to these high throughput applications, EPYC also delivers the highest per-core performance, and you're going to see that shortly. But now let me hand it over to AMD's CTO, Mark Papermaster, to give you more details on how we designed Zen 3 and what it brings to the data center.

Thank you, Lisa. I'm incredibly proud to talk about the new features and performance gains and the latest addition to the Zen family for the data center. AMD has been on a relentless journey to bring competition to the x86 server market, and with the Zen roadmap of high-performance x86 CPUs, we set up the team to develop processor generations in parallel, with designers leapfrogging from one generation to another. This results in consistent delivery as promised. Zen is not just for targeted workloads. It's designed to take on the broad set of applications needed to run your business and to scale as you add more processors. I'm excited to share with you today detail about what Zen 3 will bring to the data center. Zen brought AMD back into high performance with a 52% improvement in instruction per clock in a single generation. It was a strong base for us to build our future upon. Zen 2 brought a double-digit percent increase in IPC and debuted 7-nanometer and chiplet technology. That 7-nanometer efficiency enabled leadership multi-core performance and a doubling of core density. Our innovative chiplet implementation enabled ease of manufacturing in a small 7-nanometer die and scalability for leadership core counts. And now Zen 3. Zen 3 achieves a repeat of double-digit performance gain with a 19% average increase in instructions per clock tick, all of that performance in the same power envelope and 7-nanometer technology as the prior generation. There are design improvements across every unit of the CPU, including a doubling of [ N8 ] throughput to speed inferencing workloads. There are also 2 key security improvements. First, we added shadow stack, hardware mitigation against control flow attacks that leverage return-oriented programming techniques. And second, we further extended our leadership in secure encrypted virtualization by adding secure nested paging to protect against malicious hypervisors. Altogether, Zen 3 is our most significant architectural update to the Zen era. A new processor layout brings all cores on a die into a unified 8-core complex, accelerating core-to-core communication. The consolidation allows every core to directly access 32 megabytes of L3 cache, which dramatically accelerates latency-sensitive tasks, like high-frequency trading or database applications. And finally, significant changes throughout the Zen 3 micro architecture generate a 19% IPC uplift as compared to the prior generation. Achieving this level of CPU performance gain takes focused engineering on the key levers and bringing innovation to each of those levers to make for a great micro architecture. Zen 3 is that great micro architecture with a total front-to-back redesign and a laser focus on core performance and improvements across caching, load/store execution, prefetching and dispatch. We made the Zen 3 floating point in integer execution units wider, more flexible, allowing us to deliver more execution capability for the user at lower latency. We increased the number of loads and stores per cycle versus the prior generation, helping to feed those wider execution units of the Zen 3 architecture. More branch prediction bandwidth also allows Zen 3 to tackle back-to-back predictions much more quickly, reducing delays. These innovations keep Zen 3 ahead of the industry norm on generational improvements and positions Zen 3-based EPYC CPUs with clear server performance leadership, leadership in both single-thread and multi-threaded workloads. We set this goal years ago on our road map and executed the strategy with intense focus. Let's take a closer look at the differences in the core complexes. Zen 2 was organized with 2 4-core cache complexes that were tightly integrated on the core die. Zen 3 transitions to a single unified 8 cores on the die. Every core has direct access to the 32 megabytes of L3 cache rather than 16 megabyte with Zen 2. Every core can now communicate directly to the cache without traversing across the die, reducing latency. Finally, the direct access to a cache of this size is good for applications that make frequent use of the memory subsystem. Many server workloads take advantage of multiple cores and have multiple copies of a workload running on a core die. For instance, database workloads run in this mode with a fixed-size shared data. With a doubled 32-megabyte shared L3 in Zen 3, the effective cache capacity of the data per core increases significantly, directly improving cache hits rates and as a result, performance. At AMD, security remains foundational in our designs, and I'm proud of the data center first with Zen 3's Infinity Guard security feature. It adds strong additional layers to our security offering. And our 1st Gen EPYC secure encrypted virtualization-enabled encrypted memory for each virtual machine and to protect that confidential information. In the second generation, we added SEV-ES to protect the VM control registers on the CPU from compromise. And now with Zen 3 and 3rd Gen EPYC, we add secure nested paging to protect data and use at the virtual machine level. Secure nested paging enables the industry's best virtualized environment to date, and AMD is the first to market with this solution. It's a new layer of protection in multi-tenant cloud. It mitigates an attack vector, meaning the cloud providers' hypervisor controller can't see into those VMs or make changes to the tenant, and that protects data in use even from cloud administrators. Moreover, all of these confidential computing capabilities can be seamlessly implemented as a lift and shift with no application modification needed. We will continue our priority focus on security and plan to add more hardening of AMD security offerings going forward. Our Zen road map is executing to plan. We continue our intense execution focus to bring not only leadership performance over a broad range of workloads but server-class reliability, availability and serviceability, along with leadership security features. Zen 3 is shipping as planned. 5-nanometer Zen 4 are well underway and on track to come to market in 2022. Zen 5 is in design phase. The AMD product development team is hitting on all cylinders. There is no shortage of innovations for future generations of Zen processors to keep AMD at the forefront and deliver the best compute experience for our customers. Thank you very much. And I'm excited to welcome Forrest Norrod to the stage to share more details of the 3rd Gen EPYC family.

Thank you, Mark. Your teams have done an amazing job in developing the industry's highest performance CPU technology. I'm very proud to announce the full family of third generation EPYC server CPUs available today. The latest in the EPYC line, the third-generation 7003 family spans a wide range of performance, starting at the 8-core 72F3 all the way up to a 64-core behemoth, the 7763, which, as Lisa already showed you, is the highest performance server CPU in the world. All of these processors, regardless of their core count, include the complete set of EPYC features: 8 channels of high-performance memories supporting up to 4 terabytes of DRAM per CPU; 128 or more lanes of PCIe generation 4 IO to connect to networking, storage and accelerators without the cost or performance impacts of external bridge chips. Crucially, this EPYC family offers an unmatched set of security features to help keep your data, your users and your customers safe. Third-generation EPYC processors are socket compatible with the second generation, enabling our customers and partners to easily embrace this new level of performance. And the performance we're offering is incredible. When we launched the second-generation EPYC processors, I was very proud to be able to show over 80 world records across a wide range of applications and workloads, and today, the best just keeps getting better. EPYC has over 200 world records across nearly every type of application, workload and customer, and many of these world records are from a single company, Lenovo. To tell you more about what they've accomplished with third-generation EPYC, here's Kamran Amini from Lenovo.

Thank you, Forrest. Lenovo is committed to being a trusted partner to our customers and delivering industry-leading infrastructure solutions for the data center and beyond. With a new portfolio of 10 Lenovo Think system servers and ThinkAgile hyperconverged solutions built on third-generation EPYC processor, we achieved more than 25 new world records across a broad set of industry standard workload benchmarks. In fact, we achieved over 100 total world record benchmarks with Lenovo servers built on both 2nd Gen and 3rd Gen AMD EPYC CPUs. One of the most important benchmark for many of our customers is around virtualization. So I'm excited today to announce that Lenovo ThinkSystem SR665 with the 3rd Gen EPYC CPU 7763 achieved a new world record on VMmark version 3 benchmark for 2-socket match pair servers. This means Lenovo customers can benefit from getting more VMs per data center floor tile. We know how important it is for IT investment and Lenovo to deliver the best possible performance, especially when it comes to mission-critical applications. So I'm equally excited to share that our Lenovo Thinksystem servers with 3rd Gen AMD EPYC achieved multiple world records in key computational intensive benchmarks, including SPECspeed floating point, SPEC accelerate, OMP and MPI. For our customers, this means they can accelerate the speed of innovation and discovery, from vaccine research to climate change to renewable energy. Our customers are the real heroes. We're proud of the collaboration between Lenovo and AMD to deliver infrastructure solutions to help solve some of humanity's greatest challenges.

Thanks to the team at Lenovo for your partnership. It's incredible to see what you've done with the third-generation EPYC processors. That's a testament to the performance that the third-generation EPYC family brings, performance that will deliver real results for our customers. This has showed the leadership performance at the top of the stack. Our 64-core EPYC processors have more than twice the performance of the competition. But many customers prefer lower core counts in the middle of the stack. There as well we're delivering performance that beats allcomers. Before today, in the heart of the stack, the integer performance leader has been the second-generation AMD EPYC 7532, 12% faster than the best our competitor has to offer. Today, the third-generation 75F3 32-core part takes it up to another level and drives 40% more integer performance than the competition. That performance that delivers better results for enterprises, databases, data analytics, Java workloads, a wide range of applications. There is no place where performance is more important than in high-performance computing, HPC. Once again, the previous floating point performance 32-core leader was the second-generation AMD EPYC 7532, 40% faster than our competitors' best. Today, with the third-generation 75F3, we are achieving 70% higher performance on the SPECfp rate benchmark than the competition. With the performance of the EPYC processor line, AMD is now leading across many areas of HPC. From natural resource exploration to financial modeling, from advanced robotics to product design for a wide range of industrial, consumer and automotive products, AMD is the processor of choice. Let me show you what that level of performance leadership means to customers. Companies designing cars live in a highly competitive market where success depends on developing the most comfortable, economical, efficient and attractive vehicles while always making them as safe as possible. One critical task is testing how the car performs in a collision. Our friends at ANSYS have a great simulation tool, LS-DYNA, that allows car companies to design a vehicle, crash it in a large number of virtual scenarios and then iterate on the design and testing until the car is as safe as possible. This is all done in servers before they bend one piece of metal. What you see on the screen is LS-DYNA running a simulation of a 3-car collision, one of many such tests that need to be run. It's running on 2 different systems: one powered by a pair of third-generation EPYC processors and the other using top-of-the-line dual-socket processors from our competitor. The EPYC-based server performs the simulation faster, allowing the designers to speed up product design and achieve a safer car. In fact, the EPYC-based system performed the simulation over twice as fast. Now you may have noticed the EPYC-system had 32 cores in each processor while the Intel had 28 cores. But even if we normalize for the core count, the EPYC system delivers 98% more performance per core in this HPC application. That performance means shorter development time while running far more simulations, hence improved product quality. Leadership in HPC means getting a better, safer car to your driveway sooner. At the apex of HPC is supercomputing. These systems are the largest computers in the world, the computers that perform complex genomics and proteomics research that researchers use to predict our weather and climate. Today, supercomputers harness the power of thousands or tens of thousands of CPUs and increasingly, compute-optimized GPUs working in concert. We are now in the petascale era, where the largest supercomputers in the world are able to execute a million billion-floating point operations per second, but we're about to break through to the next level, to the exascale era, where supercomputers will be 1,000x faster, able to execute over a quintillion, a billion billion-floating point operations per second. These systems will tackle problems that have hitherto been beyond our reach. AMD is leading the journey to exascale. With our third-generation EPYC CPUs and Instinct compute-optimized GPUs, we are helping companies and organizations around the world build pre-exascale systems that are laying the foundation for the exascale era. This year, we will be helping our partners install LUMI, the largest supercomputer in Europe based on EPYC and Instinct; and build Australia's largest supercomputer, again, using third-generation EPYC CPUs and next-generation Instinct GPUs. AMD is immensely proud to be part of the effort later this year that will build the industry's first system to break the exascale barrier, the Frontier supercomputer at Oak Ridge National Labs. The Frontier system is being built in collaboration with the Cray team at Hewlett Packard Enterprise. Powered by third-generation EPYC processors and next-generation Instinct GPUs, Frontier will be capable of over 1.5 exaflops, 1.5 quintillion calculations every second. What Frontier can do in 1 second would take 7 years if every human did 1 math calculation a second. That is computation in an unprecedented scale and shatters the limits of today's supercomputers. When we launched the EPYC processor family, AMD promised to deliver a road map of high-performance innovative server CPUs. The third-generation EPYC processors continue to deliver to that promise with unmatched performance, incredible scalability and advanced security features. Those attributes are important to many customers and applications but perhaps are most important in the servers that power the cloud. Now here's Lisa to tell you more.

Thank you, Forrest. The cloud has never been more important to the world, and with the spike in demand, cloud providers are focused more than ever on making the right technology investments to deliver the best performance, scalability and security, which is exactly why leading cloud providers are choosing AMD EPYC processors. We take incredible pride that the world's largest and most important cloud services are powered by AMD. Billions of users around the world rely on AMD EPYC-powered services, including Microsoft 365 and Teams, Tencent Meeting, Twitter or Zoom. From Amazon to Alibaba, Google, IBM, Microsoft, Oracle and Tencent, the world's leading cloud companies use AMD EPYC. We're very pleased with our progress in the cloud, and today, we have over 200 first and second-generation EPYC instances available. As we turn to the third-generation EPYC, we see even more opportunity to accelerate our cloud deployments. Now I'd like to invite a few of our cloud partners to talk about how they are deploying 3rd Gen EPYC. Let's start first with Microsoft Azure, a partner who is taking full advantage of all of the benefits of EPYC. To tell us more, here is Executive Vice President of Microsoft Azure, Jason Zander.

Thank you, Lisa. Let me be among the first to congratulate you and your team on the launch of AMD EPYC third-gen processors. We're proud of our deep partnership with AMD. Together, we provide the essential computing power for our customers around the world to run a wide range of general-purpose and specialized workloads in Azure while also enabling us to operate Microsoft services at a global scale, such as Microsoft 365 and Teams. Today, Microsoft and AMD are jointly announcing general availability of high-performance computing, HB-series VMs in Azure running on AMD EPYC third-gen processors. Our vision for HPC and AI in Azure is based on a foundation of genuine and recognized expertise that uses proven HPC technology and design principles enhanced with the best features of the cloud. This provides application scaling that is 12x higher than on other public clouds, with performance that rival some of the most powerful supercomputers on the planet. As a result, more and more engineers, researchers and scientists are turning to Azure to accelerate their most important HPC workloads. Azure remains at the forefront as the first public cloud to have the most scalable and powerful set of HPC solutions available on the latest AMD EPYC processors. Our HPC customers have achieved some amazing success with our AMD-based VMs. In Azure, your data is your data and maintaining privacy is a core tenet of our approach. As technology continues to evolve, we will always look for new ways to help protect your data. Today, Microsoft and AMD are jointly announcing the private preview of confidential computing VMs in Azure running on AMD EPYC third-gen processors. Confidential computing builds on the strong encryption at rest and in transit capabilities to keep your data encrypted all the way to the CPU. Customers can easily take advantage of this added protection for their most sensitive workloads without the need to rewrite or recompile them. With confidential computing, the run time state of your VMs is fully encrypted by CPU-generated keys. And so the contents of the virtual machine are opaque to administrators. This unprecedented protection of data in use adds a new level of isolation in addition to the existing technical and operational controls that are already in place with Azure. Lastly, we're also excited to share our plans to introduce general-purpose Azure VMs featuring AMD EPYC third-gen processors later this year. Our continued partnership and the innovations that we've seen along the way have brought us to where we are today. And looking forward, I think this partnership will be critically important to our future accomplishments.

Thanks so much, Jason. Microsoft has been an incredible partner, and we have made so much progress over the last year. We're very excited that Azure is announcing immediate availability of the new HBv3 VMs with leadership HPC performance and a private preview of new confidential computing VMs as well as general-purpose VMs launching later this year, all powered by our 3rd Gen EPYC processors. Now let's hear from our friends at Google. Google has adopted EPYC to power their global data centers as well as many of their Google Cloud offerings. To provide more insight on what we're doing together, let's hear it from Engineering Fellow and VP of Systems Infrastructure, Amin Vahdat.

Thank you, Lisa. Congratulations to you and the team. We are very excited to be a part of the launch of the third-generation AMD EPYC CPUs. AMD technologies have been important for us and our customers to scale up and out. Google has been using AMD EPYC processors in our data centers around the world to run some of the most demanding hyperscale applications, cloud infrastructure and web properties. Our services have some of the broadest reach on the planet, helping billions of users around the world, and these processors now power many of these services. Google strives to improve people's lives around the world, and with AMD's help, we are able to improve the ecosystem to help achieve this goal. At Google Cloud, we are committed to partnering with customers on their digital transformation journeys. Our partnership with AMD helps us accelerate this commitment in 2 very important ways. First, we brought AMD EPYC-based VMs to Google compute engine last year, giving our customers a wider range of VM families to choose from at improved price performance. Second, a key element of any digital transformation is security and isolation of workloads. That's why we introduced confidential VMs and Confidential GKE Nodes, the first products in our confidential computing portfolio. They have breakthrough technologies that encrypt data in use while it's being processed, and they run on AMD EPYC. I'm excited to announce we are building on our existing partnership with AMD and plan to bring the third-generation AMD EPYC processors to Google Cloud customers later this year. Google Cloud will introduce a new compute-optimized VM called [ C2D ] based on third-generation EPYC processors. This offers new machine sizes for compute-intensive workloads such as high-performance computing. We will also extend our current general-purpose offering, N2D, to third-generation EPYC processors. When that launches, customers will be able to auto upgrade to the new CPU generation. Finally, confidential computing will be available on both [ C2D ] and N2D on the latest AMD EPYC processors at the time of launch. We are excited to continue partnering with AMD to bring the latest technology to market and deliver increased performance and functionality to all Google users, customers and services. Back to you Lisa, and thank you very much.

Thank you, Amin, for the strong partnership. We're thrilled with how broadly Google is using EPYC to power many of their internal properties as well as Google Cloud, and we're very excited that Google will bring the power of 3rd Gen AMD EPYC to Google Cloud customers with new [ C2D ] and N2D VMs with support for confidential computing. One of the other large cloud providers that has adopted EPYC processors to rapidly scale their applications and services is Tencent Cloud. To talk a little bit more about the work we're doing together and what comes next, let's hear from Senior Executive Vice President of Tencent Cloud, Dowson Tong.

This is Dowson Tong from Tencent. Tencent has been partnering with AMD for the past few years. The partnership of AMD and Tencent will lead the industry with high-performance infrastructure products and services that deliver the best ROI to our customers. Today, we are very excited to announce Tencent Cloud has a 3-server instance together with the third-generation AMD processor launch. The series of SA1, SA2 and SA3 has been supporting many application scenarios, ranging from database to big data analytics, from on-demand video to real-time live streaming, from public cloud to hybrid cloud environments. These applications are used by over 1 billion people every month, and we will enable even more business scenarios on SA3 instances. With Milan's [ N8 ] performance enhancements, SA3 customers can now enjoy a substantial performance boost for their AI applications such as recommendation engines, computer vision, natural language processing and more. AMD and Tencent Cloud have been able to demonstrate significant performance improvement on AI inference using AMD's optimized libraries for SA3 in TensorFlow or Resnet50. Today, Tencent Cloud's SA3 instance is ready for our customers to order online. With the acceleration of the industrial Internet development, cloud computing has become an expansible foundation of the digital economy. Tencent will continue to invest in the cloud and data center technologies and work closely with our partners to support the digital transformation of industries. Once again, congratulations to AMD. Thank you very much for the support to the partnership.

Thank you, Dowson. It's wonderful to see the growing use of EPYC in Tencent Cloud, and we're proud that more than 1 billion people every month are now using Tencent services powered by EPYC. We're thrilled that starting today, Tencent is offering customers new SA3 instances powered by 3rd Gen EPYC. I'm looking forward to working together to bring the benefits of EPYC processors to even more Tencent Cloud users and customers. This is just a preview of what you can expect with 3rd Gen EPYC in the cloud. When we look at the number of instances that we're announcing today and that we have in development with our partners, we're on track to more than double with 3rd Gen EPYC, which translates into over 400 AMD-powered instances across all EPYC families by the end of 2021. That's just tremendous progress, and we're incredibly proud that the industry's leaders are increasingly choosing EPYC processors to power their businesses. Now let's turn to enterprise adoption with AMD Senior Vice President and Server General Manager, Dan McNamara.

Thank you, Lisa. For the next few minutes, I will focus on how we work with our ecosystem partners to develop solutions that solve enterprise's biggest challenges, and those challenges are complex as we learn through constant conversation with IT leaders. They need to be able to respond quickly to rapidly changing business conditions. They are constantly under pressure to meet the ever-increasing workload demands. They are being asked to not just keep the infrastructure running but to improve the efficiency of the operation. Now on top of all these challenges, there are ever-increasing security threats that need to be defended against in real time. With this environment in mind, we developed 3rd Gen EPYC to help enterprises overcome these challenges and thrive. It's the industry's highest performance server CPU, and that translates to lower total cost of ownership, helping our customers do more with less. And we're laser-focused on delivering that value quickly and seamlessly. We are partnering with the best in the industry across operating system, application software and server infrastructure to harness all these innovations and transform them into solutions. We're focused on the solution areas that are top of mind for enterprise IT leaders, starting with hyperconverged infrastructure for hybrid and multi-cloud environments, moving to relational databases for transactional data processing and to data analytics for teasing out insights from data. Let me start with hyperconverged infrastructure, which is gaining tremendous momentum in private cloud and hosted environments. It offers the flexibility, scalability and operational efficiency that our customers need. We have partnered with industry leaders to bring to market a set of certified solutions as well as fully integrated appliances. This includes solutions from VMware, Nutanix, HPE and Microsoft. One of the key advantages that 3rd Gen EPYC brings to HCI is unmatched VM density and performance for a wide range of applications. Let me highlight one of them. With our 3rd Gen EPYC, we deliver 180% higher performance and industry-leading VM density with the popular VMmark benchmark on a comparable configuration based on our competition. We have made significant generational improvements working closely with our ecosystem partners. The net effect allows customers to improve performance and consolidate infrastructure. Now let's talk about the sea of data that needs to be processed daily in the enterprise and look at how infrastructure consolidation drives increased value to IT leaders. Relational databases continue to be the foundation for many enterprise core applications and require the highest level of reliability, performance and scalability. AMD is partnered with the best in the ecosystem to develop world-class solutions to meet these demanding requirements. Today, we are announcing several performance proof points for relational databases. One of them is a TPC-H benchmark world record. This benchmark is a popular yardstick in measuring the performance of decision support systems. I'm very proud to show you that our 2-socket 3rd Gen EPYC-based system is outperforming a 4-socket Cascade Lake-based system by 14% in performance with Microsoft's SQL Server 2019. Think about this value. You can do more with 2 EPYC processors than you can do with 4 processors from our competition. This is a good indication of how enterprises can achieve high performance for database applications while lowering the cost of ownership with EPYC. Continuing with the access to the mountains of data available today. Enterprises need the compute power to analyze the data and get to actionable insights quickly. In today's data-driven economy, the ability to analyze data is often a key competitive advantage. AMD is partnering with leaders in data analytics to optimize and tune their software and solutions on EPYC processors. Over the years, Apache Hadoop has become one of the foundations of data analytics in the enterprise and cloud. Today, we are announcing a world record for Hadoop with 60% higher throughput for one of the popular industry standard benchmarks, demonstrating the unique performance and TCO advantages of our single-socket servers. Getting more data analytics performance from a single-socket EPYC-based server versus a dual-socket Xeon server means enterprises can do 1 of 2 things with tremendous savings: one, they can reduce the overall cost of the solution; or two, they can invest those savings back into the infrastructure to process even more data, getting deeper and more nuanced insights. Either way, the enterprise that adopts 3rd Gen EPYC is gaining a competitive advantage. If I take a step back, I've shown you leadership performance and total cost of ownership benefits driven by 3rd Gen EPYC on HCI, relational databases and data analytics and how we're working with our ecosystem to bring compelling solutions to market. Next, I'd like to invite some of our closest partners to talk about how they're using 3rd Gen EPYC to enable these solutions for you. Let's hear from Ravi Pendekanti on how Dell is delivering HCI and other solutions based on our newest EPYC processors.

Thanks, Dan. Advances in IT have freed workloads from residing in a given location. Applications have moved to where they deliver their best performance. Video games can be run and streamed from the cloud. Cars are driving themselves. And most documents on your laptop really aren't on your laptop. This shift has put significant pressure on organizations to modernize their infrastructure. The Dell Technologies partnership with AMD is closer than ever as we jointly deliver on top performance with PowerEdge servers with third-generation AMD EPYC processors. Dell Technologies is not only delivering some of the world's highest performing servers with AMD inside of PowerEdge but will be throughout our portfolio, including our Ready Solutions for databases at HPC, VxRail HCI platforms and Azure Stack HCI solutions. Modern infrastructure design must match the unique requirements driven by artificial intelligence. The PowerEdge XE8545 is AI infrastructure without compromise, delivering optimized CPU and GPU performance, accelerated IO throughput and simplified management. The XC8545 trains Resnet50 image classification to top-quality accuracy in under half the time of the prior generation PowerEdge systems. Customers can order and receive PowerEdge with their third-generation AMD EPYC processors today. Whether you are pushing the boundaries of your existing applications or building a data center that can handle your future growth, Dell is delivering these solutions now. Dan, back to you.

Thanks to Ravi and the entire Dell team for their partnership. We are excited to have their new EPYC-based PowerEdge servers available for sale today. And as Ravi mentioned, they are also expanding their Ready Solutions with 3rd Gen EPYC across databases, HPC and HCI. Earlier, Mark showed you the progression of security innovation in our EPYC road map to protect enterprise data. Let's hear from Krish Prasad on how VMware is bringing the value of these modern security features to private and hybrid cloud deployments.

I'd like to thank AMD for the opportunity to participate in the Milan launch. VMware has been a very strong supporter of AMD EPYC since the beginning. In 2019, I had the pleasure of sharing the stage with Mark Papermaster to announce support for VMware products on AMD EPYC Rome. Today, I'm very excited to announce day 0 support for Milan, for VMware's flagship vSphere as well as the vSAN product lines. Security is top of mind for our customers. Working closely with AMD, we have leveraged the industry-leading technology in this area to provide an added layer of security for our customers, not only on the hypervisor but also for the VMs and containers that run on top of it. Last year, VMware announced vSphere with Tanzu, where we built in natively Kubernetes as well as containers into our platform along with our VMs. Today, we are announcing SEV-ES support for vSphere with Tanzu in version 7.0 U2. The collaboration between VMware and AMD has never been better. Our engineers have worked together, and we have done a lot of optimization for the vSphere scheduler as well as the vSAN code for the AMD micro architecture. As a result, I expect to see stellar top-notch performance for the VMware products, both vSphere and vSAN running on servers with AMD Rome as well as AMD Milan. Our joint customers are interested in increasing VM density as well as continuous improvement in total cost of ownership. With the launch of Milan and the joint work that we have done to announce day 0 support for VMware vSphere and vSAN, we expect the adoption to continue and even accelerate. With that, I'd like to thank AMD for the opportunity to participate in the Milan launch. Thank you again.

Thanks to Krish for his partnership. As Krish said, VMware has added secure encrypted virtualization support in vSphere for VMs and containers. And our joint engineering collaboration has also delivered enhanced performance levels on vSAN and vSphere. We are very excited about how our partnership has grown as EPYC continues to see broad adoption. Earlier, Lisa showed you the power of 64-core EPYC to drive significant value in deploying VDI clients. Let's hear from Neil MacDonald on how HPE is tapping into the power of 3rd Gen EPYC to deliver a broad set of solutions across enterprise and high-performance computing.

Thank you, Dan. We couldn't be more excited about AMD's new third-generation EPYC processor and what it means for our already strong and growing partnership. With today's announcement, HPE will double the number of AMD-based ProLiant servers in our lineup, offering greater choice to match business needs. And working closely with AMD, we're delivering these new servers tuned for the most critical workloads. HPE is also extending AMD's new third-generation EPYC processors into our HPE Apollo and HPE Cray high-performance computing portfolio. For data- and IO-intensive workloads, we have the new ProLiant DL345. And to support a remote workforce and workplace strategy, we have the new dense 2-socket ProLiant DL365 for VDI. Both servers will be available stand-alone or as part of our solutions portfolio. We are integrating security as a core capability into every HPE ProLiant server. Our third-generation AMD EPYC-based products will be no different. As your data center becomes more virtualized, servers should be the strongest defense, providing the latest innovations to prevent, detect and recover from security attacks. That's why we continue to develop and enable holistic security technologies, providing a foundation to protect data and our customers' business. HPE silicon root of trust, combined with the inherent security and encryption of AMD EPYC processors, delivers unmatched foundational security at the hardware and firmware level. The new secure nested paging on third-generation AMD EPYC processors has memory integrity protection to help prevent malicious hypervisor-based attacks and creates an isolated execution environment, making the servers even more secure. With AMD, we will continue to help our customers accelerate their digital transformations, empower their workforce, deploy new resilient IT solutions and extract insights from critical data while leveraging and consuming these solutions more dynamically as a service.

I want to thank Neil and the entire HPE team for our partnership across platforms and solution development for the enterprise and the high-performance computing segment. We are excited to have Neil announce workload-focused platforms based on 3rd Gen EPYC processes across ProLiant, Apollo and Cray server products. I hope you got a feel from our partners about the accelerated value AMD brings to their solution offerings. As you can see, there is significant enthusiasm for EPYC in the industry. And if you want to see a broader range of solutions available, you can browse the more than 20 solutions overviews on our launch site. To sum it up for enterprise IT, when you put together the performance and value of 3rd Gen EPYC, combined with these solutions, you will be able to do more, do it faster and do it for less. Now that is transformational value for the industry. Thank you. And now let's turn it back to Lisa.

Thanks, Dan. Let me now bring this to a close. I hope you now see why we're so excited about 3rd Gen EPYC. EPYC is the world's fastest server processor, powering infrastructure and services from the largest cloud providers and delivering superior value while offering lower TCO for enterprise customers, all with the confidence of modern security features. Whether you have an on-prem infrastructure or you're a cloud user or you have a hybrid environment, 3rd Gen EPYC processors are simply the best in the industry. I'm so thankful for our amazing partners who joined us today. They are just a tremendous set of partners and really just the tip of the iceberg in terms of the broad ecosystem we've built as we bring the benefits of EPYC processors to as many cloud, enterprise and HPC customers as possible. You can expect to hear and see much more about 3rd Gen EPYC in 2021. For the enterprise, we'll have more than 100 EPYC server platforms available from all the leading providers. For the cloud, there will be more than 400 EPYC-based instances available. And we've also worked across the entire ecosystem with a broad range of software providers and system integrators to deliver solutions for nearly every application and customer segment. We are incredibly proud of the deep partnerships we have built over the past several years across the entire server ecosystem, and you can count on us to continue to push the envelope in high-performance computing. I couldn't be more excited about the opportunities ahead. Thank you for joining us today.