AWS and AMD
Powerful processors to right size your workload
Amazon Elastic Compute Cloud (Amazon EC2) provides customers the most options for compute so they can tailor the infrastructure to their business needs, and AMD is integral to these offerings. AMD-powered EC2 instances give customers the ability to run general purpose, memory intensive, burstable, compute intensive, and graphics intensive workloads, all at a significant price advantage relative to comparable offerings.
AWS and AMD have collaborated to give customers more choice and value in cloud computing, starting with the first generation AMD EPYC processors in 2018, expanding to the second generation AMD EPYC processors in 2020, and most recently with third generation AMD EPYC in the Amazon EC2 C6a, M6a, and R6a instances . Customers can use EC2 instances powered by AMD and enjoy scalable performance for a broad variety of workloads such as databases, mission critical enterprise applications, web scale in-memory databases and caches, big data analytics, batch processing, gaming, log analysis, remote workstations, transcoding, and rendering.
Workloads
General Purpose
These workloads benefit from a balance of compute, memory, networking, and storage resources. Use cases include business critical applications, web and application servers, back-end servers for enterprise applications, gaming servers, caching fleets, and application development environments.
Memory Optimized
Fast performance for workloads that process large data sets in memory. Examples include high performance databases, distributed web scale in-memory caches, mid-size in-memory databases, real time big data analytics, and other enterprise applications.
Graphics Intensive
Customers looking to use remote workstations in the cloud for running graphics applications can use G4ad instances to get the flexibility to provision resources on a per project basis rather than being limited by on-premises capacity.
Compute Intensive
Compute intensive applications benefit from high-performance processors and are bound by the amount of compute power that can be delivered. Examples include high-performance web servers, high-performance computing, batch processing, ad serving, highly scalable multiplayer gaming, video encoding, scientific modeling, distributed analytics and machine/deep learning inference.
Spiky Workloads
Spiky workload applications have variable CPU usage that experience occasional spikes in demand that need to burst seamlessly to meet temporary traffic peaks and then scale back down to operate at typical traffic levels. Examples include microservices, low-latency interactive applications, small and medium databases, virtual desktops, development environments, code repositories, and business critical applications.
Customer Success Stories
“Dropbox is a leader in personal and enterprise online storage and is in a unique position to utilize AMD technology on-premise and in the cloud. We’re able to deliver on our promise to help customers work smarter, in part, through the technology partnerships we’ve established. AMD’s EPYC processors deliver the performance we need on our on-prem workloads, while Amazon EC2 instances powered by AMD processors offer us significant savings.”
Capital One Financial Corporation, a leading information-based technology company and digital banking innovator, has taken a cloud-first approach to software development that provides highly-personalized customer experiences in the areas of credit cards, auto loans, banking, and savings products. Capital One has been able to leverage the utilization and cost benefits of AWS using EC2 instances powered by AMD EPYC processors. Capital One is always looking for ways to optimize our cloud utilization, taking advantage of Amazon EC2 R5a, M5a, T3a and C5a instances, which are all based on AMD’s EPYC processors.
“Sprinklr is a provider of enterprise software for customer experience management. More than 1,200 brands in 150 countries create happier customers with Sprinklr, including L’Oreal and McDonald’s. We are always looking for ways to improve our cloud utilization and after using AWS’s EC2 M5a, R5a and C5a, which are all based on AMD’s EPYC processors, we’ve secured substantial savings on our cloud costs.”
“Wego is the leading online travel marketplace in Asia Pacific and the Middle East and North Africa (MENA) and it harnesses powerful yet simple technology that automates the process of searching and comparing results from hundreds of airline, hotel and online travel agency websites. We operate 60 country sites in over 22 languages and 58 currencies supporting millions of travelers every month. We had been using Amazon EC2 M5a instances since last year for our microservices, and soon after the EC2 C5a instances were launched, we migrated our Java-based metasearch services to these instances. With no code changes necessary, the transition was seamlessly completed in just a few days and it allowed us to achieve more than 30% cost savings.”
“Thatgamecompany studio (best known for beautifully designed games such as Flow, Flower and Journey), has recently seen the major benefits of using C5a to support our games. Sky: Children of the Light, which Apple named 2019’s iPhone Game of the Year, leans into C5a to support Sky’s backend services including Account Services, Load Balancing, Matchmaking, and more. Making the migration to C5a was straight-forward and easy, and we immediately saw the cost and performance benefits of doing so. As a longtime fan of AMD, I was excited to see their new 2nd Gen EPYC processors offered in the C5 family on AWS.”
“Elementum is the company behind the first cloud-native supply chain orchestration platform. In an age where instant gratification is the new norm and customer expectations are continuing to grow exponentially, successful operations need to adapt faster than ever to unplanned exceptions. Amazon Web Services provide our SRE team with the breadth and depth of services needed to scale our business while managing costs. For example, we recently moved our DC/OS and CoreOS based micro-services platform from 3rd generation EC2 instances to current generation AMD EPYC based instances. The transition was seamless, we're able to give our Java developers lots of memory with the same number of vCPU's which delivers snappy performance, and we achieved the 20-25% cost savings we estimated at the start of this project.”
“Asurion helps people protect, connect and enjoy the latest tech – to make life a little easier. We’ve built a global scale presence and are still growing. Every day our 19,000 experts help more than 300 million people around the world solve the most common and uncommon tech issues. Amazon Web Services provides a broad and deep portfolio of cloud services giving us the agility to scale our business without compromising our dedication to service, and commitment to exceptional customer experience. We recently migrated our memory intensive customer claim and fraud analysis services from a mixture of Amazon EC2 M4 and M5 instances and consolidated them on Amazon EC2 M5a instances featuring AMD EPYC processors. The M5a instances provide a seamless migration path, offering the same instance sizes as M5 and application compatibility while delivering cost savings and exceeding our performance expectations.”
“Druva delivers data protection and management for the cloud era. We operate in 14 global AWS regions, serving 4,000+ customers and protect a global footprint of 150+ PB of data stored completely within AWS. What began as a thought experiment for our global footprint of compute resources with AMD-powered instances transformed into a broad initiative resulting in powerful outcomes. In less than two weeks we were able to successfully transfer our global capacity from legacy instances to AMD-powered ones, without any additional effort required from our engineering teams. We unilaterally drove this change between the Operations and Performance teams, and in the process achieved a remarkable saving of 10-15 percent and in some cases as much as 50 percent.”
“Auvik's cloud-based network management software keeps IT networks around the world running optimally. We partnered with AWS to allow us to create a scalable infrastructure, and continue our commitment of delivering a phenomenal customer experience. We recently migrated to AMD powered instances using M5a, T3a, and C5a. Auvik seamlessly migrated over six hundred servers to AMD instances which allowed us to reduce EC2 costs by 10%, without sacrificing performance.”
“Dream11 is a fantasy sports platform with 100 million+ users. We are running in-memory databases on Amazon EC2 instances powered by AMD EPYC processors. We performed benchmark testing on the Amazon EC2 R5a instances and observed performance and throughput as expected. We run multiple clusters on Amazon EC2 instances powered by AMD EPYC, using r5a.large, r5a.xlarge, r5a.2xlarge, and r5a.4xlarge instance sizes. The adoption of EC2 instances with AMD EPYC processors was seamless because we used the same AMIs from our R5 instances. There were no software compatibility issues , and we were able to adopt AMD processors in just one week."
Instances
Amazon EC2 M7a
Amazon EC2 M7a instances, powered by 4th generation AMD EPYC processors, deliver up to 50% higher performance compared to M6a instances. M7a instances support AVX3-512, VNNI, and BFloat16, and feature DDR5 memory, which provides 50% higher memory bandwidth compared to DDR4 memory to enable high-speed access to data in memory. M7a instances are ideal for applications that benefit from high performance and high throughput such as financial applications, application servers, simulation modeling, gaming, mid-size data stores, application development environments, and caching fleets.
Amazon EC2 M6a
Amazon EC2 R6a
Amazon EC2 R6a instances are powered by 3rd generation AMD EPYC processors with an all-core turbo frequency of 3.6 GHz, deliver up to 35% better price performance compared to R5a instances, and offer 10% lower cost than comparable x86-based Amazon EC2 instances. R6a instances also offer new larger sizes with up to 192 vCPUs and 1,536 GiB of memory. These instances are SAP-Certified and are an ideal fit for memory-intensive workloads, such as SQL and NoSQL databases; distributed web scale in-memory caches, such as Memcached and Redis; in-memory databases and real-time big data analytics, such as Apache Hadoop and Apache Spark clusters; and other enterprise applications.
Amazon EC2 G4ad
G4ad instances, powered by AMD Radeon Pro V520 GPUs, provide the best price performance for graphics intensive applications in the cloud. These instances offer up to 45 percent better price performance compared to G4dn instances, which were already the lowest cost instances in the cloud, for graphics applications. They provide up to 4 AMD Radeon Pro V520 GPUs, 64 vCPUs, 25 Gbps networking, and 2.4 TB local NVMe-based SSD storage.
Amazon EC2 C5a
C5a and C5ad instances offer leading x86 price/performance for a broad set of compute-intensive workloads and feature 2nd generation 3.3GHz AMD EPYC 7002 series processors built on a 7nm process node for increased efficiency. C5a and C5ad instances provide up to 20 Gbps of network bandwidth and 9.5 Gbps of dedicated bandwidth to Amazon EBS. With C5ad instances, local NVMe-based SSDs are physically connected to the host server and provide block-level storage that is coupled to the lifetime of the instance.
Amazon EC2 T3a
The T3a instances feature AMD EPYC 7000 series processors with an all core turbo clock speed of 2.5 GHz. The AMD-based instances provide additional options for customers that do not fully utilize the compute resources and can benefit from a cost savings of 10 percent.
Amazon EC2 C6a
C6a instances are powered by 3rd generation AMD EPYC processors with an all-core turbo frequency of 3.6 GHz, deliver up to 15% better price performance compared to C5a instances, and offer 10% lower cost than comparable x86-based EC2 instances. C6a instances also offer new larger sizes with up to 192 vCPUs and 384 GiB of memory, enabling you to consolidate workloads on fewer instances and save on per core licensing costs. These instances feature more than twice the network bandwidth of C5a instances, and are designed for compute-intensive workloads such as batch processing, distributed analytics, high performance computing (HPC), ad serving, highly-scalable multiplayer gaming, and video encoding.
