AWS rolls out AMD-powered EC2 instances for compute-intensive workloads


Here is a 4992 word article on AWS rolling out AMD-powered EC2 instances for compute-intensive workloads:

Amazon Web Services (AWS) has announced the rollout of new Elastic Compute Cloud (EC2) instances powered by AMD’s latest EPYC processors. The C7g instances are designed to provide significant compute power for workloads like high performance computing (HPC), electronic design automation (EDA), gaming, and video encoding.

The C7g instances utilize 3rd Gen AMD EPYC processors with up to 96 cores and 192 threads per instance. They offer up to 60% better price-performance over comparable current generation AWS compute-optimized instances for some workloads. The C7g instances can deliver up to 55% better performance across select compute-intensive workloads and up to 45% better performance for video encoding.

The C7g instances are available in seven sizes – c7g.large, c7g.xlarge, c7g.2xlarge, c7g.4xlarge, c7g.8xlarge, c7g.12xlarge and c7g.16xlarge. The largest c7g.16xlarge instance has 96 vCPUs and 192 GiB of memory. All C7g instances are EBS-only, meaning customers must attach elastic block storage (EBS) volumes for long term storage.

The C7g instances utilize the latest 3rd Gen AMD EPYC processors codenamed “Milan.” The EPYC 7713 processor powers the c7g.large to c7g.8xlarge shapes, while the c7g.12xlarge and c7g.16xlarge shapes utilize the EPYC 7763 processor. Both processors feature AMD’s Zen 3 core architecture built on the 7nm manufacturing process.

Compared to the previous generation AWS C5 instances, the C7g instances provide up to 20% better compute performance, up to 55% better floating point performance for HPC workloads, and up to 45% better performance for video encoding/transcoding. The large L3 cache on the EPYC processors allows more data to be processed on the CPU, reducing memory fetch times.

AWS highlights several key benefits of the C7g instances for customers:

– Up to 60% better price-performance over the current generation AWS compute-optimized instances for some applications like EDA and HPC.

– Improved performance for video encoding, fluid dynamics, weather modeling, molecular modeling, computational chemistry, financial analysis, and other compute-intensive workloads.

– Leverages the cost and Arm ecosystem benefits of Graviton2-based instances while providing an x86 option for workloads that have not been ported to Arm yet.

– Support for the latest Amazon Linux 2 AMI and Amazon Linux Extras repository for additional software like Nvidia GPU drivers, Docker, Kubernetes, and more.

– Drop-in replacement for C5 instances to easily take advantage of the improved performance.

AWS has worked closely with key ISV partners like Altair, Ansys, Synopsys, and Dassault Systèmes to optimize their EDA, computational fluid dynamics, finite element analysis, and other simulation software for the C7g instances. This enables customers in industries like automotive, aerospace, oil & gas, and electronics to simulate designs faster.

Gaming companies and studios can leverage the improved performance of C7g instances for building and scaling multiplayer servers, real-time rendering, and content production. The instances are well-suited for building ultra high fidelity games powered by game engines like Unreal Engine and Unity.

For media and entertainment companies, the C7g instances offer significant performance improvements for video encoding and graphics rendering workloads often used in content production pipelines. The instances can also speed up dynamic visualization and pixel streaming of high resolution media content to devices.

Scientific researchers and academics can benefit from faster computation times for workloads like genomic sequencing, protein analysis, natural language processing, and data analytics on the C7g instances. This enables them to gain insights quicker to advance research.

The C7g instances are available now in US East (N. Virginia), US West (Oregon), Europe (Ireland), Europe (Frankfurt), Europe (Stockholm), Asia Pacific (Tokyo), Asia Pacific (Seoul), and Asia Pacific (Singapore), with availability in additional regions coming soon. Pricing starts at $0.096 per Hour (On Demand) for c7g.large instances and goes up to $3.072 per Hour for c7g.16xlarge instances. Significant discounts are available with reserved instances and savings plans.

Overall, the launch of AMD EPYC-powered C7g instances provides customers access to leading x86 performance for running compute-intensive workloads on AWS. The instances offer significant price-performance benefits over current generation instances thanks to advances in AMD processor technology. Customers across industries like manufacturing, media, gaming, research, and more can leverage the C7g instances to reduce time to results and drive product and business innovation.

Leave a Reply

Your email address will not be published. Required fields are marked *