Advanced Computing in the Age of AI|Tuesday, September 29, 2020
  • Subscribe to EnterpriseAI Weekly Updates:  Subscribe by email

AWS Graviton2 Arm-Based EC2 Instances Now Available 

Amazon Web Services (AWS) has announced the general availability of sixth-generation Amazon EC2 C6g and R6g instances powered by Arm-based AWS Graviton2 processors. The new instances are available starting today in a variety of sizes across four regions alongside the general-purpose M6g, which became available last month.

The Graviton2 chip. Image courtesy of AWS.

The 7nm Graviton2 processors, which utilize 64-bit Arm Neoverse N1 cores and contain around 30 billion transistors, were announced in December 2019. The chips are custom-built for cloud workloads by Israeli-based engineering firm Annapurna Labs, which was acquired by AWS around five years ago. Amazon promises a “major leap in performance and capabilities” compared to its first-generation Graviton processors, with seven times the performance, four times the compute cores, two times the cache size and five times faster memory and two times the floating point performance per core for scientific and HPC workloads. The chip also includes always-on 256-bit DRAM encryption that runs 50 percent faster than the first generation.

Amazon reports that both the new Graviton2-based C6g and R6g instances offer substantial price and performance improvements compared to x86-based instances: the new C6g instances offer up to 40 percent better price performance versus x86 for compute-intensive workloads like HPC and distributed analytics, Amazon says, and the new R6g instances offer up to 40 percent better price performance compared to x86 for memory-intensive workloads like real-time big data analytics. These more specific instances are complemented by the Graviton2-based general-purpose M6g instances (available in "preview" since late last year and launched last month), said to offer the same 40 percent price performance improvement for workloads like gaming servers, mid-size databases and microservices.

“Today more than ever, customers are looking for innovative ways to increase performance and reduce cost, and Arm processors have emerged as an exciting and mainstream alternative to x86 processors for a wide variety of existing and emerging workloads,” said David Brown, vice president of Amazon EC2. “The new Amazon EC2 instances powered by AWS-designed, Arm-based, Graviton2 processors represent a significant generational leap for customers … and already we’ve seen a broad set of customers embrace them across a wide variety of general purpose, compute optimized, and memory optimized workloads.”

These customers include heavy-hitters like Nielsen, LexisNexis and Netflix.

“We use Amazon EC2 M instance types for a number of workloads inclusive of our streaming, encoding, data processing, and monitoring applications,” said Ed Hunter, director of performance and operating systems at Netflix. “We tested the new M6g instances using industry standard LMbench and certain Java benchmarks and saw up to 50% improvement over M5 instances. We’re excited about the introduction of AWS Graviton2-based Amazon EC2 instances.”

Instance sizes range from one to 64 vCPUs (in addition to a bare-metal option) and support configurations including up to 256 GB of memory, 25 Gbps of networking and 19 Gbps of Elastic Block Store (EBS) bandwidth. Soon, the instances will also be available with local NVMe-based SSD storage options. The instances can be reserved, purchased as spot instances, purchased on demand or purchased as part of savings plans. They are available in the US East, US West, Europe and Asia Pacific regions. Amazon says that the instances already support a “broad ecosystem of operating systems and services,” including Linux distributions, container services, agents and developer/automation tools.

The availability of Graviton2 Arm-based instances is another move away from an Intel-dominated cloud market and toward cloud instance diversification from AWS, which expanded AMD-based instances to global markets less than a year ago.

Add a Comment

Do NOT follow this link or you will be banned from the site!
Share This