Amazon is taking the chip business seriously, but it might not bode well for other prominent processor manufacturers like Intel and AMD. At the recently-concluded AWS re:Invent event, the tech giant announced the launch of Graviton2 — a server chip for streamlining processes in the cloud. This announcement invoked interest from around the world as the processor has set a benchmark in the cloud chip technology.
The History
Amazon had revealed the first generation (A1) of Arm-based, Graviton-powered EC2 instances at re:Invent 2018. Graviton was based on 64bit Armv8 Cortex-72 microarchitecture. It contained 2MB of L2 cache for the four quad-core clusters.
Over the last couple of years, many organisations deployed it to run different types of scale-out workloads, including containerised microservices, web servers, and data/log processing. Graviton was designed for running a few instances and was limited to only handle workloads such as messaging and serverless Lambda functions.
However, Amazon was surprised by the proliferation of Graviton adoption among many companies. The Operating System Vendors (OSV) and Independent Software Vendor (ISV) communities were early to embrace the Arm architecture and the A1 instances. They could choose an Amazon Machine Image (AIM) from a wide range of options such as Linux & Unix distributions including Amazon Linux 2, Ubuntu, Red Hat, SUSE, Fedora, Debian, and FreeBSD.
Besides, one could also choose between three container services (Docker, Amazon ECS, and Amazon Elastic Kubernetes Service), multiple system agents, including developer tools like AWS Developer Tools or Jenkins, among others.
Graviton2
Amazon provided a sneak peek at the next generation of Arm-based EC2 instances. These instances are built on AWS Nitro System, which will be powered by the new Graviton2 processor. Amazon designed a bespoke chip that is built using a 7 nm (nanometer) manufacturing process. The chip is based on 64-bit Arm Neoverse cores that can deliver up to 7x the performance of the A1 instances, including twice the floating-point performance per core for high-performance workloads. The additional memory channels and double-sized per-core caches help in expediting the memory access by up to 5x.
All of these performance enhancements come together to give these new instances a significant performance benefit over the 5th generation (M5, C5, R5) of EC2 instances. The initial benchmarks show the following per-vCPU performance improvements over the M5 instances:
- HTTPS load balancing with Nginx
- Memcached
- X.264 video encoding
- EDA simulation with Cadence Xcellium
The idea is to use those instances to power Amazon EMR, Elastic Load Balancing, Amazon ElastiCache, and other AWS services, for assisting customers in innovating on the cloud. “We are going big for our customers and our internal workloads with the new chip,” said Raj Pai, vice president of AWS EC2.
What It Means To The Chip Manufacturing Firms
Microsoft has embraced Graphcore for empowering users to render machine learning-based solutions. And now, Amazon, with the new Graphcore2, will push high-performance computing in the cloud computing for its customers. With this, both the major cloud service providers have now moved away from Intel and AMD server chips. However, AMD’s cloud processor has its relevance in Dell’s cloud computing offerings. Dell has recently adopted AMD Epyc to reduce latency in their services for delivering superior computing. But Intel is losing grounds as the Xeon is not helping it in gaining clients.
On the other hand, Intel, in the short term might not witness any significant loss as it generates only a few billion from AWS. For an organisation that roughly gets $71 billion in revenue in a quarter, AWS’s new chip will not have a major impact. But, in years to come, AWS may deviate from Intel’s processor as Graviton2 has laid the foundation to make advanced chips.
Outlook
Of late, there have been a lot of developments in the processor manufacturing; and blue-chip companies are a race to take the lead. The pioneers of chip manufacturing are under lots of pressure due to the increase in competition. However, it has only helped firms to innovate and make advanced processors, which is currently the need of the hour in the artificial intelligence landscape.
High-performance computing is paramount for businesses to develop AI-based solutions, and AWS seems to have made the right move in integrating state-of-the-art server processor.
