Amazon Web Services has fired up two new EC2 instance types running its homegrown Arm-based Graviton2 processors, repeated the claims that they significantly out-perform the x86 silicon on which it built its business, and reckoned Arm will outpace other architectures.
In May 2020, AWS unleashed Graviton2 in the EC2 M6g instance type it recommends for “application servers, microservices, gaming servers, mid-size data stores, and caching fleets." Plenty of users need that sort of application, though AWS classifies the M6g instance type as a general-purpose option.
Now the cloud colossus has added Graviton2-powered instances to its compute-optimized and memory-optimized server menu.
Gravitons, Neoverse... you'd be forgiven for thinking AWS's second-gen 64-core Arm server processor was a sci-fiREAD MORE
The compute-optimized option is called the C6g, and AWS said it offers “up to 40 percent better price performance” than the C5 instance type running second-generation Intel Cascade Lake and first generation Intel Skylake-SP Xeons. As ever, the C6g instance comes with lots of variants, starting at a single vCPU with 2GB of RAM and 10GE networking and a 4,750 Mbps pipe to the Amazon Elastic Block Store (EBS) service. At the upper end you you can rent 64 vCPUs – the Graviton2 has 64 cores – with 128GB of RAM, 25GE networking and a 19,000 Mbps EBS link.
The memory-optimized instances are called R6g, and are again said to perform up to 40 percent better than current AWS instances; the R5 type runs Intel Xeon Platinum 8175 CPU. R6g instances have the same upper and lower bounds as the C6g.
The Graviton2 powering these instances is built with a 7nm process node, and offers 64-bit Arm Neoverse N1 CPU cores each with 1MB of L2 cache plus 32MB of shared L3 cache.
The launch of the new instances was accompanied by a video – see below – in which AWS veep and distinguished engineer James Hamilton explained that scale is the reason the company is betting on Arm. Hamilton goes over old ground by explaining that Arm dominates mobile devices and has enormous market share in embedded devices, then explains that he feels that market share translates into greater innovation.
“This vast volume of ARM processors is particularly important in our industry because nothing drives low cost economics better than volume, nothing funds the R&D investment required to create server processors better than volume, builds devtool ecosystems faster than volume and it's volume that brings application developers," Hamilton said.
“I started in this industry back in about 1986 working on mainframe ADA compilers. I subsequently spent years working on IBM db2 hosted on UNIX super servers and then moved to Microsoft sequel server hosted on x86 processors.
“I joined Amazon Web Services bit more than eleven years ago. Each of these progressive role changes was to a higher-volume, faster-growing platform. Each product had more happy customers and produced far more revenue than the previous. Betting on volume has treated me well over the years and as a consequence more than a decade ago I got super interested in ARM cores and using ARM cores and servers.”
Hamilton also revealed that Nitro, the virtualisation sub-system that sits beneath EC2, runs Arm kit that was AWS’s first serious foray into designing its own silicon. He also said Nitro removes the need for a “virtualisation tax,” a phrase often used to deride AWS partner VMware! ®