Google Cloud Tau T2A Ampere Altra vs. T2D AMD EPYC Performance
Over the summer Google announced Tau T2A as the first Google Cloud Compute Engine VM to run on Arm. The T2A series is powered by Ampere Altra processors and complement the Tau T2D series powered by AMD EPYC 7003 "Milan" processors that launched last year. I have been trying out the Tau T2A series for the past several weeks and have some initial benchmarks to share today for showing how the Ampere Altra Arm VMs can perform against the existing T2D series.
The Tau T2D series launched in late 2021 as a collaboration between Google and AMD for offering higher performance over alternatives from other public cloud providers -- whether they be Arm or x86_64 based instances. Since then Google has been working with Ampere Computing for bringing Ampere Altra to their public cloud and coming up with a contender worthy of their Tau branding.
Current Tau T2A pricing as of publishing.
Like with the AMD EPYC powered T2D series, T2A is also optimized for cost-effective performance of scale-out workloads. The T2A VMs come up to 48 vCPUs per virtual machine -- each vCPU backed by a physical CPU core. There is also 4GB of memory per vCPU, matching that of the T2D VMs and Amazon's M6g Graviton2-based instances. Google has talked up the T2A series, as it has with T2D, as being great for web servers, containerized microservices, data log processing, multimedia encoding/transcoding, and Java workloads.
Current Tau T2D pricing as of publishing.
Google is supporting all the major AArch64-compatible Linux distributions on the T2A series from the usual contenders like Ubuntu, RHEL, and SUSE Linux Enterprise to newer entrants like Rocky Linux. With the T2A series is also integration with other Google Cloud offerings like the Google Kubernetes Engine, Dataflow, and other software long available to their x86_64 VMs.
More background information on the Google Tau T2A series can be found via their July announcement.
For those wondering how the Google Tau T2A line-up compares to the existing AMD EPYC powered T2D series, I ran a number of benchmarks across various Tau instances while running Ubuntu 22.04 LTS.
The Ubuntu 22.04 LTS instance was at its defaults but with upgrading to the GCC 12 compiler as packaged in the Jammy archive for both the x86_64 and AArch64 VMs in order to ensure an up-to-date compiler for both AMD EPYC and Ampere Altra processors.