Next-Generation Ampere® Instances Available on Oracle Cloud Infrastructure

Like OCI Ampere A1 Compute instances, OCI Ampere A2 Compute instances show strong performance for multiple AI functions, including generative AI. This performance is made possible through joint development efforts between Ampere Computing and OCI that have recently delivered up to 152% performance gain over the previous upstream llama.cpp open-source implementation.

Beyond AI, OCI Ampere A1 and A2 Compute instances are also well-suited for other cloud native workloads, such as analytics and databases, media services, video streaming, and web services. They offer the linear scalability, low latency, density and predictable performance these workloads need, bringing more performance and higher cost savings. For example, when deploying a typical web service on OCI Ampere A2 Compute, using very popular applications such as MySQL, NGINX, Cassandra, and Redis, the savings can be very compelling. An enterprise spending $50M annually across a weighted blend of these popular web service components could save up to $21.4M in cloud infrastructure costs compared with OCI E5 x86 based shapes. The result can save up to 43% less in infrastructure costs, 30% reduction in power consumption, and 33% less carbon emissions.

OCI Ampere A1 and A2 Compute shapes represent a significant advancement in cloud computing by lowering general purpose cloud computing costs, addressing AI computing efficiency, providing a more predictable and linearly scalable compute resource, and helping companies achieve ESG goals faster. Visit the OCI Ampere Compute website to get started.