The tech landscape is undergoing a seismic shift, with artificial intelligence (AI) emerging as the defining technology of our era. As this AI revolution unfolds, the demand for computing infrastructure requires superior performance, better scalability, and energy efficiency to handle the immense computational requirements of cloud native workloads.

Oracle just announced their new OCI Ampere A2 Compute Shapes based on the AmpereOne® processor.

Ampere launched its next generation AmpereOne® processor, with custom-designed cores, to empower cloud providers, enterprises, and edge computing applications to harness the full potential of cloud native and AI inference workloads.

Since running applications rather than synthetic benchmarks is what matters to cloud developers, we recently benchmarked and analyzed performance of a wide variety of applications commonly used in the cloud – from web hosting, media encoding, AI inference to storage and database. Some of those workloads and the performance, as well as price-performance results, are shared below.

Llama3 8B – Large Language Model Inference

Meta’s Llama 3, the most popular open-source Large Language Model (LLM) to date with the latest advancements in open-source LLMs, is now optimized for AmpereOne® based A2 instances on Oracle Cloud Infrastructure (OCI), offering an unprecedented performance and flexibility.

AmpereOne-based A2 instances showed 27% performance improvement relative to AMD's Genoa-based E5 instances. The performance was measured as total tokens generated per sec. AmpereOne A2 instances showed 95% better price-performance over AMD’s Genoa based E5 instance on OCI.

MySQL – Relational database management system
MySQL is the world’s most popular open-source database. According to DB-Engines, MySQL ranks second, behind Oracle Database, empowering the most popular applications, including Facebook, Twitter, Netflix, Uber, Airbnb, Shopify, and Booking.com.

The AmpereOne® based A2 OCPUs delivered 40% higher performance and 2x better price-performance compared to AMD EPYC E5 OCPUs at similar latencies.
More Info about MySQL-A2

NGINX- Web serving

NGINX is a popular web server that can also be used for other functions such as load balancing and reverse proxying. By some counts, close to one-third of today’s websites are powered by NGINX.

In our tests, the AmpereOne® based A2 instances delivered up to 20% higher throughput and 80% better price-performance compared to AMD EPYC E5 instances at a p99 latency of 1ms.
More Info about NGINX

Postgres
Postgres is a popular open-source object-Relational Database Management System (“RDBMS”), that uses and extends the SQL language combined with many features that safely store and scale the most complicated data workloads.
More Info about Postgres

As the charts show, the price-performance of the AmpereOne A2 VM is 1.6X of the AMD E5 VM.

We set out to design AmpereOne® to specifically address the needs of the modern cloud and AI. The metrics that are valued by cloud developers and CSPs are different from those in other segments – predictable, high performance while sipping power, high core count for superior density, and a design built from the ground up for the Cloud Native software paradigm.

In this blog, I have compared the performance of AmpereOne® based A2 instances on OCI to the current state-of-the-art AMD Genoa based E5 instances. The performance and price-performance benefits of AmpereOne® based A2 instances for modern cloud workloads make AmpereOne® A2 an extremely competitive and compelling product.

At Ampere Computing, we are laser focused on innovation. We will continue to deliver the world’s most performant, scalable, and power efficient processors that are uniquely designed for the needs of the modern cloud and AI.

For more information:

• Introducing OCI Ampere A2: Oracle's Next-Generation Arm-Based Cloud Compute
• AmpereOne® now Generally Available on Oracle Cloud Infrastructure
• Ampere OCI Partner Page
• OCI Ampere Compute

Disclaimer
All data and information contained herein is for informational purposes only and Ampere reserves the right to change it without notice. This document may contain technical inaccuracies, omissions and typographical errors, and Ampere is under no obligation to update or correct this information. Ampere makes no representations or warranties of any kind, including express or implied guarantees of noninfringement, merchantability, or fitness for a particular purpose, and assumes no liability of any kind. All information is provided “AS IS.” This document is not an offer or a binding commitment by Ampere.

System configurations, components, software versions, and testing environments that differ from those used in Ampere’s tests may result in different measurements than those obtained by Ampere.

©2024 Ampere Computing LLC. All Rights Reserved. Ampere, Ampere Computing, AmpereOne and the Ampere logo are all registered trademarks or trademarks of Ampere Computing LLC or its affiliates. All other product names used in this publication are for identification purposes only and may be trademarks of their respective companies.