Innovation at the Core of Ampere’s Roadmap and Vision


Jeff Wittich, Chief Product Officer 

 

A few weeks ago, Ampere unveiled our annual strategy and roadmap update. It’s clear that innovation is at Ampere’s core. Since day one, we’ve been hard at work building a true cloud native environment from the ground up and a fundamental part of that is our newly-revealed Ampere core that will be used in our future processors.

Since our roadmap update, Oracle announced its OCI instances of Ampere® Altra®. We are pleased with the positive user feedback for the Ampere A1. It’s been fun to see what developers are building on top of Ampere A1 and to see the awesome feedback. GitLab, Jenkins, Rancher, Datadog, OnSpecta, NGINX, and Genymobile are among those working with Oracle on this technology. You can also read Oracle’s blog here for advanced users for A1 compute tuning.

To learn more about our vision and roadmap, you can view Renee James, Atiq Bajwa, and myself discuss it here.  We were also joined by some of the top cloud providers such as Microsoft, Oracle and Tencent who discussed their excitement and commitment to Ampere® Altra® and our next generation Ampere processor.

We finally have a chance to talk about the development of our own core, which we’ve been working on for the past several years. This next generation Ampere processor, scheduled to be released in 2022, will deliver the highest performance, most scalable, and most power efficient processor with new capabilities desired by cloud providers. Our 5nm processor will support more cores and increased memory bandwidth and IO/network bandwidth.

So, why design our own core? To meet the innovation requirements of the cloud providers, we wanted to control the microarchitecture to continue delivering the features they need. The microarchitecture itself must be designed with multi-tenancy and scalability in mind. The Arm ISA is a great base for this, and our own core delivers differentiated IP to continue to deliver the best cloud processors in the industry.

Innovation at all levels from the silicon to the firmware to the server and out across the entire software ecosystem is what we do at Ampere, and the core is where it all starts. With our world-class silicon design team, we will introduce some amazing capabilities to our Ampere processors. By designing our own core, we can deliver higher performance with more cores and lower power at a faster cadence.

 

Proven History of New Cores

Ampere has been developing our own cores since it was founded. Leading cloud providers are deploying advanced services based on Ampere® Altra® and preparing for Altra® Max. While this has been a tremendous achievement, we must continue innovating aggressively in areas that are critical for cloud-native workloads. The world needs a true cloud-native processor to run their cloud-based software environments and that will only happen with Ampere’s modern-day CPU that delivers:

  • Predictable High Performance
  • Platform Scalability
  • Power Efficient Compute Density

These are the critical attributes for truly cloud-native processing. As you can see from our current line-up, roadmap and next generation processor core, Ampere is committed to delivering on all three! It’s clear that legacy technologies were not designed to power true cloud native and Ampere, along with its growing ecosystem of partners and providers, is filling that gaping void. We’ve already demonstrated our ability to execute on our product roadmap and we will continue to release new products every year that deliver the highest performance per vCPU, server, and rack; the highest core count; and the highest performance per Watt. These are the metrics that matter to cloud native computing.

We look forward to sharing more details on our next generation Ampere processor in the future. Until then, our teams will continue doing what we do best – Innovating! After all, that is the core of Ampere.



Related Post

  • Register For Early Release Product Information
    Name*
    This field is for validation purposes and should be left unchanged.