Welcome to OpenCAPI Consortium

The OpenCAPI Consortium is an open forum to manage the OpenCAPI specification and ecosystem. OpenCAPI is a not-for-profit organization formed in October 2016 by OpenCAPI Strategic Members AMD, Google, IBM, Mellanox Technologies and Micron to create an open coherent high performance bus interface based on a new bus standard called Open Coherent Accelerator Processor Interface (OpenCAPI) and grow the ecosystem that utilizes this interface. This initiative is being driven by the emerging accelerated computing and advanced memory/storage solutions that have introduced significant system bottlenecks in today’s current open bus protocols and requires a technical solution that is openly available.

Join Now

Strategic Members

View All Members

Join OpenCAPI

To join OpenCAPI, please download and review the OpenCAPI Consortium Membership Kit.

Once we receive your signed Membership Agreement, we will immediately begin on-boarding you into the OpenCAPI activities and adding you to our public membership roster. If you have any membership questions, please contact us at membership@opencapi.org

Latest News

The Rambus Labs Hybrid Memory Research Platform using the Power9 and the OpenCAPI Architecture to demonstrate New Emerging Memories

April 5, 2019

Media Alert: Rambus Demonstrates Collaboratively Developed Hybrid Memory Research Platform with Alpha Data, IBM, Wistron and Xilinx for Future Data Center Architectures

March 15, 2019

OpenPOWER Summit Europe 2018: Introduction to the OpenCAPI Interface – Brian Allison Presents

February 19, 2019

OpenCAPI In Action

What companies say about OpenCAPI

As the world moves into the age of artificial intelligence and cognitive computing, OpenCAPI has created a new interface standard for every player in every industry.

View Video

Download The Standard

The new standard, called OpenCAPI provides an open, high-speed pathway for different types of technology – advanced memory, accelerators, networking and storage – to more tightly integrate their functions within servers. This data-centric approach to server design, which puts the compute power closer to the data, removes inefficiencies in traditional system architectures to help eliminate system bottlenecks and significantly improve server performance.

These specifications were contributed to the OpenCAPI Consortium pursuant of the IPR Policy as input specifications and will be reviewed and updated appropriately following the work group process.

The OpenCAPI consortium looks forward to publishing draft specifications when this process completes.

OpenCAPI 3.0 Transaction Layer Specification (updated*)

OpenCAPI 3.1 Transaction Layer Specification (updated*)

*The changes in the V3.0 specification can be found in the revision log. A new section on profiles has been added with some new terms and clarifications. The V3.1 specification is a revision of the V3.0 specification (includes all the V3.0 changes) and adds new content. The revision log shows the changes. In both specifications, changes from the October drop have change bars noted.

OpenCAPI 3.0 Data Link Layer Specification

OpenCAPI 4.0 Transaction Layer Specification

Fill out the form to download the specifications