OpenAI has announced a collaboration with AMD, NVIDIA, and others to launch the Open Network Protocol MRC, aiming to improve AI training efficiency and reduce GPU waste.

BlockBeats News, May 6th, OpenAI announced a new open networking protocol called MRC (Multipath Reliable Connection) in collaboration with AMD, Broadcom, Intel, Microsoft, and NVIDIA. OpenAI stated that the protocol can help large-scale AI training clusters achieve faster and more stable operation, while reducing GPU resource wastage.

Currently, MRC has been deployed on all of OpenAI's large supercomputers used to train cutting-edge models, including its facility in Abilene, Texas with Oracle OCI, and the Microsoft Fairwater supercomputer. Furthermore, MRC has been open for industry-wide use and development through the Open Compute Project.