chainerMN ChainerMN depends on MPI for inter-node communication, such that it can benefit from faster network infrastructure such as InfiniBand. NVIDIA’s NCCL is then used for intra-node communication between GPUs.

The following shows the benefits of leveraging InfiniBand for performance and allows ChainerMN to achieve ~72 % accuracy.

For more information on ChainerMN, read more here.

NVIDIA Mellanox Cookie Policy

This website uses cookies which may help to deliver content tailored to your preferences and interests, provide you with a better browsing experience, and to analyze our traffic. You may delete and/or block out cookies from this site, but it may affect how the site operates. Further information can be found in our Privacy Policy.