Skip to content

Performance enhancements #149

Open
Open
@Maria1Petrova

Description

@Maria1Petrova

oneCCL exists in order to facilitate high performance communications through collective operations spanning high node count CPU and/or GPU clusters.
We seek enhancements to maximize the throughput and minimize latency for all collectives and message sizes of interest to the community. We are open to all enhancements that improve performance without degrading performance elsewhere.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions