Skip to content

[vllm] Investigate whether NCCL_CUMEM_ENABLE should be set to 0 #1630

@erictang000

Description

@erictang000

We tried to remove this in the following PR: #1600, but this caused OOMs on previously working configs. However, it should in theory be possible to remove this to enable NCCL optimizations. A deeper investigation into when this causes increased memory pressure/fragmentation is needed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions