Skip to content

Not a bug - crediting the hit/miss split design #297

Description

@sangharshadhyeta

Wanted to flag this properly rather than silently benefit from it: took the hit/miss split idea for MoE prefill weight uploads from FreeToken - device-to-device gather for experts already cache-resident, one batched transfer for the rest, instead of always paying full PCIe cost. Re-derived it against our own LFRU pool rather than porting the CUDA kernels directly (different engine shape - single unified cache vs. ours), but the design is yours. Thank you.

Context: https://github.com/sangharshadhyeta/solid.cpp

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions