vllm
[GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill
#38361
Merged
Overview
Commits
14
Changes
View On GitHub
Commits
  • [GDN] Eliminate GPU->CPU sync in prepare_chunk_indices during prefill
    arpera committed 101 days ago
  • Fix gemini-code issues: extract _insert helper in tensor_cache, add TODO for chunk_size constant
    arpera committed 101 days ago
  • Fix mypy: add type: ignore for dynamic register attribute
    arpera committed 101 days ago
  • Extract hardcoded chunk_size=64 into FLA_CHUNK_SIZE constant
    arpera committed 101 days ago
  • Fix: skip chunk_indices pre-registration on pure decode steps
    arpera committed 101 days ago
  • [GDN] Pre-compute chunk_indices/chunk_offsets in metadata builder
    arpera committed 101 days ago
  • Remove dead register() code and duplicate prefill block
    arpera committed 101 days ago
  • Fix Claude review comments: backend guard, kda chunk_size, BT simplification
    arpera committed 101 days ago
  • Remove use_flashinfer backend guard for chunk_indices pre-computation
    arpera committed 101 days ago
  • Merge branch 'main' into artem/remove-extra-d2h-copy
    vadiklyutiy committed 101 days ago
  • Merge branch 'main' into artem/remove-extra-d2h-copy
    arpera committed 99 days ago
  • fix CI: lazy-import FLA ops to avoid CUDA init in forked subprocess
    arpera committed 99 days ago
  • Merge branch 'main' into artem/remove-extra-d2h-copy
    arpera committed 98 days ago
  • Merge branch 'main' into artem/remove-extra-d2h-copy
    arpera committed 98 days ago
FAQ Terms Privacy Refunds Impressum
Loading