Skip to content

DSA indexer forward: add a lean SM100 fast path for the head_dim=128 / qhead_per_kv_head=64 regime (up to 1.46x) - #416

Open
zkyue wants to merge 4 commits into
NVIDIA:developfrom
zkyue:feat/dsa-indexer-fwd-lean
Open

DSA indexer forward: add a lean SM100 fast path for the head_dim=128 / qhead_per_kv_head=64 regime (up to 1.46x)#416
zkyue wants to merge 4 commits into
NVIDIA:developfrom
zkyue:feat/dsa-indexer-fwd-lean

lift THD gate for indexer-forward lean fast path

91db75c
Select commit
Loading
Failed to load commit list.

Workflow runs completed with no jobs