Skip to content

cuda: fold the recurrent-state gather into the GATED_DELTA_NET kernel, per-context registry (+3.8% TG on Ada) - #220

Closed
professorpalmer wants to merge 2 commits into
PrismML-Eng:prismfrom
professorpalmer:cuda-gdn-gather-fusion
Closed

cuda: fold the recurrent-state gather into the GATED_DELTA_NET kernel, per-context registry (+3.8% TG on Ada)#220
professorpalmer wants to merge 2 commits into
PrismML-Eng:prismfrom
professorpalmer:cuda-gdn-gather-fusion

cuda: stop claiming the GDN gather fusion frees the GET_ROWS allocation.

80132df
Select commit
Loading
Failed to load commit list.
Sign in for the full log view

Annotations

1 warning
labeler
succeeded Sep 21, 2026 in 11s