Source/Packages

Realization.Nvidia.SM86.StreamingAttentionSM86

packages/realizations/cooperative/nvidia-sm86/src/Realization/Nvidia/SM86/StreamingAttentionSM86.alpha

3,230 lines493 declarations193.3 KiBSHA-256 cc0366b3d857

def · lines 3109–3113

streamingAttentionForwardHeavyFirstSM86

Full file
Order-1 forms (heads fastest; forward and query tiles counted down). The launch grid is (K/V heads, groups, tiles); see saAxis.
3109def streamingAttentionForwardHeavyFirstSM86 =
3110  (lambda unrestricted seq : Nat . (lambda unrestricted heads : Nat .
3111    (lambda unrestricted keyValueHeads : Nat .
3112      (lambda erased admitted : (equal Nat (streamingAttentionGroupedSM86Admitted seq heads keyValueHeads) 1) .
3113        (streamingAttentionForwardGroupedOrderedUncheckedSM86 1 seq heads keyValueHeads)))))

The compiler supplied declaration spans and resolved links from this source snapshot. This page does not assert that this file belongs to a checked closure.