AlphaGitHub
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
Download exact fileView at revisionbcc11a4e7647d31628

def · lines 3129–3135

streamingAttentionKeyHeadsFirstFixedProductsSM86

Full file
The same with fixed-latency products (SB2 and SB3 freed) before the late-read guard serializes and compacts it.
3129def streamingAttentionKeyHeadsFirstFixedProductsSM86 =
3130  (lambda unrestricted seq : Nat . (lambda unrestricted heads : Nat .
3131    (lambda unrestricted keyValueHeads : Nat .
3132      (lambda erased admitted : (equal Nat (streamingAttentionGroupedSM86Admitted seq heads keyValueHeads) 1) .
3133        (sm86GuardLateReads
3134          (sm86TensorProductsFixed 12
3135            (streamingAttentionKeyGroupedOrderedUncheckedSM86 1 seq heads keyValueHeads 0 1)))))))

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.