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 251–256

saScores

Full file
S = Q K^T: slice by slice, the eight n-tiles of a slice in flight together
251def saScores = (lambda unrestricted tail : (family SM86Program) .
252  (saFor 4 (lambda unrestricted k : Nat .
253    (saFor 8 (lambda unrestricted nt : Nat .
254      (saHmma (saS nt) (saQ k) (saB k nt) saSB2
255        (naturalSelect (naturalIsZero nt) (naturalSelect (naturalIsZero k) (naturalAdd saWait0 saWait1) saWait2) saWaitNone)))))
256    tail))

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.