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 595–598

sfCopyBody

Full file
LDGSTS of chunk j (its pointer pair) to shared sfStore + offset. SB5 is its read barrier, one live token: each chunk waits for the previous one's.
595def sfCopyBody = (lambda unrestricted offset : Nat . (lambda unrestricted j : Nat . (lambda unrestricted wait : Nat .
596  (constructor SM86InstructionBody SM86LoadGlobalToShared (saR sfStore) (saU offset) (saR (sfPair j)) (saU 0)
597    (constructor SM86Control SM86ControlValue (byte 15) (constructor SM86YieldMode SM86Continue)
598      saNone saSB5 (nat-to-byte wait) (byte 0))))))

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.