Reorder channelwise gated delta rule chunked hot loops for autovectorization (#21021) (#21021)
Summary: Reorder the chunked prefill inner loops (steps 1, 4, 5, 6) so the innermost loop runs contiguously over the head dimension (k or v) instead of striding down a column of the state / pv. This lets the compiler autovectorize the now-unit-stride AXPYs; hand-written at::vec was tried and was slower than the compiler output, so the loops stay scalar. Reviewed By: billmguo Differential Revision: D112598714
J
Jacob Stevens committed
17c17e428a9e1486c9e11df8505944c2809f4936
Parent: 96e87e3
Committed by GitHub <noreply@github.com>
on 7/23/2026, 10:42:16 PM