pub fn gdn_exact_chain(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
h_state: DevicePtr,
query: DevicePtr,
key: DevicePtr,
value: DevicePtr,
gate: DevicePtr,
beta: DevicePtr,
output: DevicePtr,
h_inter: [DevicePtr; 3],
num_k_heads: u32,
num_v_heads: u32,
k_dim: u32,
strides: [u32; 4],
stream: u64,
) -> Result<()>Expand description
2026-10-01: The single-sequence exact verify (gdn_exact_chain{2,3,4}, chosen by the caller’s
kernel): K rows of one sequence’s strided-decode chain in one launch. h_state is updated to
the state after the last row, h_inter[t] receives the state after row t (t < K - 1; the other
entries must still be valid pointers and are not written). Rows are strides = [qk, v, gate|beta, out] floats apart.