pub fn gdn_decode_f32_norm(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
h_state: DevicePtr,
query: DevicePtr,
key: DevicePtr,
value: DevicePtr,
gate: DevicePtr,
beta: DevicePtr,
z_gate: DevicePtr,
norm_weight: DevicePtr,
output: DevicePtr,
batch_size: u32,
num_k_heads: u32,
num_v_heads: u32,
k_dim: u32,
v_dim: u32,
eps: f32,
stream: u64,
) -> Result<()>Expand description
FP32 GDN decode fused with gated RMS norm.
Produces the same BF16 post-gated-norm output that a separate
gdn_decode_f32 + gated_rms_norm_f32_input pair would produce, while
avoiding the intermediate FP32 global write/read.