pub fn gdn_prefill_split4_batched(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
h_state_ptrs: DevicePtr,
query: DevicePtr,
key: DevicePtr,
value: DevicePtr,
gate: DevicePtr,
beta: DevicePtr,
output: DevicePtr,
batch_size: u32,
seq_len: u32,
num_k_heads: u32,
num_v_heads: u32,
k_dim: u32,
v_dim: u32,
qk_stride: u32,
v_stride: u32,
gb_stride: u32,
stream: u64,
) -> Result<()>Expand description
Batched split4 GDN prefill (uses gated_delta_rule_prefill_split4_batched).