Function moe_decode_atomic_c4_silu_down_accum
Source pub fn moe_decode_atomic_c4_silu_down_accum(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
gate_out: DevicePtr,
up_out: DevicePtr,
packed_ptrs: DevicePtr,
scale_ptrs: DevicePtr,
scale2_vals: DevicePtr,
expert_indices: DevicePtr,
expert_weights: DevicePtr,
routed_accum: DevicePtr,
sh_gate_in: DevicePtr,
sh_up_in: DevicePtr,
sh_down: &QuantizedWeight,
sh_down_out: DevicePtr,
hidden: u32,
inter: u32,
top_k: u32,
num_tokens: u32,
stream: u64,
) -> Result<()>