pub fn w4a16_gemv_silu_input(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
gate_out: DevicePtr,
up_out: DevicePtr,
weight: &QuantizedWeight,
output: DevicePtr,
n: u32,
k: u32,
stream: u64,
) -> Result<()>Expand description
2026-09-25: W4A16 GEMV whose activation is silu(gate_out) * up_out,
computed per element from the two [k] BF16 vectors inside the kernel.