pub fn silu_mul(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
gate: DevicePtr,
up: DevicePtr,
output: DevicePtr,
num_elements: u32,
stream: u64,
) -> Result<()>Expand description
Fused SiLU activation: output = SiLU(gate) * up.
Kernel: silu_mul_separate(gate, up, output, n)
Grid: (ceil(n/256), 1, 1) Block: (256, 1, 1)