pub fn moe_weighted_sum_blend_batch2(
gpu: &dyn GpuBackend,
kernel: KernelHandle,
output: DevicePtr,
expert_out: DevicePtr,
expert_weights: DevicePtr,
shared_out: DevicePtr,
input: DevicePtr,
gate_weight: DevicePtr,
hidden: u32,
top_k: u32,
k: u32,
stream: u64,
) -> Result<()>Expand description
Fused weighted sum + sigmoid blend for K=2 tokens.
blockIdx.y = token index (0 or 1). Each block computes gate scalar independently from per-token input and shared gate weight.
Grid: (ceil(hidden/256), 2, 1) Block: (256, 1, 1)