Function w8a16_gemv_batch16_strided 

pub fn w8a16_gemv_batch16_strided(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    input: DevicePtr,
    weight: DevicePtr,
    block_scale: DevicePtr,
    output: DevicePtr,
    m: u32,
    n: u32,
    k: u32,
    a_row_stride: u32,
    c_row_stride: u32,
    stream: u64,
) -> Result<()>

MAX_M=16 sibling of w8a16_gemv_batch4_strided, for decode concurrency 5..=16. Same template, same launch geometry, same per-row accumulation order; the wider register array is the only difference.

Kernel: w8a16_gemv_batch16_strided (module w8a16_gemv_batch4). Grid: (ceil(N/4), 1, 1) Block: (256, 1, 1)