Function w8a8_proj 

pub fn w8a8_proj(
    gpu: &dyn GpuBackend,
    kernels: &W8a8Kernels,
    w: &W8a8Weight,
    x: DevicePtr,
    ldx: u32,
    rows: usize,
    out: DevicePtr,
    ldc: u32,
    scratch: &W8a8Scratch,
    stream: u64,
) -> Result<()>

2026-09-28: One W8A8 projection: quantize x ([rows, ldx] BF16) and multiply, writing out ([rows, ldc] BF16, the first w.n() columns).