mla_q_rope_writeback

Function mla_q_rope_writeback 

Source
pub fn mla_q_rope_writeback(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    q_rope_direct: DevicePtr,
    q_absorbed_buf: DevicePtr,
    nq: u32,
    rope: u32,
    kv_lora: u32,
    mla_cache_dim: u32,
    stream: u64,
) -> Result<()>
Expand description

MLA Q_rope writeback: scatter RoPE’d rope portions to strided layout. 1 kernel replaces 32 D2D copies.