[FA2] tiling-qk + fully smem swizzle✔️

[DefTruth]

• no padding, 0 ldsm bank_conflicts

void flash_attn_mma_stages_split_q_tiling_qk_fully_swizzle_kernel<64, 16, 8, 16, 4, 1, 4, 1, 1, 8, 1, 8, 1, 0, 0, 0>(__half *, __half *, __half *, __half *, int, int) (16, 48, 1)x(128, 1, 1), Context 1, Stream 7, Device 0, CC 8.9
    Section: Command line profiler metrics
    ------------------------------------------------------------------ ----------- ------------
    Metric Name                                                        Metric Unit Metric Value
    ------------------------------------------------------------------ ----------- ------------
    sm__sass_l1tex_data_bank_conflicts_pipe_lsu_mem_shared_op_ldsm.avg                        0
    sm__sass_l1tex_data_bank_conflicts_pipe_lsu_mem_shared_op_ldsm.max                        0
    sm__sass_l1tex_data_bank_conflicts_pipe_lsu_mem_shared_op_ldsm.min                        0
    sm__sass_l1tex_data_bank_conflicts_pipe_lsu_mem_shared_op_ldsm.sum                        0
    ------------------------------------------------------------------ ----------- ------------

• no padding, 0 ld & st bank_conflicts

  void flash_attn_mma_stages_split_q_tiling_qk_fully_swizzle_kernel<64, 16, 8, 16, 4, 1, 4, 1, 1, 8, 1, 8, 2, 0, 0, 0>(__half *, __half *, __half *, __half *, int, int) (16, 48, 1)x(128, 1, 1), Context 1, Stream 7, Device 0, CC 8.9
    Section: Command line profiler metrics
    -------------------------------------------------------- ----------- ------------
    Metric Name                                              Metric Unit Metric Value
    -------------------------------------------------------- ----------- ------------
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_ld.avg                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_ld.max                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_ld.min                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_ld.sum                        0
    -------------------------------------------------------- ----------- ------------

  void flash_attn_mma_stages_split_q_tiling_qk_fully_swizzle_kernel<64, 16, 8, 16, 4, 1, 4, 1, 1, 8, 1, 8, 2, 0, 0, 0>(__half *, __half *, __half *, __half *, int, int) (16, 48, 1)x(128, 1, 1), Context 1, Stream 7, Device 0, CC 8.9
    Section: Command line profiler metrics
    -------------------------------------------------------- ----------- ------------
    Metric Name                                              Metric Unit Metric Value
    -------------------------------------------------------- ----------- ------------
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_st.avg                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_st.max                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_st.min                        0
    l1tex__data_bank_conflicts_pipe_lsu_mem_shared_op_st.sum                        0
    -------------------------------------------------------- ----------- ------------