##SHAPE out_proj 27
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         8.43
    gpu__time_duration.sum                                                                   ms         1.42
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          201
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        21.90
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        48.86
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        62.51
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        16.32
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         4.60
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.19
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         1.67
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.39
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         0.95
##SHAPE out_proj 60
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         9.30
    gpu__time_duration.sum                                                                   ms         1.29
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          199
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        24.10
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        36.56
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        68.78
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        16.32
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         5.12
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.15
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         1.95
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.33
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.16
##SHAPE out_proj 64
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         8.37
    gpu__time_duration.sum                                                                   ms         1.43
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          201
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        21.89
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        48.85
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        62.50
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        16.32
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         4.60
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.19
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         1.67
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.39
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         0.95
##SHAPE out_proj 67
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         9.20
    gpu__time_duration.sum                                                                   ms         1.30
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          199
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        24.11
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        36.58
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        68.81
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        16.32
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         5.12
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.15
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         1.95
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.33
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.16
##SHAPE out_proj 40
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %        10.10
    gpu__time_duration.sum                                                                   ms         1.19
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          198
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        26.22
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        37.44
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        74.83
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        16.31
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         6.30
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.18
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         2.46
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.28
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.75

[exited with code 0]
