TENANCY-BEFORE: []
##SHAPE out_proj 30
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         8.61
    gpu__time_duration.sum                                                                   ms         1.39
    launch__occupancy_limit_registers                                                     block            3
    launch__occupancy_limit_shared_mem                                                    block            4
    launch__registers_per_thread                                                register/thread          168
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        23.93
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        51.42
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        63.97
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        24.09
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         6.31
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.43
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         2.12
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.58
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.58
##SHAPE out_proj 61
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         9.23
    gpu__time_duration.sum                                                                   ms         1.29
    launch__occupancy_limit_registers                                                     block            3
    launch__occupancy_limit_shared_mem                                                    block            4
    launch__registers_per_thread                                                register/thread          168
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        25.77
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        37.90
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        68.91
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        24.16
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         7.13
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.44
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         2.37
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.46
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.72
##SHAPE out_proj 32
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         8.78
    gpu__time_duration.sum                                                                   ms         1.37
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          121
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        36.94
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        53.74
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        64.69
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        32.64
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         7.75
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         0.90
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         2.66
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         1.00
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.32
##SHAPE out_proj 63
    dram__throughput.avg.pct_of_peak_sustained_elapsed                                        %         9.32
    gpu__time_duration.sum                                                                   ms         1.28
    launch__occupancy_limit_registers                                                     block            2
    launch__occupancy_limit_shared_mem                                                    block            2
    launch__registers_per_thread                                                register/thread          122
    sm__inst_executed_pipe_lsu.avg.pct_of_peak_sustained_active                               %        39.44
    sm__pipe_alu_cycles_active.avg.pct_of_peak_sustained_active                               %        40.12
    sm__pipe_tensor_op_imma_cycles_active.avg.pct_of_peak_sustained_active                    %        69.08
    sm__warps_active.avg.pct_of_peak_sustained_active                                         %        32.63
    smsp__average_warp_latency_per_inst_issued.ratio                                      cycle         8.64
    smsp__average_warps_issue_stalled_barrier_per_issue_active.ratio                       inst         1.17
    smsp__average_warps_issue_stalled_math_pipe_throttle_per_issue_active.ratio            inst         2.88
    smsp__average_warps_issue_stalled_not_selected_per_issue_active.ratio                  inst         0.82
    smsp__average_warps_issue_stalled_wait_per_issue_active.ratio                          inst         1.56
TENANCY-AFTER: []

[exited with code 0]
