Skip to content

P
Paddle

PaddlePaddle / Paddle
1 年多 前同步成功

通知 2302
Star 20931
Fork 5422
P
Paddle

fix print bug of profile report !22804

!22804 已合并 由 saxon_zh@saxon_zh 创建

Created by: zhangting2020

fix print bug of profile report:

  • before：some events of kOrdinary is not printed. image
  • after:
------------------------->     Profiling Report     <-------------------------

Place: All
Time unit: ms
Sorted by total time in descending order in the same thread

Total time: 3.0977
  Computation time       Total: 2.34272     Ratio: 75.6277%
  Framework overhead     Total: 0.754981    Ratio: 24.3723%

-------------------------     GpuMemCpy Summary     -------------------------

GpuMemcpy                Calls: 3           Total: 0.125296    Ratio: 4.04481%
  GpuMemcpyAsync         Calls: 3           Total: 0.125296    Ratio: 4.04481%

-------------------------       Event Summary       -------------------------

Event                                                                         Calls       Total       CPU Time (Ratio)        GPU Time (Ratio)        Min.        Max.        Ave.        Ratio.
thread0::lookup_table                                                         1           1.58523     1.579889 (0.996629)     0.005344 (0.003371)     1.58523     1.58523     1.58523     0.511745
thread0::recurrent                                                            1           1.23992     1.157229 (0.933310)     0.082690 (0.066690)     1.23992     1.23992     1.23992     0.400271
  thread0::recurrent/mul                                                      2           0.455774    0.388764 (0.852975)     0.067010 (0.147025)     0.146808    0.308966    0.227887    0.147133
  thread0::recurrent/sum                                                      1           0.133077    0.130037 (0.977156)     0.003040 (0.022844)     0.133077    0.133077    0.133077    0.0429599
  thread0::recurrent/elementwise_add                                          1           0.114945    0.111457 (0.969655)     0.003488 (0.030345)     0.114945    0.114945    0.114945    0.0371066
  thread0::recurrent/relu                                                     1           0.081749    0.078773 (0.963596)     0.002976 (0.036404)     0.081749    0.081749    0.081749    0.0263902
  thread0::recurrent/rnn_memory_helper                                        2           0.134996    0.130548 (0.967051)     0.004448 (0.032949)     0.046149    0.088847    0.067498    0.0435794
    thread0::recurrent/rnn_memory_helper/GpuMemcpyAsync(same_gpu):GPU->GPU    2           0.091876    0.087428 (0.951587)     0.004448 (0.048413)     0.030929    0.060947    0.045938    0.0296594
  thread0::recurrent/GpuMemcpyAsync(same_gpu):GPU->GPU                        1           0.03342     0.031692 (0.948294)     0.001728 (0.051706)     0.03342     0.03342     0.03342     0.0107886
thread0::transpose2                                                           1           0.162743    0.158871 (0.976208)     0.003872 (0.023792)     0.162743    0.162743    0.162743    0.0525367
thread0::fill_constant_batch_size_like                                        1           0.084823    0.082391 (0.971329)     0.002432 (0.028671)     0.084823    0.084823    0.084823    0.0273826
thread0::feed                                                                 1           0.024982    0.024982 (1.000000)     0.000000 (0.000000)     0.024982    0.024982    0.024982    0.00806469