==8807== Profiling result:
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
 GPU activities:   57.35%  17.3676s      1665  10.431ms  276.76us  155.83ms  void applyHaarClassifierAnchorParallel<bool=0, bool=1, bool=0, bool=1, bool=1>(unsigned int*, unsigned int, float*, unsigned int, HaarFeature64*, HaarClassifierNode128*, HaarStage64*, unsigned int*, unsigned int*, unsigned int, unsigned int, NcvSize32u, unsigned int, unsigned int, float)
                   34.84%  10.5522s      1665  6.3376ms  103.40us  94.287ms  void applyHaarClassifierAnchorParallel<bool=1, bool=1, bool=0, bool=0, bool=1>(unsigned int*, unsigned int, float*, unsigned int, HaarFeature64*, HaarClassifierNode128*, HaarStage64*, unsigned int*, unsigned int*, unsigned int, unsigned int, NcvSize32u, unsigned int, unsigned int, float)
                    2.97%  900.80ms      1665  541.02us  7.8440us  6.0432ms  void applyHaarClassifierClassifierParallel<bool=1, bool=0, bool=1>(unsigned int*, unsigned int, float*, unsigned int, HaarFeature64*, HaarClassifierNode128*, HaarStage64*, unsigned int*, unsigned int*, unsigned int, unsigned int, NcvSize32u, unsigned int, unsigned int, float)
                    0.86%  261.30ms       333  784.69us  622.89us  5.8686ms  void transpose<unsigned int>(unsigned int*, unsigned int, unsigned int*, unsigned int, NcvSize32u)
                    0.63%  192.17ms       111  1.7313ms  1.4662ms  5.6529ms  void scanRows<unsigned int, __int64, bool=0>(unsigned int*, unsigned int, unsigned int, unsigned int, __int64*, unsigned int)
                    0.53%  159.24ms      1665  95.639us  1.9200us  2.5828ms  void rectStdDev_32f_C1R<bool=1>(unsigned int*, unsigned int, __int64*, unsigned int, float*, unsigned int, NcvSize32u, NcvRect32u, float)
                    0.52%  158.58ms      5109  31.039us     320ns  3.6070ms  [CUDA memcpy HtoD]
                    0.46%  139.81ms      1665  83.970us  4.5460us  1.4372ms  void decimate_C1R<__int64, bool=1>(__int64*, unsigned int, __int64*, unsigned int, NcvSize32u, unsigned int)
                    0.34%  102.00ms       111  918.90us  693.73us  1.8336ms  void transpose<__int64>(__int64*, unsigned int, __int64*, unsigned int, NcvSize32u)
                    0.30%  90.375ms      1665  54.279us  3.9050us  1.1182ms  void decimate_C1R<unsigned int, bool=1>(unsigned int*, unsigned int, unsigned int*, unsigned int, NcvSize32u, unsigned int)
                    0.30%  90.174ms      1665  54.158us  1.6000us  414.83us  [CUDA memcpy DtoD]
                    0.30%  89.864ms       111  809.59us  577.59us  7.5202ms  void scanRows<unsigned int, unsigned int, bool=0>(unsigned int*, unsigned int, unsigned int, unsigned int, unsigned int*, unsigned int)
                    0.23%  71.084ms       111  640.40us  495.63us  3.3368ms  void scanRows<unsigned char, unsigned int, bool=1>(unsigned char*, unsigned int, unsigned int, unsigned int, unsigned int*, unsigned int)
                    0.19%  58.290ms       111  525.13us  496.85us  3.3756ms  void scanRows<unsigned char, unsigned int, bool=0>(unsigned char*, unsigned int, unsigned int, unsigned int, unsigned int*, unsigned int)
                    0.16%  47.204ms       111  425.26us  401.93us  2.8850ms  void cv::cudev::grid_transform_detail::transformSimple<cv::cudev::GlobPtr<uchar3>, unsigned char, cv::cudev::BGR_to_GRAY_func<unsigned char>, cv::cudev::WithOutMask>(uchar3, cv::cudev::GlobPtr<cv::cudev::GlobPtr<uchar3>>, unsigned char, unsigned char, int, int)
                    0.01%  2.3638ms      5103     463ns     320ns  4.7060us  [CUDA memcpy DtoH]
                    0.00%  754.43us       262  2.8790us  2.1770us  3.6810us  growDetectionsKernel(unsigned int*, unsigned int, NcvRect32u*, unsigned int, unsigned int, float)
      API calls:   78.84%  30.1433s     16978  1.7754ms  7.8450us  155.95ms  cudaStreamSynchronize
                   12.18%  4.65492s         3  1.55164s  65.662us  4.65246s  cudaMallocHost
                    5.15%  1.97074s     11251  175.16us  92.424us  2.8773ms  cudaLaunchKernel
                    0.92%  352.35ms      4995  70.539us  36.341us  424.22us  cudaMemcpyFromSymbolAsync
                    0.69%  263.78ms       111  2.3764ms  1.6450ms  4.3629ms  cudaMemcpy2D
                    0.68%  260.51ms      4995  52.154us  32.909us  312.98us  cudaMemcpyToSymbolAsync
                    0.46%  175.17ms      8547  20.494us  12.676us  267.89us  cudaBindTexture
                    0.43%  164.43ms      1776  92.582us  52.176us  754.31us  cudaMemcpy
                    0.31%  116.97ms         5  23.394ms  51.263us  59.521ms  cudaMalloc
                    0.19%  73.963ms       111  666.34us  482.53us  3.3831ms  cudaDeviceSynchronize
                    0.04%  17.101ms         1  17.101ms  17.101ms  17.101ms  cudaHostAlloc
                    0.04%  14.154ms     11251  1.2570us     769ns  70.962us  cudaGetLastError
                    0.02%  8.8179ms      8547  1.0310us     448ns  67.047us  cudaCreateChannelDesc
                    0.02%  6.0676ms         2  3.0338ms  2.1710ms  3.8966ms  cudaMallocPitch
                    0.01%  3.4590ms       225  15.373us  4.7720us  101.19us  cudaGetDevice
                    0.01%  3.2501ms         6  541.69us  157.23us  2.0153ms  cudaFree
                    0.01%  2.0453ms         4  511.33us  108.74us  1.3174ms  cudaFreeHost
                    0.00%  446.41us         3  148.80us  124.64us  170.97us  cudaGetDeviceProperties
                    0.00%  365.39us        96  3.8060us  2.2400us  87.900us  cuDeviceGetAttribute
                    0.00%  18.399us         1  18.399us  18.399us  18.399us  cuDeviceTotalMem
                    0.00%  17.440us         3  5.8130us  2.6560us  7.4880us  cuDeviceGetCount
                    0.00%  9.4110us         6  1.5680us     800ns  3.7120us  cudaGetDeviceCount
                    0.00%  7.3270us         2  3.6630us  3.4870us  3.8400us  cuDeviceGet
                    0.00%  4.8320us         1  4.8320us  4.8320us  4.8320us  cuDeviceGetName
                    0.00%  2.9110us         1  2.9110us  2.9110us  2.9110us  cuDeviceGetUuid
                    0.00%  1.6960us         1  1.6960us  1.6960us  1.6960us  cudaDriverGetVersion
                    0.00%  1.4400us         1  1.4400us  1.4400us  1.4400us  cudaRuntimeGetVersion