==9237== Profiling result:
            Type  Time(%)      Time     Calls       Avg       Min       Max  Name
 GPU activities:   46.58%  5.50405s      1404  3.9203ms  1.7501ms  31.241ms  void cv::cuda::device::hog::classify_hists_kernel_many_blocks<int=256, int=1>(int, int, int, int, float const *, float const , float, float, unsigned char*)
                   31.72%  3.74813s      1404  2.6696ms  1.2654ms  30.930ms  void cv::cuda::device::hog::compute_hists_kernel_many_blocks<int=4>(int, cv::cuda::PtrStep<float>, cv::cuda::device::hog::compute_hists_kernel_many_blocks<int=4, unsigned char>, float, float*, int, int, int, int, int, int)
                   11.42%  1.34879s      1404  960.68us  463.04us  25.903ms  void cv::cuda::device::hog::compute_gradients_8UC4_kernel<int=256, int=1>(int, int, cv::cuda::PtrStep<unsigned char>, float, cv::cuda::device::hog::compute_gradients_8UC4_kernel<int=256, int=1, float>, cv::cuda::PtrStep)
                    5.36%  632.82ms      1404  450.73us  216.64us  4.8603ms  void cv::cuda::device::hog::normalize_hists_kernel_many_blocks<int=64, int=1>(int, int, float*, float)
                    2.74%  323.22ms      1296  249.40us  161.57us  1.0611ms  cv::cuda::device::hog::resize_for_hog_kernel(float, float, cv::cuda::PtrStepSz<uchar4>, int)
                    2.17%  255.98ms     15553  16.458us     320ns  4.3931ms  [CUDA memcpy HtoD]
                    0.02%  2.7399ms      1404  1.9510us     800ns  9.2480us  [CUDA memcpy DtoH]
      API calls:   45.43%  8.33981s      2700  3.0888ms  104.61us  31.576ms  cudaDeviceSynchronize
                   23.31%  4.27799s         2  2.13899s  2.1431ms  4.27584s  cudaMalloc
                   18.61%  3.41536s       542  6.3014ms  335.81us  65.724ms  cudaFree
                    3.70%  679.45ms     15444  43.994us  30.624us  2.5147ms  cudaMemcpyToSymbolAsync
                    3.40%  623.60ms      6912  90.219us  48.064us  2.3099ms  cudaLaunchKernel
                    3.31%  607.14ms      1513  401.28us  68.160us  6.4725ms  cudaMemcpy2D
                    1.73%  317.03ms       541  586.00us  328.54us  1.2748ms  cudaMallocPitch
                    0.19%  34.909ms      1296  26.935us  15.840us  297.09us  cudaBindTexture2D
                    0.10%  18.182ms      1404  12.949us  6.7200us  86.079us  cudaFuncSetCacheConfig
                    0.08%  14.688ms      2918  5.0330us  3.0080us  118.98us  cudaGetDevice
                    0.08%  14.608ms      1296  11.271us  7.5200us  74.688us  cudaUnbindTexture
                    0.06%  10.525ms      6912  1.5220us     736ns  1.1695ms  cudaGetLastError
                    0.01%  1.6014ms      1296  1.2350us     736ns  16.800us  cudaCreateChannelDesc
                    0.00%  431.20us        96  4.4910us  1.8560us  114.88us  cuDeviceGetAttribute
                    0.00%  315.71us         2  157.85us  125.95us  189.76us  cudaGetDeviceProperties
                    0.00%  20.927us         1  20.927us  20.927us  20.927us  cuDeviceTotalMem
                    0.00%  17.888us         3  5.9620us  2.5920us  8.3200us  cuDeviceGetCount
                    0.00%  6.7840us         2  3.3920us  2.8800us  3.9040us  cuDeviceGet
                    0.00%  5.2470us         5  1.0490us     704ns  1.8880us  cudaGetDeviceCount
                    0.00%  4.6080us         1  4.6080us  4.6080us  4.6080us  cuDeviceGetName
                    0.00%  2.4320us         1  2.4320us  2.4320us  2.4320us  cuDeviceGetUuid
                    0.00%  1.6000us         1  1.6000us  1.6000us  1.6000us  cudaDriverGetVersion
                    0.00%  1.4080us         1  1.4080us  1.4080us  1.4080us  cudaRuntimeGetVersion