Platform: Intel(R) OpenCL
  Device: Intel(R) Many Integrated Core Acceleration Card
    Driver version  : 1.2 (Linux x64)
    Compute units   : 240
    Clock frequency : 1333 MHz

    Global memory bandwidth (GBPS)
      float   : 68.00
      float2  : 49.67
      float4  : 89.31
      float8  : 101.30
      float16 : 1.84

    Single-precision compute (GFLOPS)
      float   : 2130.58
      float2  : 2262.12
      float4  : 2255.07
      float8  : 2247.39
      float16 : 2213.89

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 1157.62
      double2  : 1156.44
      double4  : 1153.55
      double8  : 1146.15
      double16 : 354.26

    Integer compute (GIOPS)
      int   : 1158.72
      int2  : 1161.01
      int4  : 1158.69
      int8  : 1158.58
      int16 : 1147.02

    Integer compute Fast 24bit (GIOPS)
      int   : 1158.82
      int2  : 1160.21
      int4  : 1158.02
      int8  : 1150.32
      int16 : 1136.05

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 2.25
      enqueueReadBuffer               : 5.62
      enqueueWriteBuffer non-blocking : 4.92
      enqueueReadBuffer non-blocking  : 5.57
      enqueueMapBuffer(for read)      : 137.80
        memcpy from mapped ptr        : 3.71
      enqueueUnmap(after write)       : 6.55
        memcpy to mapped ptr          : 3.43

    Kernel launch latency : 79.60 us


Platform: NVIDIA CUDA

Platform: AMD Accelerated Parallel Processing