Platform: Portable Computing Language
  Device: Orin
    Driver version  : 5.0 (Linux ARM64)
    Compute units   : 16
    Clock frequency : 1300 MHz

    Global memory bandwidth (GBPS)
      float   : 174.18
      float2  : 178.27
      float4  : 178.72
      float8  : 179.86
      float16 : 114.04

    Single-precision compute (GFLOPS)
      float   : 5173.41
      float2  : 5141.03
      float4  : 5052.36
      float8  : 5210.57
      float16 : 5170.51

    Half-precision compute (GFLOPS)
      half   : 2638.87
      half2  : 9957.23
      half4  : 7488.28
      half8  : 8151.71
      half16 : 7179.72

    Double-precision compute (GFLOPS)
      double   : 83.12
      double2  : 83.04
      double4  : 82.84
      double8  : 82.44
      double16 : 81.60

    Integer compute (GIOPS)
      int   : 1767.72
      int2  : 1777.34
      int4  : 1816.11
      int8  : 1772.77
      int16 : 1781.42

    Integer compute Fast 24bit (GIOPS)
      int   : 1767.65
      int2  : 1777.16
      int4  : 1816.12
      int8  : 1772.98
      int16 : 1781.72

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 8.52
      enqueueReadBuffer               : 7.30
      enqueueWriteBuffer non-blocking : 7.30
      enqueueReadBuffer non-blocking  : 7.29
      enqueueMapBuffer(for read)      : 129757.33
        memcpy from mapped ptr        : 7.32
      enqueueUnmap(after write)       : 8.98
        memcpy to mapped ptr          : 7.32

    Kernel launch latency : -410.81 us