Platform: Portable Computing Language
  Device: NVIDIA Tegra X1
    Driver version  : 3.0-rc2 (Linux ARM64)
    Compute units   : 1
    Clock frequency : 921 MHz

    Global memory bandwidth (GBPS)
      float   : 17.84
      float2  : 20.57
      float4  : 21.38
      float8  : 19.87
      float16 : 17.72

    Single-precision compute (GFLOPS)
      float   : 220.80
      float2  : 228.89
      float4  : 230.41
      float8  : 229.79
      float16 : 229.13

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 7.32
      double2  : 7.32
      double4  : 7.30
      double8  : 7.28
      double16 : 7.22

    Integer compute (GIOPS)
      int   : 76.86
      int2  : 77.52
      int4  : 77.30
      int8  : 76.07
      int16 : 74.52

    Integer compute Fast 24bit (GIOPS)
      int   : 76.87
      int2  : 77.51
      int4  : 77.29
      int8  : 76.04
      int16 : 74.89

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 2.35
      enqueueReadBuffer               : 2.36
      enqueueWriteBuffer non-blocking : 2.37
      enqueueReadBuffer non-blocking  : 2.27
      enqueueMapBuffer(for read)      : 7635.37
        memcpy from mapped ptr        : 0.71
      enqueueUnmap(after write)       : 9.07
        memcpy to mapped ptr          : 3.56

    Kernel launch latency : -14.24 us