Platform: Portable Computing Language
  Device: Xavier
    Driver version  : 3.0-rc2 (Linux ARM64)
    Compute units   : 6
    Clock frequency : 1109 MHz

    Global memory bandwidth (GBPS)
      float   : 47.13
      float2  : 48.76
      float4  : 49.66
      float8  : 44.09
      float16 : 48.35

    Single-precision compute (GFLOPS)
      float   : 844.51
      float2  : 846.66
      float4  : 844.92
      float8  : 842.05
      float16 : 835.65

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 26.57
      double2  : 26.53
      double4  : 26.46
      double8  : 26.27
      double16 : 25.99

    Integer compute (GIOPS)
      int   : 841.91
      int2  : 844.46
      int4  : 840.22
      int8  : 843.51
      int16 : 843.97

    Integer compute Fast 24bit (GIOPS)
      int   : 841.91
      int2  : 844.08
      int4  : 839.84
      int8  : 843.38
      int16 : 842.70

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 6.64
      enqueueReadBuffer               : 6.68
      enqueueWriteBuffer non-blocking : 6.67
      enqueueReadBuffer non-blocking  : 6.68
      enqueueMapBuffer(for read)      : 21204.90
        memcpy from mapped ptr        : 6.71
      enqueueUnmap(after write)       : 10.64
        memcpy to mapped ptr          : 6.69

    Kernel launch latency : -7.18 us