Platform: Portable Computing Language
  Device: pthread-POWER9, altivec supported
    Driver version  : 3.0-rc2 (Linux unknown)
    Compute units   : 160
    Clock frequency : 3800 MHz

    Global memory bandwidth (GBPS)
      float   : 30.30
      float2  : 59.39
      float4  : 63.92
      float8  : 60.26
      float16 : 57.05

    Single-precision compute (GFLOPS)
      float   : 73.11
      float2  : 179.68
      float4  : 411.74
      float8  : 739.41
      float16 : 910.81

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 85.08
      double2  : 151.08
      double4  : 275.05
      double8  : 401.79
      double16 : 456.30

    Integer compute (GIOPS)
      int   : 112.89
      int2  : 189.39
      int4  : 440.41
      int8  : 708.03
      int16 : 748.61

    Integer compute Fast 24bit (GIOPS)
      int   : 149.56
      int2  : 226.40
      int4  : 407.09
      int8  : 721.65
      int16 : 755.17

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 5.88
      enqueueReadBuffer               : 5.37
      enqueueWriteBuffer non-blocking : 5.52
      enqueueReadBuffer non-blocking  : 5.24
      enqueueMapBuffer(for read)      : 901.70
        memcpy from mapped ptr        : 7.58
      enqueueUnmap(after write)       : 734.74
        memcpy to mapped ptr          : 11.31

    Kernel launch latency : 76.72 us