Platform: AMD Accelerated Parallel Processing
  Device: gfx1012:xnack- (RX 5500XT)
    Driver version  : 3361.0 (HSA1.1,LC) (Linux x64)
    Compute units   : 11
    Clock frequency : 1900 MHz

    Global memory bandwidth (GBPS)
      float   : 190.36
      float2  : 182.03
      float4  : 171.64
      float8  : 157.88
      float16 : 154.64

    Single-precision compute (GFLOPS)
      float   : 5046.67
      float2  : 4936.51
      float4  : 4887.78
      float8  : 4871.37
      float16 : 4796.57

    Half-precision compute (GFLOPS)
      half   : 2544.83
      half2  : 9875.69
      half4  : 9771.45
      half8  : 9731.20
      half16 : 9533.27

    Double-precision compute (GFLOPS)
      double   : 323.84
      double2  : 323.33
      double4  : 322.61
      double8  : 321.09
      double16 : 318.06

    Integer compute (GIOPS)
      int   : 1025.25
      int2  : 1025.22
      int4  : 1021.86
      int8  : 1018.56
      int16 : 1012.24

    Integer compute Fast 24bit (GIOPS)
      int   : 4738.49
      int2  : 4805.52
      int4  : 4799.40
      int8  : 4682.88
      int16 : 4766.09

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 15.64
      enqueueReadBuffer               : 15.34
      enqueueWriteBuffer non-blocking : 15.70
      enqueueReadBuffer non-blocking  : 15.44
      enqueueMapBuffer(for read)      : 613566.81
        memcpy from mapped ptr        : 15.37
      enqueueUnmap(after write)       : 1227133.62
        memcpy to mapped ptr          : 15.70

    Kernel launch latency : 12.77 us