./clpeak

Platform: AMD Accelerated Parallel Processing
  Device: gfx1150
    Driver version  : 3635.0 (HSA1.1,LC) (Linux x64)
    Compute units   : 8
    Clock frequency : 2900 MHz

    Global memory bandwidth (GBPS)
      float   : 89.48
      float2  : 95.48
      float4  : 96.20
      float8  : 93.38
      float16 : 96.39

    Single-precision compute (GFLOPS)
      float   : 4199.26
      float2  : 4300.34
      float4  : 4285.72
      float8  : 4205.67
      float16 : 4236.92

    Half-precision compute (GFLOPS)
      half   : 4290.57
      half2  : 7295.99
      half4  : 7111.85
      half8  : 7103.91
      half16 : 6717.18

    Double-precision compute (GFLOPS)
      double   : 181.86
      double2  : 181.30
      double4  : 180.56
      double8  : 180.13
      double16 : 178.96

    Integer compute (GIOPS)
      int   : 1074.42
      int2  : 1008.43
      int4  : 1031.80
      int8  : 904.98
      int16 : 1079.03

    Integer compute Fast 24bit (GIOPS)
      int   : 3266.01
      int2  : 3493.31
      int4  : 3482.08
      int8  : 3331.33
      int16 : 3383.28

    Integer char (8bit) compute (GIOPS)
      char   : 3244.71
      char2  : 2195.23
      char4  : 2145.75
      char8  : 1948.54
      char16 : 1953.09

    Integer short (16bit) compute (GIOPS)
      short   : 3371.77
      short2  : 3419.20
      short4  : 3380.22
      short8  : 3337.97
      short16 : 3374.91

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 26.51
      enqueueReadBuffer               : 26.45
      enqueueWriteBuffer non-blocking : 26.52
      enqueueReadBuffer non-blocking  : 26.48
      enqueueMapBuffer(for read)      : 380085.59
        memcpy from mapped ptr        : 26.44
      enqueueUnmap(after write)       : 681740.81
        memcpy to mapped ptr          : 26.54

    Kernel launch latency : 5.83 us