Platform: AMD Accelerated Parallel Processing
  Device: Cypress
    Driver version : 1348.4 (Linux x64)
    Compute units  : 20

    Global memory bandwidth (GBPS)
      float   : 127.53
      float2  : 125.89
      float4  : 94.18
      float8  : 64.77
      float16 : 34.23

    Single-precision compute (GFLOPS)
      float   : 542.26
      float2  : 1077.85
      float4  : 2139.41
      float8  : 2130.85
      float16 : 2133.98

    Double-precision compute (GFLOPS)
      double   : 540.76
      double2  : 539.13
      double4  : 535.31
      double8  : 537.66
      double16 : 534.95

    Integer compute (GIOPS)
      int   : 270.28
      int2  : 540.54
      int4  : 540.12
      int8  : 540.88
      int16 : 541.23

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 3.61
      enqueueReadBuffer          : 3.36
      enqueueMapBuffer(for read) : 206.81
        memcpy from mapped ptr   : 3.41
      enqueueUnmap(after write)  : 584.32
        memcpy to mapped ptr     : 3.32

    Kernel launch latency : 117.88 us