Platform: AMD Accelerated Parallel Processing
  Device: gfx1036
    Driver version  : 3635.0 (HSA1.1,LC) (Linux x64)
    Compute units   : 1
    Clock frequency : 2200 MHz

    Global memory bandwidth (GBPS)
      float   : 37.70
      float2  : 47.62
      float4  : 49.18
      float8  : 43.33
      float16 : 35.86

    Single-precision compute (GFLOPS)
      float   : 512.54
      float2  : 516.92
      float4  : 520.29
      float8  : 509.35
      float16 : 502.36

    Half-precision compute (GFLOPS)
      half   : 510.31
      half2  : 1067.73
      half4  : 1069.36
      half8  : 1037.82
      half16 : 1048.27

    Double-precision compute (GFLOPS)
      double   : 34.19
      double2  : 34.10
      double4  : 34.13
      double8  : 33.89
      double16 : 33.54

    Integer compute (GIOPS)
      int   : 107.46
      int2  : 107.12
      int4  : 108.21
      int8  : 107.37
      int16 : 106.68

    Integer compute Fast 24bit (GIOPS)
      int   : 533.57
      int2  : 536.99
      int4  : 535.24
      int8  : 522.02
      int16 : 478.56

    Integer char (8bit) compute (GIOPS)
      char   : 541.26
      char2  : 288.72
      char4  : 278.60
      char8  : 256.80
      char16 : 254.43

    Integer short (16bit) compute (GIOPS)
      short   : 535.69
      short2  : 535.77
      short4  : 535.26
      short8  : 532.39
      short16 : 509.55

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 17.56
      enqueueReadBuffer               : 4.97
      enqueueWriteBuffer non-blocking : 17.72
      enqueueReadBuffer non-blocking  : 4.96
      enqueueMapBuffer(for read)      : 244032.22
        memcpy from mapped ptr        : 4.87
      enqueueUnmap(after write)       : 370255.78
        memcpy to mapped ptr          : 16.53

    Kernel launch latency : 1545461120.00 us