Platform: NVIDIA CUDA
  Device: GeForce RTX 2080
    Driver version  : 410.73 (Linux x64)
    Compute units   : 46
    Clock frequency : 1815 MHz

    Global memory bandwidth (GBPS)
      float   : 362.93
      float2  : 382.42
      float4  : 391.26
      float8  : 400.79
      float16 : 364.98

    Single-precision compute (GFLOPS)
      float   : 11258.41
      float2  : 11248.28
      float4  : 11228.37
      float8  : 11166.76
      float16 : 11064.75

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 354.32
      double2  : 353.24
      double4  : 351.23
      double8  : 349.27
      double16 : 346.67

    Integer compute (GIOPS)
      int   : 11085.63
      int2  : 11005.45
      int4  : 11002.92
      int8  : 10991.37
      int16 : 10955.21

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 5.88
      enqueueReadBuffer          : 6.49
      enqueueMapBuffer(for read) : 5.96
        memcpy from mapped ptr   : 14.68
      enqueueUnmap(after write)  : 6.18
        memcpy to mapped ptr     : 14.82

    Kernel launch latency : 3.85 us