Platform: NVIDIA CUDA
  Device: GeForce RTX 2060
    Driver version  : 440.33.01 (Linux x64)
    Compute units   : 30
    Clock frequency : 1200 MHz

    Global memory bandwidth (GBPS)
      float   : 277.69
      float2  : 292.44
      float4  : 299.43
      float8  : 307.26
      float16 : 295.93

    Single-precision compute (GFLOPS)
      float   : 6660.58
      float2  : 6652.10
      float4  : 6616.57
      float8  : 6490.34
      float16 : 6415.59

    No half precision support! Skipped

    Double-precision compute (GFLOPS)
      double   : 225.49
      double2  : 222.73
      double4  : 223.70
      double8  : 222.83
      double16 : 220.67

    Integer compute (GIOPS)
      int   : 5287.57
      int2  : 5311.86
      int4  : 5324.80
      int8  : 5347.13
      int16 : 5331.60

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 6.02
      enqueueReadBuffer          : 6.20
      enqueueMapBuffer(for read) : 5.98
        memcpy from mapped ptr   : 13.10
      enqueueUnmap(after write)  : 6.15
        memcpy to mapped ptr     : 13.22

    Kernel launch latency : 3.44 us