Platform: NVIDIA CUDA
  Device: GeForce GTX 465
    Driver version : 325.15 (Linux x86)
    Compute units  : 11

    Global memory bandwidth (GBPS)
      float   : 85.28
      float2  : 86.04
      float4  : 87.34
      float8  : 45.51
      float16 : 22.70

    Single-precision compute (GFLOPS)
      float   : 843.19
      float2  : 835.39
      float4  : 836.46
      float8  : 831.69
      float16 : 827.41

    Double-precision compute (GFLOPS)
      double   : 106.75
      double2  : 106.66
      double4  : 106.41
      double8  : 106.04
      double16 : 105.21

    Integer compute (GIOPS)
      int   : 426.07
      int2  : 425.49
      int4  : 426.23
      int8  : 426.26
      int16 : 426.24

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 0.69
      enqueueReadBuffer          : 0.47
      enqueueMapBuffer(for read) : 0.35
        memcpy from mapped ptr   : 0.53
      enqueueUnmap(after write)  : 1.57
        memcpy to mapped ptr     : 0.49

    Kernel launch latency : 15.87 us