Platform: NVIDIA CUDA
  Device: GeForce GT 650M
    Driver version: 331.82 (Win32)

    Global memory bandwidth (GBPS)
      float   : 25.16
      float2  : 26.84
      float4  : 26.84
      float8  : 13.42
      float16 : 12.39

    Single-precision compute (GFLOPS)
      float   : 486.09
      float2  : 644.06
      float4  : 613.40
      float8  : 613.40
      float16 : 613.40

    Double-precision compute (GFLOPS)
      double   : 30.38
      double2  : 30.34
      double4  : 30.27
      double8  : 30.13
      double16 : 29.92

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 4.36
      enqueueReadBuffer          : 4.29
      enqueueMapBuffer(for read) : 2.95
        memcpy from mapped ptr   : 8.01
      enqueueUnmap(after write)  : 4.33
        memcpy to mapped ptr     : 8.26

    Kernel launch latency : 25.48 us