Platform: ARM Platform
  Device: Mali-T628
    Driver version  : 1.2 (Linux ARM)
    Compute units   : 4
    Clock frequency : 600 MHz

    Global memory bandwidth (GBPS)
      float   : 4.14
      float2  : 5.19
      float4  : 6.85
      float8  : 6.25
      float16 : 5.29

    Single-precision compute (GFLOPS)
      float   : 17.20
      float2  : 5.94
      float4  : 5.83
      float8  : 33.96
      float16 : 7.20

    Double-precision compute (GFLOPS)
      double   : 9.01
      double2  : 1.74
      double4  : 16.80
      double8  : 16.84
      double16 : 16.75

    Integer compute (GIOPS)
      int   : 5.29
      int2  : 6.07
      int4  : 6.18
      int8  : 7.38
      int16 : 34.14

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer         : 4.70
      enqueueReadBuffer          : 2.87
      enqueueMapBuffer(for read) : 441.72
        memcpy from mapped ptr   : 2.21
      enqueueUnmap(after write)  : 634.30
        memcpy to mapped ptr     : 2.31

    Kernel launch latency : 233.30 us