Platform: Intel(R) OpenCL Graphics
  Device: Intel(R) Arc(TM) Graphics
    Driver version  : 24.35.030872 (Linux x64)
    Compute units   : 128
    Clock frequency : 2350 MHz

    Global memory bandwidth (GBPS)
      float   : 72.71
      float2  : 73.58
      float4  : 75.40
      float8  : 78.40
      float16 : 79.53

    Single-precision compute (GFLOPS)
      float   : 4774.18
      float2  : 4753.14
      float4  : 4757.33
      float8  : 4734.26
      float16 : 4490.65

    Half-precision compute (GFLOPS)
      half   : 9473.49
      half2  : 9375.61
      half4  : 9460.43
      half8  : 9379.66
      half16 : 9290.04

    Double-precision compute (GFLOPS)
      double   : 149.58
      double2  : 147.29
      double4  : 148.95
      double8  : 147.74
      double16 : 145.22

    Integer compute (GIOPS)
      int   : 1259.85
      int2  : 1216.70
      int4  : 1210.95
      int8  : 1206.91
      int16 : 1204.12

    Integer compute Fast 24bit (GIOPS)
      int   : 1211.44
      int2  : 1206.41
      int4  : 1208.73
      int8  : 1214.26
      int16 : 1196.85

    Integer char (8bit) compute (GIOPS)
      char   : 2895.98
      char2  : 2909.44
      char4  : 2912.20
      char8  : 2845.54
      char16 : 2735.94

    Integer short (16bit) compute (GIOPS)
      short   : 7731.06
      short2  : 7241.28
      short4  : 7279.79
      short8  : 7210.25
      short16 : 7299.28

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 12.24
      enqueueReadBuffer               : 12.29
      enqueueWriteBuffer non-blocking : 30.95
      enqueueReadBuffer non-blocking  : 30.22
      enqueueMapBuffer(for read)      : 25.39
        memcpy from mapped ptr        : 21.12
      enqueueUnmap(after write)       : 33.72
        memcpy to mapped ptr          : 21.11

    Kernel launch latency : 37.37 us