Platform: Intel(R) OpenCL Graphics
  Device: Intel(R) Arc(TM) A770 Graphics
    Driver version  : 24.52.032224 (Linux x64)
    Compute units   : 512
    Clock frequency : 2400 MHz

    Global memory bandwidth (GBPS)
      float   : 398.72
      float2  : 403.84
      float4  : 406.51
      float8  : 411.07
      float16 : 417.07

    Single-precision compute (GFLOPS)
      float   : 13018.30
      float2  : 11116.62
      float4  : 10404.04
      float8  : 10026.77
      float16 : 9709.70

    Half-precision compute (GFLOPS)
      half   : 19549.44
      half2  : 19485.39
      half4  : 19516.85
      half8  : 19454.13
      half16 : 19336.65

    No double precision support! Skipped

    Integer compute (GIOPS)
      int   : 5494.20
      int2  : 5485.00
      int4  : 5481.39
      int8  : 5455.18
      int16 : 5424.15

    Integer compute Fast 24bit (GIOPS)
      int   : 5495.92
      int2  : 5490.93
      int4  : 5477.65
      int8  : 5448.69
      int16 : 5423.26

    Integer char (8bit) compute (GIOPS)
      char   : 11373.47
      char2  : 11309.92
      char4  : 11281.28
      char8  : 11044.53
      char16 : 10582.53

    Integer short (16bit) compute (GIOPS)
      short   : 18022.99
      short2  : 17908.35
      short4  : 17980.90
      short8  : 17852.69
      short16 : 17671.18

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 5.41
      enqueueReadBuffer               : 5.26
      enqueueWriteBuffer non-blocking : 5.81
      enqueueReadBuffer non-blocking  : 5.64
      enqueueMapBuffer(for read)      : 5.43
        memcpy from mapped ptr        : 21.11
      enqueueUnmap(after write)       : 5.83
        memcpy to mapped ptr          : 21.43

    Kernel launch latency : 87.89 us