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   : 400.33
      float2  : 405.39
      float4  : 406.40
      float8  : 413.38
      float16 : 419.15

    Single-precision compute (GFLOPS)
      float   : 13008.71
      float2  : 11142.17
      float4  : 10408.81
      float8  : 10030.98
      float16 : 9713.66

    Half-precision compute (GFLOPS)
      half   : 19566.62
      half2  : 19503.51
      half4  : 19533.32
      half8  : 19470.99
      half16 : 19356.95

    No double precision support! Skipped

    Integer compute (GIOPS)
      int   : 5504.73
      int2  : 5491.85
      int4  : 5474.75
      int8  : 5440.34
      int16 : 5408.72

    Integer compute Fast 24bit (GIOPS)
      int   : 5477.15
      int2  : 5470.67
      int4  : 5456.10
      int8  : 5422.03
      int16 : 5397.73

    Integer char (8bit) compute (GIOPS)
      char   : 11393.08
      char2  : 11335.06
      char4  : 11307.85
      char8  : 11060.50
      char16 : 10578.37

    Integer short (16bit) compute (GIOPS)
      short   : 18225.60
      short2  : 18155.14
      short4  : 18200.54
      short8  : 18091.05
      short16 : 17910.72

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 5.43
      enqueueReadBuffer               : 5.25
      enqueueWriteBuffer non-blocking : 5.82
      enqueueReadBuffer non-blocking  : 5.62
      enqueueMapBuffer(for read)      : 5.43
        memcpy from mapped ptr        : 20.67
      enqueueUnmap(after write)       : 5.84
        memcpy to mapped ptr          : 20.65

    Kernel launch latency : 4.40 us