clpeak benchmark using intel uhd 630 gpu

clpeak.log

spervaiz@roche ~/clpeak/build (git)-[master] % ./clpeak

Platform: Intel(R) OpenCL HD Graphics
  Device: Intel(R) Graphics [0x9bc8]
    Driver version  : 21.04.18912 (Linux x64)
    Compute units   : 23
    Clock frequency : 1100 MHz

    Global memory bandwidth (GBPS)
      float   : 13.65
      float2  : 14.82
      float4  : 15.80
      float8  : 16.35
      float16 : 16.38

    Single-precision compute (GFLOPS)
      float   : 399.31
      float2  : 394.52
      float4  : 397.46
      float8  : 395.00
      float16 : 388.30

    Half-precision compute (GFLOPS)
      half   : 784.32
      half2  : 778.45
      half4  : 784.29
      half8  : 779.06
      half16 : 769.07

    Double-precision compute (GFLOPS)
      double   : 100.31
      double2  : 99.12
      double4  : 99.85
      double8  : 99.11
      double16 : 97.30

    Integer compute (GIOPS)
      int   : 133.33
      int2  : 132.89
      int4  : 133.24
      int8  : 132.70
      int16 : 131.22

    Integer compute Fast 24bit (GIOPS)
      int   : 133.33
      int2  : 132.91
      int4  : 133.23
      int8  : 132.65
      int16 : 131.10

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 6.78
      enqueueReadBuffer               : 7.01
      enqueueWriteBuffer non-blocking : 6.61
      enqueueReadBuffer non-blocking  : 6.63
      enqueueMapBuffer(for read)      : 245061.25
        memcpy from mapped ptr        : 7.02
      enqueueUnmap(after write)       : 32838206.00
        memcpy to mapped ptr          : 6.78

    Kernel launch latency : 32.87 us

./clpeak  55.27s user 37.86s system 45% cpu 3:26.01 total