Platform: Apple
  Device: Apple M1 Ultra
    Driver version  : 1.2 1.0 (Macintosh)
    Compute units   : 48
    Clock frequency : 1000 MHz

    Global memory bandwidth (GBPS)
      float   : 699.31
      float2  : 716.93
      float4  : 729.42
      float8  : 703.35
      float16 : 375.06

    Single-precision compute (GFLOPS)
      float   : 7205.15
      float2  : 7340.24
      float4  : 7365.57
      float8  : 6082.06
      float16 : 7706.36

    No half precision support! Skipped

    No double precision support! Skipped

    Integer compute (GIOPS)
      int   : 3955.97
      int2  : 3956.72
      int4  : 3954.73
      int8  : 3937.65
      int16 : 3958.39

    Integer compute Fast 24bit (GIOPS)
      int   : 3957.59
      int2  : 3956.45
      int4  : 3953.95
      int8  : 3937.62
      int16 : 3954.19

    Transfer bandwidth (GBPS)
      enqueueWriteBuffer              : 34.98
      enqueueReadBuffer               : 34.05
      enqueueWriteBuffer non-blocking : 45.08
      enqueueReadBuffer non-blocking  : 41.00
      enqueueMapBuffer(for read)      : 795364.31
        memcpy from mapped ptr        : 33.96
      enqueueUnmap(after write)       : 346368.34
        memcpy to mapped ptr          : 34.27

    Kernel launch latency : 2.14 us