There is still a few example doing manual profiling. Can we fix those? Just search for `num_iterations = 126 # Match Iris benchmark iterations`

[mawad-amd]

There is still a few example doing manual profiling. Please use iris.do_bench. Can we fix those? Just search for num_iterations = 126 # Match Iris benchmark iterations