You must log in or register to comment.
I mean if the CPU that’s running these instructions is super low power then 0.02 might be worth it
Or if you’re scaling a large cluster of CPUs for parallel computations where a 0.02% increase can make a tangible difference in runtimes.
Only if you’d removed and fixed all other bottlenecks that would gain you more than 0.02%. And I’m not convinced there are many if any projects of any reasonable size where that has been the case.
