Let a, b be two integers with n digits. I am wondering does the computing time of the square of a is shorter than a*b.
Thank you for your help.
Let a, b be two integers with n digits. I am wondering does the computing time of the square of a is shorter than a*b.
Thank you for your help.
I don't think there's a way to square A without using an IMUL on x86. I could be wrong.
To find out how long something takes, microbenchmark it!
Edit: oh wait, I've got it! a*b takes two memory reads and a*a takes one! So a*a is faster :-).
True answer: there's no reason a*b would be slower unless you have some outside factor influencing things.