NZEC in TMUL Python3.1

callmeLG 2011-12-28 12:52:38 UTC #1

Hello,
i want ask why nobody make TMUL in Python 3. There are few algorithms for Python 2.5.
I tried it with Karatsuba algorithm but i get NZEC all time even if it is working on my local computer.

hendrik 2011-12-28 17:26:14 UTC #2

If you believe you have a solution that does the job try Py-2.5. If you then still believe that can be solved with 3.1 just try. For the majority of the problems 2.5 is faster than 3.1.
Also performance analysis of your code on your home computer is required and recommended.

numerix 2011-12-28 20:38:19 UTC #3

Now there is a Python 3.1 solution ...

callmeLG 2011-12-29 17:11:43 UTC #4

Thanks you both for answers. I already made AC. So numerix lets do a MUL problem

numerix 2011-12-29 18:06:09 UTC #5

Just to get it right: You implemented your own Karatsuba multiplication for the AC code?

According to the best C/C++ runtimes it should be possible at least if you use psyco.
The approach I used for my TMUL solutions won't suffice.

callmeLG 2011-12-29 21:36:49 UTC #6

No. I am not sure why, but my Karatsuba multiplication is slower than python multiplication.

numerix 2011-12-30 05:54:35 UTC #7

Perhaps you didn't know that: Python's builtin multiplication uses Karatsuba algorithm ...