This week I added a lot of comments and some documentation. Next I processed all compiler warnings and fix couple of bugs (in tests).
I created a new branch which contains all "clear" code (I remove Montgomery form and some extra files), so that we can review it and if everything is okay merge it..
![]() |
timings from profiling code |
I add profiling code with five primes of each size = 50, 60, ..., 300. I took a numbers from List of small primes. So anyone can now easily look at timings on his computer. I also run test for 2^2000+841 (603 digits) and 2^3515+159 (1059 digits) it took 60 and 700 seconds respectively.
In PARI/GP implementation for Jacobi sum computation them use q / 2 size table and I use two q size tables (to compute f(x) such that 1 - g^x = f(x), g - modulo q primitive root). Now I'm working on it. It will save a few megabytes of ram for large numbers and maybe give small speed up.
No comments:
Post a Comment