GSoC Pencils Down

The pencils down deadline for GSoC was 21st Aug, the previous Friday. It has been quite a journey, working with FLINT for over three months. I've learnt quite a lot, not just math but how to write good quality code and testing.

Although I haven't been able to do everything I had mentioned in my proposal, I worked on a lot of stuff which I didn't mention. I worked extensively on the ECM, more than I had planned and ended up writing a very good stage I and II using mpn's. Apart from that I worked on the Pollard Rho Brent algorithm, and the MPQS. The MPQS has some problems and is not functional though.

This week I tried to debug the MPQS. The problem with MPQS is that the sieve value is not exceeding a limit, and hence does not get evaluated. No relations are being found. I believe this is because the roots of the polynomials are being computed incorrectly. I will be looking into this eventually, once I get a break in the semester.

Apart from the MPQS, I also worked on the fmpz_factor_smooth() function. It aims to find a factor of a number, given that it is not semi prime (around n^1/3). It would be a combination of trial division, Pollard Rho Brent, p + 1 and ECM. The profiling of Pollard Rho Brent is yet to be done. Once I do that, I can finish this off.

What's next? I have a lot of stuff which I plan to do. First, is obviously to finish off MPQS and adding a double large prime version. Some other things which I want to work on is to improve the ECM stage II, and the fmpz_factor_smooth() function. I will be working on them, however at a slower speed due to the course load.

Show Comments