Hi!
I missed out last week's blog, couldn't write it as I was traveling and had poor internet connection. In the last two weeks, I have been trying out different implementations of stage II, using multi point evaluation.
Mid term evaluations began on the 26th of June, and I have met most of the targets I had set. I have added an optimized version of pollard rho, and the ECM code should be completed by the end of this week. Although there are still a lot of things I'd like to try to make it faster, I don't want to deviate a lot from the timeline now. I'll probably get back to it after GSoC.
The current ECM timings are close to GMP-ECM, I'll be posting some timings next week, once I code it using mpn's. My main worry right now is the fact that using multi point evaluation is slowing down ECM, contrary to the expected speedup.
This week, I plan to find the reason why the multi point evaluation technique is not turning out to be faster, as expected. Also I would be rewriting ECM, at least the stage I, using mpn's.