Re: found my problem, subtle error occurring with insert into a set

[American Citizen <website.reader3@gmail.com>]

For a rank=3 curve, the ellpool program is running over 61x faster. I am 
very pleased with this speedup


On 3/4/24 17:44, American Citizen wrote:
> To all
>
> More hours of painful troubleshooting step by step finally revealed 
> that I was making a subtle error, doing a set insert into a current 
> set, while operating ON that set.
>
> I now get the 67 points for the congruent number curve of 
> [0,0,0,-1156,0] of height <= 13.
>
> qfminim() command is correct, it is giving me 8 vectors, so I must 
> take the negative point also, giving 16 overall for the points of 
> infinite order. When I use the 3 torsion points for this curve, I 
> eventually come up to 67 points = 16*4 + 3 (torsion points).
>
> Thank you for bearing with me on this troubleshooting, as my 
> reduce_to_basis algorithm showed that combinations NOT on the qfminin 
> vector list were occuring, but this was due to the negative point 
> (i.e. ellneg(e,pt))
>
> I am happy now with the new rewrite of my ellpool(curve,basis_pts, 
> max_height) as it is running 8x as fast as my previous program for a 
> rank 2 curve.
>
> Randall