Re: branch bill-bnfinit-parallel

[Bill Allombert <Bill.Allombert@math.u-bordeaux.fr>]

On Sun, Sep 21, 2025 at 11:39:44PM +0200, Bill Allombert wrote:
> Dear PARI developers,
> 
> I have pushed a branch bill-bnfinit-parallel which an attempt to merge the parallel bnfinit code
> to master.

I pushed this branch to master wiht minor changes (now the default is [0,0,0,0,0,0]).

This is the conclusion of more than 10 year of works on bnfinit, with Loïc Grenié and Karim.
Ideally bnfinit should pick up good parameters in more situations, and this tuning should be
local default. But at least now, we can gather more experience without having to rebuild
PARI all the time.

> x^28+1154*x^26+489873*x^24+100211629*x^22+11699125060*x^20+854076746718*x^18+40959569036260*x^16+1321233290968720*x^14+28785594434947700*x^12+417246803388550268*x^10+3867079700626715505*x^8+21181990617517025086*x^6+59006973863083141311*x^4+63878093656658485059*x^2+21292697885552828353

For some reason this field is pathological. The default value of idex does not
work, but otherwise it is easy.

There are different two strategies that works fine:

? bnfinit(P,,[0,0,-1,100,0,1]);
  ***   last result: cpu time 2min, 31,465 ms, real time 57,565 ms.

? bnfinit(P,,[0,0,0,100,6,1]);
  ***   last result: cpu time 3min, 16,368 ms, real time 47,966 ms.

Cheers,
Bill.