Re: Abnormous memory use for gaussian gcd()?

[hermann@stamm-wilbrandt.de]

> > 1)
> > Running "gp < 36401.gp" does have >3G resident ram during execution, 
> > that is
> > abnormous memory.
>
> Indeed, I fixed that in the master branch.
Thank you, I did build from last stable on i7-11850H.
How can I build from master branch?

> You should be able to replace sq2(p) by qfbcornacchia(1, p)
I did that:

$ tail -5 3.gp
print("foobar");

[x,y] = qfbcornacchia(1, p)
##
assert(x^2 + y^2 == p, "x,y", "not sum of squares");
$

And it completed correctly:

$ gp < 3.gp

foobar
%20 =
...
508147421190861611913185178774089527222223208482508012472161373459203[+++]
  ***   last result computed in 1min, 17,373 ms.
Goodbye!
stammw:pari$

But again taking more than 1 minute.
And in between "top" showed 4.8GB resident memory for "gp" process.


Regards,

Hermann Stamm-Wilbrandt.