Re: Pi with MT

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

On Mon, Aug 28, 2023 at 11:51:50PM +0200, Jean-Luc ARNAUD wrote:
> I'm interested in a Pi function using MT.
> 
> Pari/GP Pi is very fast but does not use MT.
> 
> Did anyone code this Pi function in MT and would be so kind as to share
> his/her code?

There is a library for fast parallel multiplication of integers that can be used
to implement a parallel Brent-Salamin AGM method for Pi, but I do not remember
its name.

Maybe we should consider implementing parallel multiplication for very very
large integers, using a two-stage NTT process.

Cheers,
Bill.