Max Alekseyev on Sat, 07 Aug 2010 23:54:00 +0200


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: nfinit() is eager for memory


On Sat, Aug 7, 2010 at 12:59 PM, Bill Allombert
<Bill.Allombert@math.u-bordeaux1.fr> wrote:
> On Sat, Aug 07, 2010 at 10:19:32AM -0400, Max Alekseyev wrote:
>> nfinit(polcyclo(49321,y)) wants more than 12GB stack.
>
> Does it actually complete ?

Not really.
12GB was the maximum stack size I've tried.

>> Is that expected? Does nfinit() indeed needs that much memory?
>
> Independently of the algorithm used, nfinit output (amon other things) the
> multiplication table which is a NxNxN tensor where N is the degree of the polynomial.
> Here N = 45360, so it has 93*10^12 entries. This will not fit in memory.

Is there a way to overcome this?
My goal is to compute
nffactor(x^3+x^2-16440*x+80375,nfinit(polcyclo(49321,y)))
that is, to express roots of x^3+x^2-16440*x+80375 in terms of
trigonometric functions (by later substituting y=exp(2*Pi*I/49321)).

Thanks,
Max