American Citizen on Thu, 03 Jul 2025 05:14:14 +0200


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

Re: question on execution time for qfbsolve


I ran over the first 100,000 integers

? for(i=1,100000,qfbsolve(Qfb(1,0,1),i,3))
cpu time = 3,394 ms, real time = 3,394 ms.
? for(i=1,100000,qfbsolve(Qfb(1,0,1),[i,factor(i)],3))
cpu time = 3,348 ms, real time = 3,348 ms.

so it is about the same, not much improvement, 46 milliseconds.

On 7/2/25 20:11, American Citizen wrote:
Hello:

Currently I am using qfbsolve(Qfb(1,0,1),N,3) to find pairs of squares summing to N (if possible)

The web page https://pari.math.u-bordeaux.fr/dochtml/html/Arithmetic_functions.html#qfbsolve states that "The integer n can also be given by its factorization matrix fa = factor(n) or by the pair [n, fa]."

Currently my values for N < 1,000,000 and in fact are only primes or composites of primes of 1 mod 4 only.

How much of a speed up would I obtain by doing qfbsolve(Qfb(1,0,1),[N,factor(N)],3)  instead?

Randall