speed of unbalanced multiplication

[bodrato at mail.dm.unipi.it]

Ciao Paul,

Il Ven, 25 Gennaio 2013 9:24 pm, Zimmermann Paul ha scritto:
> in GMP 5.1.0, a multiplication of n x m limbs for m < n can be slower than
> a multiplication of n x n limbs. Compare for example the line starting

>         mpn_mul.1000000 mpn_mul.1000000
> 775660   #0.740046000   0.744046000

>             mpn_mul_n     mpn_mul_n
> 1000000    0.736046000  #0.728045000

Is the problem due to the unbalanced size of operands, or it depends on
the fact that FFT is poorly tuned for very big sizes?

I mean, which timing do you obtain with the following?
./speed -s $((1000000+775660)/2) mpn_mul_n mpn_mul_n

Regards,
m

-- 
http://bodrato.it/papers/combinatorics.html