Re: Unused code in mpn_mul_fft_decompose?

2019-03-05 Thread Torbjörn Granlund
paul zimmermann writes: Anyway this code is now obsolete (see [1] and [2]). No, the current GMP code is not obsolete. We have improvements in various state of development, but it is unlikely that any of that will be ready for inclusion in GMP soon. -- Torbjörn Please encrypt, key id

Re: Unused code in mpn_mul_fft_decompose?

2019-03-05 Thread paul zimmermann
Dear Mikhail, > From: Mikhail Hogrefe > Date: Mon, 4 Mar 2019 21:44:50 -0500 > > I was digging around in the GMP 6.1.2 source, and I noticed something in > mpn/generic/mul_fft.c. > > In mpn_mul_fft_decompose, there is an if statement checking 'dif > Kl', > which is equivalent to 'nl > 2

Unused code in mpn_mul_fft_decompose?

2019-03-05 Thread Mikhail Hogrefe
I was digging around in the GMP 6.1.2 source, and I noticed something in mpn/generic/mul_fft.c. In mpn_mul_fft_decompose, there is an if statement checking 'dif > Kl', which is equivalent to 'nl > 2 * K * l'. But there is a comment above the function saying 'We must have nl <= 2*K*l.', suggesting