On 6/11/26 7:44 PM, Waldek Hebisch wrote: > On Thu, Jun 11, 2026 at 06:33:03PM +0800, Qian Yun wrote: >> Testing for equality (or inequality) of EPXR is >> a core part of Kernel and other places. >> >> I find that current situation can be improved. >> >> To test equality, we do not need to compute (x-y) >> in full, we can skip the denominator part, skip gcd, > > yes > >> skip potential "algreduc" (rationalizing the denominator), >> save a few unnecessary "reduc" calls (only reduce on >> necessary algebraic kernels). > > we definitely need 'reduc'. 'algreduc' is actually about > improving efficiency, because _correct_ alternatives are > more expensive. >
Not all 'reduc' are skipped, it used to happen for commonk(x, y), now it is commonk(numer x, denom y) and commonk(denom x, numer y) separately. (The 'reduc' is only needed for multiplication, right?) 'algreduc' is not totally skipped, it is skipped for denom(x)*denom(y). - Qian -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/fricas-devel/bd142dc4-0503-4c76-b38e-70ae76d4acb6%40gmail.com.
