loop with getSimplifyDenomsFlag?

[Qian Yun]

I find this when trying getSimplifyDenomsFlag(true)
with the 80k integrals benchmark.


f := (a^2+2*a*b*x^(1/4)+b^2*x^(1/2))^(3/2)
ratDenom(1/f) -- returns in 2s
setSimplifyDenomsFlag(true)
1/f -- doesn't return


Is there something wrong in 'algreduc'?

- Qian

[Waldek Hebisch]

On Tue, Aug 04, 2026 at 06:10:54PM +0800, Qian Yun wrote:
> I find this when trying getSimplifyDenomsFlag(true)
> with the 80k integrals benchmark.
>
>
> f := (a^2+2*a*b*x^(1/4)+b^2*x^(1/2))^(3/2)
> ratDenom(1/f) -- returns in 2s
> setSimplifyDenomsFlag(true)
> 1/f -- doesn't return

Hmm, for me it finishes after about 60s.

> Is there something wrong in 'algreduc'?

'algreduc' uses different tactic than 'ratDenom'. It makes sense
to look why this computation takes so much time, but 'algreduc'
look OK.

--
Waldek Hebisch