Re: remove-duplicates performances

[Stefan Monnier]

> I go down to a list of 10 elements and it still faster:

I'm not surprised the break-even is less than 10.

> liste de 2X10 éléments:
> remove-duplicates  1           0.000209      0.000209
> remove-dups        1           3.6e-05       3.6e-05

> liste de 2X5 éléments:
> remove-duplicates  1           7.3e-05       7.3e-05
> remove-dups        1           6.4e-05       6.4e-05

Hmm... so it's faster to do it for 20 than for 10?

I expect it is common to call remove-duplicates with very short lists
(shorter than 10 for sure) that present (almost) no duplication.


        Stefan