Thanks for the answer! (and sorry for my very late reply)

Le dimanche 3 mai 2015 18:54:43 UTC+2, Bill Hart a écrit :
>
>
>
> On 3 May 2015 at 12:51, <20100.d...@gmail.com <javascript:>> wrote:
>
>> Hello,
>>
>> I would like to know whether it it is possible to compare a mpz_t and a 
>> mpq_t more efficiently than:
>>  - creating a mpq_t with denominator 1
>>  - swap the numerator with the mpz_t
>>  - call mpq_cmp
>>  - swap back the numerator with the mpz_t
>>  - free the memory of the mpq_t
>>
>
> Sure. Take a look at the mpq_numref and mpq_denref functions. There is no 
> need to create an mpq_t, though you do need a temporary mpz_t for the 
> result of the multiplication. 
>
> Of course, any function which performed such a comparison would have to 
> create this internally anyway. So in terms of efficiency, it's about the 
> best you can do. You can special case denominator = 1 or numerator = 0 and 
> a few other cases like that. But it's about the best you can do.
>

I asked the question after I had a look at mpq_cmp. This function first 
tries some comparison based on the size of numerators and denominators. So 
I deduced that calling mpq_cmp would be faster (in simple cases) rather 
than direct cross multiplication even if I had to initialize an mpq_t with 
a denominator 1.


>> If not, would it be useful to have this as a standalone function (like 
>> "mpq_cmp_mpz")? There is a need within Sage [1,2] where often integer and 
>> rational types are mixed.
>>
>
> I'd be inclined to just add such a thing as a macro if you need it often. 
> It's only about four lines to implement.
>
 
What would be this macro? Calling mpq_cmp? Cross multiplication?

Vincent

-- 
You received this message because you are subscribed to the Google Groups 
"mpir-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to mpir-devel+unsubscr...@googlegroups.com.
To post to this group, send email to mpir-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/mpir-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to