Re: [boost] fixed pt binary lib

From: Neal Becker (ndbecker2_at_[hidden])
Date: 2008-08-01 07:00:35

Anthony Williams wrote:

> "John C. Femiani" <john.femiani_at_[hidden]> writes:
> 
>> Anthony Williams wrote:
>>> "John C. Femiani" <john.femiani_at_[hidden]> writes:
>>>   
>>>> Also, do your implementations for /= and *= actually round? You
>>>> actually seem to be using the round policy for conversions, is that
>>>> right? Have you looked at the numeric/conversion stuff? (I haven't
>>>> grokked it yet).
>>>>     
>>>
>>> The division throws away the lower "frac_bits" of the result:
>> So why doesn't it use the rounding policy? I kind of expect it to be
>> used anytime quantization occurs.
> 
> I'll leave Neal to answer that (it's his code), but getting division
> right is a pain.  If you followed the link to my code, you'll see that
> correct division is not trivial.
> 
> Anthony

Multiplication is a bit easier to understand.  I thought about mixed-arith mult, but I think it opens a can of worms.  What do you want the result to be?  So, I decided to only support arith of like-types.  If you want the result of mult to be a larger type, then you simply ask for it:

fixed<8,8> x, y;
fixed<16,16> z = fixed<16,16> (x) * fixed<16,16> (y);

As far as division, I wasn't quite sure and the implementation shown is only a guess.  Suggestions welcome.