View mode: basic / threaded / horizontal-split · Log in · Help
October 17, 2008
Re: CustomFloat
Andrei Alexandrescu Wrote:

> I found myself in need for smaller and specialized floating point types, 
> where I configure the exact configuration (sign, mantissa, exponent, 
> bias). See http://en.wikipedia.org/wiki/Half_precision for a half 
> precision number slated for inclusion in IEEE 754r.
> 
> Would it be interesting to add a CustomFloat template to phobos? I'm 
> thinking along the lines of:
> 
> template CustomFloat!(bool sign, uint mantissa,
>          uint exponent, uint bias)
> {
>      ...
> }
> 
> So half-precision numbers are:
> 
> alias CustomFloat!(true, 5, 10, 15) HalfFloat;
> 
> There are quite a few details to kink out but this is definitely doable. 
> Numbers like 24-bit floating point and even 8-bit floating point would 
> be easy to support too. For now CustomFloat would be intended 
> exclusively as a compact storage mechanism; only conversion to the 
> standard floating points would be implemented. Later, maybe we can get 
> to implement some operations natively at least on machines that support 
> them in hardware. I wanted to gauge interest in the topic.
> 
> 
> Andrei

I've done some work on a set of decimal floating point numbers consistent with the new 754r decimal formats approved in June. Some of that work would be applicable here, I think.

It would probably make sense to add the decimal formats at the same time. Let me know if I can help.

Paul
October 17, 2008
Re: CustomFloat
Paul D. Anderson Wrote:

> Andrei Alexandrescu Wrote:
> 
> > I found myself in need for smaller and specialized floating point types, 
> > where I configure the exact configuration (sign, mantissa, exponent, 
> > bias). See http://en.wikipedia.org/wiki/Half_precision for a half 
> > precision number slated for inclusion in IEEE 754r.
> > 
> > Would it be interesting to add a CustomFloat template to phobos? I'm 
> > thinking along the lines of:
> > 
> > template CustomFloat!(bool sign, uint mantissa,
> >          uint exponent, uint bias)
> > {
> >      ...
> > }
> > 
> > So half-precision numbers are:
> > 
> > alias CustomFloat!(true, 5, 10, 15) HalfFloat;
> > 
> > There are quite a few details to kink out but this is definitely doable. 
> > Numbers like 24-bit floating point and even 8-bit floating point would 
> > be easy to support too. For now CustomFloat would be intended 
> > exclusively as a compact storage mechanism; only conversion to the 
> > standard floating points would be implemented. Later, maybe we can get 
> > to implement some operations natively at least on machines that support 
> > them in hardware. I wanted to gauge interest in the topic.
> > 
> > 
> > Andrei
> 
> I've done some work on a set of decimal floating point numbers consistent with the new 754r decimal formats approved in June. Some of that work would be applicable here, I think.
> 
> It would probably make sense to add the decimal formats at the same time. Let me know if I can help.
> 
> Paul
> 

Here's a link to info regarding the decimal formats: http://speleotrove.com/decimal/
October 17, 2008
Re: CustomFloat
Paul D. Anderson wrote:
> Andrei Alexandrescu Wrote:
> 
>> I found myself in need for smaller and specialized floating point
>> types, where I configure the exact configuration (sign, mantissa,
>> exponent, bias). See http://en.wikipedia.org/wiki/Half_precision
>> for a half precision number slated for inclusion in IEEE 754r.
>> 
>> Would it be interesting to add a CustomFloat template to phobos?
>> I'm thinking along the lines of:
>> 
>> template CustomFloat!(bool sign, uint mantissa, uint exponent, uint
>> bias) { ... }
>> 
>> So half-precision numbers are:
>> 
>> alias CustomFloat!(true, 5, 10, 15) HalfFloat;
>> 
>> There are quite a few details to kink out but this is definitely
>> doable. Numbers like 24-bit floating point and even 8-bit floating
>> point would be easy to support too. For now CustomFloat would be
>> intended exclusively as a compact storage mechanism; only
>> conversion to the standard floating points would be implemented.
>> Later, maybe we can get to implement some operations natively at
>> least on machines that support them in hardware. I wanted to gauge
>> interest in the topic.
>> 
>> 
>> Andrei
> 
> I've done some work on a set of decimal floating point numbers
> consistent with the new 754r decimal formats approved in June. Some
> of that work would be applicable here, I think.
> 
> It would probably make sense to add the decimal formats at the same
> time. Let me know if I can help.

Sounds interesting, and thanks for the offer. I'd be glad to add decimal 
types too, in case there's enough interest. I think monetary 
calculations can be helped, and what don't we do nowadays to help people 
involved in monetary calculations? :o)

Andrei
Top | Discussion index | About this forum | D home