Re: [boost] Specific-Width Floating-Point Typedefs

Subject: Re: [boost] Specific-Width Floating-Point Typedefs
From: Christopher Kormanyos (e_float_at_[hidden])
Date: 2013-03-28 18:11:53

>> But would welcome comments from Boosters to help
>> us improve this proposal.


> Personally I've grown used to assume float is float32_t and
> double is  float64_t, but it could be useful to have proper
> typedefs for this.

Thank you for your comments, Mathias.

For sure, I wish it were that simple.

It sounds simple enough, but it's really a multifaceted issue.
In my microcontroller and numerical travels, I have used
* native float with 24 and 32 bits
* native double with 32, 64, 80, and 128 bits
* native long double with 64, 80, and 128 bits

It's a nightmare!

And just when you think you wrote some good code
for double, it doesn't run somewhere because double
is all of a sudden float.

We are aiming at portability improvements here.

Sincerely, Chris.