K. Frank: You raise the very important point of embedded processors that
may not have an FPU. I venture that it's the responsibility of the designer
to use processors that have an FPU when the application requires one. The
middle ground is the problem - where some FP is used but software FP is
a
Hello Jon!
On Sun, May 1, 2011 at 10:17 AM, JonY wrote:
> On 5/1/2011 22:03, James K Beard wrote:
>> Now, about that math library problem...
>> ...
>> AFAIK, all currently produced FPUs are IEEE compliant. Are there some out
>> there that are still bucking the inevitable?
>
> -- Putting mingw li
I would hope that in the interests of performance and maintainable code that
we simply allow rounding bits on any targets that use 64-bit FPU that may
still be in service somewhere. I would not expect such processors to be
doing such critical work that this would be a real problem.
One thing that
On 5/1/2011 22:03, James K Beard wrote:
> Now, about that math library problem -- it seems to me that asking full
> mantissa accuracy and agreement for 32-bit and 64-bit floating point is
> reasonable. Differences between FPUs makes full mantissa accuracy and
> agreement at 80 bits a bit of a ques
On 5/1/2011 19:01, Peter Rockett wrote:
> On 01/05/11 08:43, Keith Marshall wrote:
>> On 30/04/11 23:14, Kai Tietz wrote:
>>> long double (80-bit):
>>> digits for mantissa:64
>>> digit:18, min exp:-16381 (-4931 10th)
>>> max exp: 16384 (4932 10th)
>>> Epsilon:1.0842e-019
>>>
>>> Which means
