[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
reproducibility of tininess detection for binary formats
David Hough wrote:
However I do agree with draft 1.5 that there is no particular reason
why the underflow rule has to be the same on binary and decimal formats.
I don't think that both radices would go through the exact same rounding
and exception detection paths.
It is not a matter of implementation (path sharing etc.), but of clarity
and aesthetics. If the rules are different for binary and decimal it
will be more difficult to remember which is which.
In any case, subsequently I made a case for a common choice based on the
merits of the choice for binary on its own: I think it is significantly
easier to implement artificial tininess-before on tininess-after machines
than the other way around.
Michel.
Sent: 2007-10-14 06:22:16 UTC