[libre-riscv-dev] [isa-dev] Re: FP transcendentals (trigonometry, root/exp/log) proposal

[Hendrik Boom]

On Thu, Aug 08, 2019 at 05:37:48AM -0700, lkcl wrote:
> On Thursday, August 8, 2019 at 7:56:15 PM UTC+8, Jacob Lifshay wrote:
> 
> > 
> > One more part: hw can implement a less accurate mode as if a more accurate mode was selected, so, for example, hw can implement all modes using hw that produces correctly-rounded results (which will be the most accurate mode defined) and just ignore the mode field since correct-rounding is not less accurate than any of the defined modes.
> 
> Hmm don't know. Hendrik pointed out that Ahmdahl / IBM370 mainframe problem that extra accuracy caused.

I suspect that the number of potential customers that were worried 
about incompatibility vastly exceeded the number that actually neede 
incorrect results.  Later models of Amdahl machines aparently had a 
switch to choose between accuracy and comatibility.

-- hendrik

> 
> I don't know if that lesson from history matters [in 2019].
> 
> No clue. Don't know enough to offer an opinion either way. Anyone any recommendations?

How many potential users will be worried about bitwise compatibility 
with their previous machines, which will probably be incompatible 
anyway?  Intels, AMDs, Nvidias, Raspberries, etc?

-- hendrik