On Sat, Dec 26, 2015 at 2:44 PM, Nicolas George wrote:
> Le sextidi 6 nivôse, an CCXXIV, Ganesh Ajjanagadde a écrit :
>> Point taken though: the only clear changes are ones which are Pareto
>> improvements for speed and accuracy
>
> ... AND code complexity. For example, the pow(10, x) call in test
Le sextidi 6 nivôse, an CCXXIV, Ganesh Ajjanagadde a écrit :
> Point taken though: the only clear changes are ones which are Pareto
> improvements for speed and accuracy
... AND code complexity. For example, the pow(10, x) call in testsrc is
neither speed-critical nor accuracy-critical.
Regards,
On Sat, Dec 26, 2015 at 3:37 AM, Michael Niedermayer
wrote:
[...]
>
> i dont know if the speed difference of exp10 really matters for any
> speed relevant code (probably not) and i dont know if the accuracy
> increase really matters for any code either
Accuracy increases rarely matter for anythin
Hi,
On Sat, Dec 26, 2015 at 6:37 AM, Michael Niedermayer wrote:
> On Fri, Dec 25, 2015 at 08:10:16PM -0800, Ganesh Ajjanagadde wrote:
> > On Fri, Dec 25, 2015 at 9:11 AM, Ganesh Ajjanagadde
> > wrote:
> > > This tightens the errors by doing a first order Taylor approximation of
> > > one obviou
On Fri, Dec 25, 2015 at 08:10:16PM -0800, Ganesh Ajjanagadde wrote:
> On Fri, Dec 25, 2015 at 9:11 AM, Ganesh Ajjanagadde
> wrote:
> > This tightens the errors by doing a first order Taylor approximation of
> > one obvious source of error. This results in negligible runtime
> > slowdown, but impro
On Fri, Dec 25, 2015 at 9:11 AM, Ganesh Ajjanagadde
wrote:
> This tightens the errors by doing a first order Taylor approximation of
> one obvious source of error. This results in negligible runtime
> slowdown, but improves worst case relative error by ~ 30%, worst case
> ulp count from 673 to 39
This tightens the errors by doing a first order Taylor approximation of
one obvious source of error. This results in negligible runtime
slowdown, but improves worst case relative error by ~ 30%, worst case
ulp count from 673 to 391. Another illustration is:
arg : -303.13720760010643270798027
