[PATCH 1/2] Improve lgammaf range for very small cases
Andoni Arregi
andoni.arregui@gtd-gmbh.de
Thu Feb 10 16:16:37 GMT 2022
Hi Corinna,
On Thu, 2022-02-10 at 15:58 +0100, Corinna Vinschen wrote:
> Hi Andoni,
>
> On Feb 9 17:28, Andoni Arregi wrote:
> > The original cut for small arguments at |x|<2**-70 (copied from the
> > double version) produces that when computing nadj we get a subnormal
> > number for t*x and thus, the division of pi/subnormal will be INF and
> > the logarithm of it too, which is wrong as a result for lgammaf in this
> > range.
> > The proposed new limit seems to be safe and has been tested to
> > produce accurate results.
> > (Courtesy of Andreas Jung, ESA)
> > ---
> > newlib/libm/math/erf_lgamma.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/newlib/libm/math/erf_lgamma.c
> > b/newlib/libm/math/erf_lgamma.c
>
> This is broken. It should be on one line, so apparently your
> mailer broke it via autowrapping lines.
>
Thanks for letting me know. Now trying it again without autowrapping.
> > index f88f63092..84d02159b 100644
> > --- a/newlib/libm/math/erf_lgamma.c
> > +++ b/newlib/libm/math/erf_lgamma.c
> > @@ -168,7 +168,7 @@ static float zero= 0.0000000000e+00;
> > *signgamp = -1;
> > return one/(x-x);
> > }
> > - if(ix<0x1c800000) { /* |x|<2**-70, return -log(|x|) */
> > + if(ix<0x30800000) { /* |x|<2**-30, return -log(|x|) */
> > if(hx<0) {
> > *signgamp = -1;
> > return -__ieee754_logf(-x);
> > --
> > 2.35.1
>
> But even if I fix this manually, your patch doesn't apply, neither with
> `git am', nor with `patch -p1'. I have to admit that I don't understand
> why...
>
I just checked `git am' myself before resubmitting. Hope it works now.
> Another question, would you mind to send patch series with a cover letter,
> please?
>
Sure, sorry for that.
>
> Thanks,
> Corinna
--
Andoni Arregi
Geschäftsführer
GTD GmbH
Ravensburger Str. 32a, 88677 Markdorf
T: +49 7544 96440 22 | M: +49 151 65620499 | F: +49 7544 96440 29
http://www.gtd-gmbh.de andoni.arregi@gtd-gmbh.de
More information about the Newlib
mailing list