Date: Tue, 27 Aug 2013 00:30:12 +1000 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Steve Kargl <sgk@troutmask.apl.washington.edu> Cc: freebsd-numerics@freebsd.org, Bruce Evans <brde@optusnet.com.au> Subject: Re: (2nd time) tweaks to erff() threshold values Message-ID: <20130827002821.K2328@besplex.bde.org> In-Reply-To: <20130823171526.GA47736@troutmask.apl.washington.edu> References: <20130822213315.GA6708@troutmask.apl.washington.edu> <20130823202257.Q1593@besplex.bde.org> <20130823171526.GA47736@troutmask.apl.washington.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, 23 Aug 2013, Steve Kargl wrote:
> On Fri, Aug 23, 2013 at 09:12:33PM +1000, Bruce Evans wrote:
>>
>> @ if(ix < 0x3f580000) { /* |x|<0.84375 */
>> @ if(ix < 0x38800000) { /* |x|<2**-14 */
>> @ if (ix < 0x04000000) /* |x|<0x1p-119 */
>> @ return (8*x+efx8*x)/8; /*avoid spurious underflow */
>
> To keep the diff between s_erf.c and s_erff.c small, do you
> want my to update s_erf.c with the last line above?
Yes, including the style change from *0.125 to /8 and any similar style
changes you can find (or don't change this anywhere), and the comment
on the threshold for underflow.
Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130827002821.K2328>