Date: Mon, 1 Jun 2015 17:26:19 -0700 From: Steve Kargl <sgk@troutmask.apl.washington.edu> To: "Montgomery-Smith, Stephen" <stephen@missouri.edu> Cc: "freebsd-numerics@freebsd.org" <freebsd-numerics@freebsd.org> Subject: Re: Let's get moving Message-ID: <20150602002619.GA45253@troutmask.apl.washington.edu> In-Reply-To: <556CEB8C.90605@missouri.edu> References: <556CEB8C.90605@missouri.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jun 01, 2015 at 11:32:29PM +0000, Montgomery-Smith, Stephen wrote: > Hey people, > > I was looking at https://wiki.freebsd.org/Numerics, and we are stalling. The wiki is out-of-date. I've committed patches for erfcl, erfl, and lgammal. I'm working on powl and I'll do tgammal after I get powl done. I have versions of sincos[fl], which aren't quite ready, but I would like to have available for my work on the Bessel functions. > For example, we have working code for clog and clogf - why don't we > commit it? > > One of the reasons I bring this up is I would like to create a port for > http://librsb.sourceforge.net/, and the only barriers are the lack of > cpowf and cpow. Right now I intend to commit it with a rather > simplistic patch: >From src/math_private.h * The C99 standard intends x+I*y to be used for this, but x+I*y is * currently unusable in general since gcc introduces many overflow, * underflow, sign and efficiency bugs by rewriting I*y as * (0.0+I)*(y+0.0*I) and laboriously computing the full complex product. * In particular, I*Inf is corrupted to NaN+I*Inf, and I*-0 is corrupted * to -0.0+I*0.0. Last time I checked, clang was even worse with regards to I. -- Steve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150602002619.GA45253>