From mboxrd@z Thu Jan  1 00:00:00 1970
X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/652
Path: news.gmane.org!not-for-mail
From: Szabolcs Nagy <nsz@port70.net>
Newsgroups: gmane.linux.lib.musl.general
Subject: Re: correctly rounded sqrt
Date: Thu, 15 Mar 2012 04:23:04 +0100
Message-ID: <20120315032304.GJ5728@port70.net>
References: <20120314165604.GF5728@port70.net>
 <20120314190123.GH5728@port70.net>
 <20120315014647.GI5728@port70.net>
Reply-To: musl@lists.openwall.com
NNTP-Posting-Host: plane.gmane.org
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Trace: dough.gmane.org 1331781799 4508 80.91.229.3 (15 Mar 2012 03:23:19 GMT)
X-Complaints-To: usenet@dough.gmane.org
NNTP-Posting-Date: Thu, 15 Mar 2012 03:23:19 +0000 (UTC)
To: musl@lists.openwall.com
Original-X-From: musl-return-653-gllmg-musl=m.gmane.org@lists.openwall.com Thu Mar 15 04:23:18 2012
Return-path: <musl-return-653-gllmg-musl=m.gmane.org@lists.openwall.com>
Envelope-to: gllmg-musl@plane.gmane.org
Original-Received: from mother.openwall.net ([195.42.179.200])
	by plane.gmane.org with smtp (Exim 4.69)
	(envelope-from <musl-return-653-gllmg-musl=m.gmane.org@lists.openwall.com>)
	id 1S81HY-0001Qx-9R
	for gllmg-musl@plane.gmane.org; Thu, 15 Mar 2012 04:23:16 +0100
Original-Received: (qmail 9709 invoked by uid 550); 15 Mar 2012 03:23:15 -0000
Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm
Precedence: bulk
List-Post: <mailto:musl@lists.openwall.com>
List-Help: <mailto:musl-help@lists.openwall.com>
List-Unsubscribe: <mailto:musl-unsubscribe@lists.openwall.com>
List-Subscribe: <mailto:musl-subscribe@lists.openwall.com>
Original-Received: (qmail 9701 invoked from network); 15 Mar 2012 03:23:15 -0000
Content-Disposition: inline
In-Reply-To: <20120315014647.GI5728@port70.net>
User-Agent: Mutt/1.5.21 (2010-09-15)
Xref: news.gmane.org gmane.linux.lib.musl.general:652
Archived-At: <http://permalink.gmane.org/gmane.linux.lib.musl.general/652>

* Szabolcs Nagy <nsz@port70.net> [2012-03-15 02:46:48 +0100]:
> /* dekker exact mult u*u == hi + lo */
> static void sq(long double *hi, long double *lo, long double u)
> {
> 	static const long double c = 1.0 + 0x1p33;
this was wrong, c = 1.0 + 0x1p32; the correct value

and probably
*hi = u1*u1;
*lo = u1*u2*2.0 + u2*u2;
is enough instead of
> 	*hi = u*u;
> 	*lo = (u1*u1 - *hi) + u1*u2*2.0 + u2*u2;

but now that we figured out the fpu status register based
solution this does not matter..