From owner-svn-src-all@FreeBSD.ORG Mon Oct 22 13:21:12 2012 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 5330B5B9; Mon, 22 Oct 2012 13:21:12 +0000 (UTC) (envelope-from imp@FreeBSD.org) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:4f8:fff6::2c]) by mx1.freebsd.org (Postfix) with ESMTP id 21F098FC0A; Mon, 22 Oct 2012 13:21:12 +0000 (UTC) Received: from svn.freebsd.org (localhost [127.0.0.1]) by svn.freebsd.org (8.14.4/8.14.4) with ESMTP id q9MDLBY3018262; Mon, 22 Oct 2012 13:21:11 GMT (envelope-from imp@svn.freebsd.org) Received: (from imp@localhost) by svn.freebsd.org (8.14.4/8.14.4/Submit) id q9MDLB4p018260; Mon, 22 Oct 2012 13:21:11 GMT (envelope-from imp@svn.freebsd.org) Message-Id: <201210221321.q9MDLB4p018260@svn.freebsd.org> From: Warner Losh Date: Mon, 22 Oct 2012 13:21:11 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r241887 - head/lib/msun/src X-SVN-Group: head MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 22 Oct 2012 13:21:12 -0000 Author: imp Date: Mon Oct 22 13:21:11 2012 New Revision: 241887 URL: http://svn.freebsd.org/changeset/base/241887 Log: Revert r241756 Modified: head/lib/msun/src/e_expf.c Modified: head/lib/msun/src/e_expf.c ============================================================================== --- head/lib/msun/src/e_expf.c Mon Oct 22 13:20:31 2012 (r241886) +++ head/lib/msun/src/e_expf.c Mon Oct 22 13:21:11 2012 (r241887) @@ -21,68 +21,6 @@ __FBSDID("$FreeBSD$"); #include "math.h" #include "math_private.h" -/* __ieee754_expf - * Returns the exponential of x. - * - * Method - * 1. Argument reduction: - * Reduce x to an r so that |r| <= 0.5*ln2 ~ 0.34658. - * Given x, find r and integer k such that - * - * x = k*ln2 + r, |r| <= 0.5*ln2. - * - * Here r will be represented as r = hi-lo for better - * accuracy. - * - * 2. Approximation of exp(r) by a special rational function on - * the interval [0,0.34658]: - * Write - * R(r**2) = r*(exp(r)+1)/(exp(r)-1) = 2 + r*r/6 - r**4/360 + ... - * We use a special Remes algorithm on [0,0.34658] to generate - * a polynomial of degree 2 to approximate R. The maximum error - * of this polynomial approximation is bounded by 2**-27. In - * other words, - * R(z) ~ 2.0 + P1*z + P2*z*z - * (where z=r*r, and the values of P1 and P2 are listed below) - * and - * | 2 | -27 - * | 2.0+P1*z+P2*z - R(z) | <= 2 - * | | - * The computation of expf(r) thus becomes - * 2*r - * expf(r) = 1 + ------- - * R - r - * r*R1(r) - * = 1 + r + ----------- (for better accuracy) - * 2 - R1(r) - * where - * 2 4 - * R1(r) = r - (P1*r + P2*r) - * - * 3. Scale back to obtain expf(x): - * From step 1, we have - * expf(x) = 2^k * expf(r) - * - * Special cases: - * expf(INF) is INF, exp(NaN) is NaN; - * expf(-INF) is 0, and - * for finite argument, only exp(0)=1 is exact. - * - * Accuracy: - * according to an error analysis, the error is always less than - * 0.5013 ulp (unit in the last place). - * - * Misc. info. - * For IEEE float - * if x > 8.8721679688e+01 then exp(x) overflow - * if x < -1.0397208405e+02 then exp(x) underflow - * - * Constants: - * The hexadecimal values are the intended ones for the following - * constants. The decimal values may be used, provided that the - * compiler will convert from decimal to binary accurately enough - * to produce the hexadecimal values shown. - */ static const float one = 1.0, halF[2] = {0.5,-0.5,},