From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18776 invoked by alias); 12 Apr 2010 01:54:27 -0000 Received: (qmail 16765 invoked by uid 48); 12 Apr 2010 01:54:10 -0000 Date: Mon, 12 Apr 2010 01:54:00 -0000 Message-ID: <20100412015410.16763.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug target/32180] Paranoia UCB GSL TestFloat libm tests fail - accuracy of recent gcc math poor In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "rob1weld at aol dot com" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2010-04/txt/msg01099.txt.bz2 ------- Comment #26 from rob1weld at aol dot com 2010-04-12 01:54 ------- (In reply to comment #25) > I understand that this is INVALID because all the points raised by comment #21. > If crlibm is better than what we have, but we cannot use it, it is the same as > if it didn't exist. It is possible to use various other Programs (or Libraries with a bit of programming) and run them seperately on a test script to check that both programs give the same result. We could compile and run 'crlibm' (and some other programs) _without_ incorporating it's code into gcc and compare the results of the two Programs with each other using 'diff' . Doing the above is useful for _testing_ but ultimatley we should either 'fix' our code to use the default flags (so the math is not broken) _or_ use the flags that I suggested as the defaults and avoid the problem altogether. We could (but do not have to) use a different math Library. The existance of other code that works correctly proves that we could write our own code differently (properly) and have correct results instead of insisting that we can only have a wrong result and that we should accept it. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32180