From owner-freebsd-alpha Sat Jul 8 2:25:15 2000 Delivered-To: freebsd-alpha@freebsd.org Received: from eel.radicalmedia.com (eel.radicalmedia.com [204.254.246.9]) by hub.freebsd.org (Postfix) with ESMTP id B9B2F37B579 for ; Sat, 8 Jul 2000 02:25:12 -0700 (PDT) (envelope-from phiber@eel.radicalmedia.com) Received: (from phiber@localhost) by eel.radicalmedia.com (8.9.3/8.9.3) id FAA24815; Sat, 8 Jul 2000 05:22:37 -0400 (EDT) Date: Sat, 8 Jul 2000 05:22:36 -0400 From: Mark Abene To: Doug Rabson Cc: freebsd-alpha@FreeBSD.ORG Subject: Re: floating point troubles... Message-ID: <20000708052236.D22774@radicalmedia.com> References: <20000708040526.C22774@radicalmedia.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.1i In-Reply-To: ; from Doug Rabson on Sat, Jul 08, 2000 at 09:42:42AM +0100 Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org See my first message on the subject... Compiling with -mieee made no difference, I still got the floating exception and core dump. Disabling SIGFPE was the only sure way to stop this, fpsetmask wasn't doing the job it should be. Strange, because fpsetmask *does* work as used in the lame mp3 encoder. I don't know if fpsetmask not working is specific to overflowing a pow(), or if the problem is much broader in scope... -Mark On Sat, Jul 08, 2000 at 09:42:42AM +0100, Doug Rabson wrote: > > Try compiling the -mieee flag to gcc. For some programs, the alpha > requires slightly stricter, slower floating point code to be generated so > that unusual results (INFs, NaNs etc) can be generated by the kernel's > floating point emulator. > > -- > Doug Rabson Mail: dfr@nlsystems.com > Nonlinear Systems Ltd. Phone: +44 20 8442 9037 > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-alpha" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message