X-Recipient: archive-cygwin AT delorie DOT com X-SWARE-Spam-Status: No, hits=-1.6 required=5.0 tests=AWL,BAYES_00,SPF_PASS X-Spam-Check-By: sourceware.org Message-ID: <4A6CB4C8.9030209@aol.com> Date: Sun, 26 Jul 2009 12:55:52 -0700 From: Tim Prince Reply-To: tprince AT computer DOT org User-Agent: Thunderbird 2.0.0.22 (Windows/20090605) MIME-Version: 1.0 To: gcc-help AT gcc DOT gnu DOT org, cygwin AT cygwin DOT com Subject: [Fwd: [Fwd: Re: cygwin gcc compatibility with MSVC numerics]] Content-Type: multipart/mixed; boundary="------------020103020602040508070400" X-AOL-IP: 64.12.78.139 Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com --------------020103020602040508070400 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Leena DOT Padgaonkar AT patni DOT com wrote: > My basic problem is that the cygwin floating addition is giving different results than VS 2008 for certain float values .Both the environments are on the same PC. So I was wondering about the floating point format used in cygwin. > > Btw, the makefile which I am using are having following options OPTFLAGS = -I . -O3 -funroll-loops -mtune=pentium3 -ffast-math -mfancy-math-387 > > Only 64-bit Windows passes the same settings of x87 precision mode (53-bit) and SSE abrupt underflow mode to both gcc and MSVC built .exe. -ffast-math would not be recommended for similarity to MSVC build, where none of the aggressive options would normally be in use. Only the abrupt underflow setting matches MSVC. If /Ox is set for MSVC, similar optimization should be obtained with gcc -O3. If you are looking for full performance, and don't need compatibility with 10-year-old CPUs, you would normally set /fp:fast /arch:SSE2 in MSVC, and corresponding -march=pentium-m -mfpmath=sse (or newer -march) in gcc. If you don't set /arch:SSE2 /fp:fast in MSVC, you imply K&R style promotion of certain float expressions to double, such as you get with 387 math in gcc. --------------020103020602040508070400 Content-Type: message/rfc822; name="[Fwd: Re: cygwin gcc compatibility with MSVC numerics].eml" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename*0="[Fwd: Re: cygwin gcc compatibility with MSVC numerics].eml" Return-Path: Received: from rly-mg05.mx.aol.com (rly-mg05.mail.aol.com [172.20.83.111]) by air-mg07.mail.aol.com (v124.15) with ESMTP id MAILINMG071-a114a6cb397d2; Sun, 26 Jul 2009 15:50:59 -0400 Received: from cs-ems2.ieee.org (cs-ems2.ieee.org [140.98.193.208]) by rly-mg05.mx.aol.com (v124.15) with ESMTP id MAILRELAYINMG053-a114a6cb397d2; Sun, 26 Jul 2009 15:50:47 -0400 Received: from gemini4.ieee.org (gemini4.ieee.org [140.98.193.189]) by cs-ems2.ieee.org (8.12.11/8.12.11) with ESMTP id n6QJokf6013026 for ; Sun, 26 Jul 2009 15:50:46 -0400 Received: from gemini4.ieee.org (gemini4.ieee.org [127.0.0.1]) by postfix.imss70 (Postfix) with ESMTP id C43EC69C820 for ; Sun, 26 Jul 2009 15:43:29 -0400 (EDT) Received: from hormel7.ieee.org (hormel7.ieee.org [140.98.193.230]) by gemini4.ieee.org (Postfix) with ESMTP id B0CB569C81D for ; Sun, 26 Jul 2009 15:43:29 -0400 (EDT) Received: from sourceware.org (sourceware.org [209.132.176.174]) by hormel7.ieee.org (8.13.8/8.13.8/Debian-3) with SMTP id n6QJojvq006655 for ; Sun, 26 Jul 2009 15:50:46 -0400 Received: (qmail 23817 invoked by alias); 26 Jul 2009 19:50:35 -0000 Received: (qmail 23808 invoked by uid 22791); 26 Jul 2009 19:50:34 -0000 X-SWARE-Spam-Status: No, hits=-1.5 required=5.0 tests=AWL,BAYES_00,SPF_PASS X-Spam-Check-By: sourceware.org Received: from imo-m12.mx.aol.com (HELO imo-m12.mail.aol.com) (64.12.143.100) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Sun, 26 Jul 2009 19:50:28 +0000 Received: from imo-da02.mx.aol.com (imo-da02.mx.aol.com [205.188.169.200]) by imo-m12.mail.aol.com (v107.10) with ESMTP id RELAYIN1-24a6cb370b6; Sun, 26 Jul 2009 15:50:08 -0400 Received: from N8TM AT aol DOT com by imo-da02.mx.aol.com (mail_out_v40_r1.5.) id s.c33.59bb3904 (37581); Sun, 26 Jul 2009 15:50:16 -0400 (EDT) Received: from [192.168.1.102] (99-13-231-72.lightspeed.snjsca.sbcglobal.net [99.13.231.72]) by cia-mb05.mx.aol.com (v124.15) with ESMTP id MAILCIAMB056-92cd4a6cb3752bb; Sun, 26 Jul 2009 15:50:16 -0400 Message-ID: <4A6CB376 DOT 1030605 AT aol DOT com> Date: Sun, 26 Jul 2009 12:50:14 -0700 From: Tim Prince Reply-To: tprince AT computer DOT org User-Agent: Thunderbird 2.0.0.22 (Windows/20090605) MIME-Version: 1.0 To: cygwin AT cygwin DOT com, gcc-help AT gcc DOT org, Leena DOT Padgaonkar AT patni DOT com Subject: [Fwd: Re: cygwin gcc compatibility with MSVC numerics] Content-Type: multipart/mixed; boundary="------------060503020200030408000508" Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT cygwin DOT com Mail-Followup-To: cygwin AT cygwin DOT com Delivered-To: mailing list cygwin AT cygwin DOT com X-Spam-Score: undef - cygwin.com is whitelisted. X-CanIt-Geo: ip=209.132.176.174; country=US; region=NC; city=Raleigh; postalcode=27606; latitude=35.7463; longitude=-78.7239; metrocode=560; areacode=919; http://maps.google.com/maps?q=35.7463,-78.7239&z=6 X-IEEE-UCE-Settings: tprince AT computer_society (inherits from 32_HI_TAG-LO_BLOCK,default) X-IEEE-UCE-Stats-ID: Bayes signature not available X-Scanned-By: IEEE UCE Filtering Service (UCE . ieee . org) on 140.98.193.230 X-AOL-IP: 140.98.193.208 X-Mailer: Unknown (No Version) --------------060503020200030408000508 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Leena DOT Padgaonkar AT patni DOT com wrote: > My basic problem is that the cygwin floating addition is giving > different results than VS 2008 for certain float values .Both the > environments are on the same PC. So I was wondering about the floating > point format used in cygwin. > > Btw, the makefile which I am using are having following options > OPTFLAGS = -I . -O3 -funroll-loops -mtune=pentium3 -ffast-math > -mfancy-math-387 > > Only 64-bit Windows passes the same settings of x87 precision mode (53-bit) and SSE abrupt underflow mode to both gcc and MSVC built .exe. -ffast-math would not be recommended for similarity to MSVC build, where none of the aggressive options would normally be in use. Only the abrupt underflow setting matches MSVC. If /Ox is set for MSVC, similar optimization should be obtained with gcc -O3. If you are looking for full performance, and don't need compatibility with 10-year-old CPUs, you would normally set /fp:fast /arch:SSE2 in MSVC, and corresponding -march=pentium-m -mfpmath=sse (or newer -march) in gcc. If you don't set /arch:SSE2 /fp:fast in MSVC, you imply K&R style promotion of certain float expressions to double, such as you get with 387 math in gcc. --------------060503020200030408000508 Content-Type: message/rfc822; name="Re: cygwin gcc compatibility with MSVC numerics.eml" Content-Transfer-Encoding: 7bit Content-Disposition: inline; filename="Re: cygwin gcc compatibility with MSVC numerics.eml" Return-Path: Received: from [192.168.1.102] (99-13-231-72.lightspeed.snjsca.sbcglobal.net [99.13.231.72]) by cia-ma05.mx.aol.com (v124.15) with ESMTP id MAILCIAMA058-91464a6cabea119; Sun, 26 Jul 2009 15:18:05 -0400 Message-ID: <4A6CABEB DOT 4000701 AT aol DOT com> Date: Sun, 26 Jul 2009 12:18:03 -0700 From: Tim Prince Reply-To: tprince AT computer DOT org User-Agent: Thunderbird 2.0.0.22 (Windows/20090605) MIME-Version: 1.0 To: Leena DOT Padgaonkar AT patni DOT com Subject: Re: cygwin gcc compatibility with MSVC numerics References: <10746230 DOT 2742571248621615816 DOT JavaMail DOT nabble AT isper DOT nabble DOT com> In-Reply-To: <10746230 DOT 2742571248621615816 DOT JavaMail DOT nabble AT isper DOT nabble DOT com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-AOL-IP: 99.13.231.72 X-Mailer: Unknown (No Version) Leena DOT Padgaonkar AT patni DOT com wrote: > My basic problem is that the cygwin floating addition is giving different results than VS 2008 for certain float values .Both the environments are on the same PC. So I was wondering about the floating point format used in cygwin. > > Btw, the makefile which I am using are having following options > OPTFLAGS = -I . -O3 -funroll-loops -mtune=pentium3 -ffast-math -mfancy-math-387 > > Only 64-bit Windows passes the same settings of x87 precision mode (53-bit) and SSE abrupt underflow mode to both gcc and MSVC built .exe. -ffast-math would not be recommended for similarity to MSVC build, where none of the aggressive options would normally be in use. Only the abrupt underflow setting matches MSVC. If /Ox is set for MSVC, similar optimization should be obtained with gcc -O3. If you are looking for full performance, and don't need compatibility with 10-year-old CPUs, you would normally set /fp:fast /arch:SSE2 in MSVC, and corresponding -march=pentium-m -mfpmath=sse (or newer -march) in gcc. If you don't set /arch:SSE2 /fp:fast in MSVC, you imply K&R style promotion of certain float expressions to double, such as you get with 387 math in gcc. --------------060503020200030408000508 Content-Type: text/plain; charset=us-ascii -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple --------------060503020200030408000508-- --------------020103020602040508070400 Content-Type: text/plain; charset=us-ascii -- Problem reports: http://cygwin.com/problems.html FAQ: http://cygwin.com/faq/ Documentation: http://cygwin.com/docs.html Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple --------------020103020602040508070400--