From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nick Ing-Simmons To: richard.earnshaw AT arm.com Cc: N8TM AT aol.com, gcc AT gcc.gnu.org, Marc Espie , David Edelsohn Subject: Re: type based aliasing again Date: Tue, 14 Sep 1999 10:31:00 -0000 Message-id: <199909141726.SAA08296@tiuk.ti.com> References: <199909141706.SAA08354@sun52.NIS.cambridge> X-SW-Source: 1999-09/msg00546.html Richard Earnshaw writes: >dje@watson.ibm.com said: >> I think that -ffast-math and the behavior of compilers from other >> vendors is good precedent that -fstrict-aliasing should be removed >> from the standard -O# optimization levels and accessed orthogonally. > >-ffast-math doesn't give any precedent at all. Yes it does. Both -ffast-math and -fstrict-aliasing turn on optimizations which may break existing code. >-ffast-math enables >optimizations that the relevant standards disallow. -fstrict-aliasing >does exactly the opposite. But the math standard has had much longer to settle in and get adopted in the existing code base. When new C alias standard has been around as long as IEEE-754 then it may be time to flip the default... -- Nick Ing-Simmons Via, but not speaking for: Texas Instruments Ltd. From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nick Ing-Simmons To: richard.earnshaw@arm.com Cc: N8TM@aol.com, gcc@gcc.gnu.org, Marc Espie , David Edelsohn Subject: Re: type based aliasing again Date: Thu, 30 Sep 1999 18:02:00 -0000 Message-ID: <199909141726.SAA08296@tiuk.ti.com> References: <199909141706.SAA08354@sun52.NIS.cambridge> X-SW-Source: 1999-09n/msg00546.html Message-ID: <19990930180200.s_8-wT4j8A-x1uDzT_i8PMmAvmR__eUyd7yAZI8F81U@z> Richard Earnshaw writes: >dje@watson.ibm.com said: >> I think that -ffast-math and the behavior of compilers from other >> vendors is good precedent that -fstrict-aliasing should be removed >> from the standard -O# optimization levels and accessed orthogonally. > >-ffast-math doesn't give any precedent at all. Yes it does. Both -ffast-math and -fstrict-aliasing turn on optimizations which may break existing code. >-ffast-math enables >optimizations that the relevant standards disallow. -fstrict-aliasing >does exactly the opposite. But the math standard has had much longer to settle in and get adopted in the existing code base. When new C alias standard has been around as long as IEEE-754 then it may be time to flip the default... -- Nick Ing-Simmons Via, but not speaking for: Texas Instruments Ltd.