This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [RFC PATCH] implement fma() as builtin x87 or SSE intrinsic

From: Paolo Bonzini <bonzini at gnu dot org>
To: gcc-patches at gcc dot gnu dot org
Cc: 0 at main dot gmane dot org,0 at main dot gmane dot org,0 at main dot gmane dot org
Date: Fri, 07 May 2004 13:59:38 +0200
Subject: Re: [RFC PATCH] implement fma() as builtin x87 or SSE intrinsic
References: <409B5031.5080000@kss-loka.si> <87d65gebo8.fsf@informatik.uni-tuebingen.de> <409B722B.3030206@kss-loka.si>

Call to __builtin_fma() should be handled as BUILT_IN_FMA{,F,L} optimization in builtins.c, if flag_unsafe_math_optimizations flag is set. This way, call to fma() would be transformed to multiply and add for all architectures.

BTW: BUILT_IN_HYPOT{,F,L} could also be handled this way.

I'm not sure about which precision is requested for fma and hypot?

Paolo

References:
- [RFC PATCH] implement fma() as builtin x87 or SSE intrinsic
  - From: Uros Bizjak
- Re: [RFC PATCH] implement fma() as builtin x87 or SSE intrinsic
  - From: Falk Hueffner
- Re: [RFC PATCH] implement fma() as builtin x87 or SSE intrinsic
  - From: Uros Bizjak

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]