[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gcc-bugs
Subject:    [Bug other/32172] Behavior of -ffast-math
From:       "dominiq at lps dot ens dot fr" <gcc-bugzilla () gcc ! gnu ! org>
Date:       2007-05-31 21:34:37
Message-ID: 20070531213437.18709.qmail () sourceware ! org
[Download RAW message or body]



------- Comment #4 from dominiq at lps dot ens dot fr  2007-05-31 21:34 -------
> Note we do expand anint in this optimal way already (for x86 and SSE math). 

I forgot to mention PPC (see comment #3) for which the implementation of
anint(x) and friends (mod(x), modulo(x)) are quite bad compared to xlf.

> Note that we don't have infrastructure to obey paranthesis 
> but only to not do re-association at all.

I am not sure to understand what you mean.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32172

[prev in list] [next in list] [prev in thread] [next in thread]