This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Change definition of complex::norm
- To: dewar at gnat dot com
- Subject: Re: Change definition of complex::norm
- From: Gabriel Dos Reis <gdr at codesourcery dot com>
- Date: 01 Nov 2001 15:43:50 +0100
- Cc: gdr at codesourcery dot com, lucier at math dot purdue dot edu, gcc at gcc dot gnu dot org, hjstein at bloomberg dot com, nbecker at fred dot net
- Organization: CodeSourcery, LLC
- References: <20011101122518.D58B0F28AF@nile.gnat.com>
dewar@gnat.com writes:
| <<Must libstdc++ assume IEEE-754 airthmetic on all platform supported by
| GCC?
| >>
|
| Well as we have discussed at length, it is not clear what such an assumption
| means. Many machines are not quite complete 754 at the hardware level (e.g.
| Alpha, SGI), and for efficient code generation, there are problems on other
| machines. Furthermore, IEEE-754 is a hardware-software standard that describes
| certain operations that must be available, it does not describe the mapping
| to various languages.
|
| So it is better to say what you mean, i.e. exactly what assumptions you
| want to make, at the C level, rather than just make a broad brush statement
| about the entire standard (note that parts of C99 specifying the mapping are
| also relevant here).
I don't think I made a broad brush statement about the entire standard.
Exactly what I means is a question asking if libstdc++ should
IEEE-754 arithmetic model on all plateform supported by GCC.
-- Gaby