This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: The complex problems
- To: Greg McGary <gkm at eng dot ascend dot com>
- Subject: Re: The complex problems
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Sat, 18 Sep 1999 00:16:13 -0600
- cc: Michael Meissner <meissner at cygnus dot com>, rth at cygnus dot com, egcs at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <ms4sgsbxjx.fsf@tucson-net-82.eng.ascend.com>you write:
> I know you agree about the RTL part, but what about doing away with
> the complex machine modes? AFAICT, the only purpose these serve is to
> push builtin complex arithmetic into the backend (optabs.c). If the
> complex modes are dropped, we couldn't do it in the backend anymore,
> except by adding a flag flag on the rtx to identify it as complex.
> The alternative is to expand complex arithmetic at the tree level.
> What are your thoughts about this part?
So, how do you propose we emit code to perform complex arithmetic? Somehow,
somewhere we need to know how to build RTL those kind of operations.
jeff