[lno] [RFC] if-conversion and auto vectorizer

[Toon Moene]

Richard Henderson wrote:

> For the same reasons that we preserve a set of rules that are
> beyond C89 such that it's possible to write kernels.

Ah, I already thought so [well, actually, I was thinking "embedded 
systems"].

Can I have an option (to be set by default to "yes" by the various 
Fortran front ends) that says:

No valid (Fortran) program can cause the generation of floating point 
exceptions or memory access violations ?

Even Fortran 2003 will only allow 1.0/0.0 to cause a (legal) exception 
if you first specify:

USE IEEE_EXCEPTIONS

and the (as yet hypothetical) front end implementing that Standard could 
easily flip the switch on seeing that statement *for the translation unit*.

Thanks in advance,

-- 
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG  Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)