This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC] [autovect patch] Implement vectorization hints
- From: Zdenek Dvorak <rakdver at atrey dot karlin dot mff dot cuni dot cz>
- To: Devang Patel <dpatel at apple dot com>, stevenb at suse dot de
- Cc: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sun, 27 Feb 2005 22:22:51 +0100
- Subject: Re: [RFC] [autovect patch] Implement vectorization hints
- References: <CFE651BB-3353-4BA5-9727-F250F1767000@apple.com>
Hello,
basing the implementation of this idea on source line numbers is a sure
way to hell. You probably cannot do much better in the very early
stages. But after optimization passes, line numbers are generally very
unreliable. In particular the node that you take in the
vect_pragma_lookup function may have been moved there by some code
motion optimization from elsewhere. Or the loop itself could be created
by some optimization (jump threading may do it sometimes). And these
are just those that occured to me, I am fairly sure there are more.
Zdenek