This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug tree-optimization/65206] Vectorized version of loop is removed.
- From: "rguenth at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Thu, 26 Feb 2015 11:40:24 +0000
- Subject: [Bug tree-optimization/65206] Vectorized version of loop is removed.
- Auto-submitted: auto-generated
- References: <bug-65206-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65206
--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
For the masked load/store case we could also simply put the "real" memory
access
in place of the pointer argument. To make that valid GIMPLE we could wrap it
inside a fake VIEW_CONVERT_EXPR for example - like one converting to a
char[sizeof (ref)], this would make it appear as aggregate. Of course when
we inspect the masked load/store we'd have to strip that VIEW_CONVERT_EXPR
again. But I don't think it would harm anyone seeing that VIEW_CONVERT_EXPR.