This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Fix BIV detection with promoted loop indexes
- From: kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)
- To: rth at redhat dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Fri, 22 Feb 02 21:22:46 EST
- Subject: Re: Fix BIV detection with promoted loop indexes
Last time I looked at this closely it was still a win to have the
subregs added, rather than let optabs insert sign (or worse zero)
extensions.
Well, the comments say it's doing the *same* thing as expand_binop does and
that was my assumption. If the comments were false, then indeed this may
have been a loss in some cases. But from reading the code in expand_binop,
the variable "no_extend" is set, so no extension will indeed be done.
Therefore, I don't understand your comment.
As I said, my concern was the 16-bit case, where we end up doing *two*
sets of SUBREGs.