From: Segher Boessenkool <segher@kernel.crashing.org>
To: Xi Ruoyao <xry111@mengyan1223.wang>
Cc: 172060045@hdu.edu.cn, gcc-help <gcc-help@gcc.gnu.org>,
bin.cheng@linux.alibaba.com, rguenther@suse.de, ook@ucw.cz,
Jan Hubicka <hubicka@ucw.cz>
Subject: Re: Why vectorization didn't turn on by -O2
Date: Sun, 9 May 2021 08:44:37 -0500 [thread overview]
Message-ID: <20210509134436.GX10366@gate.crashing.org> (raw)
In-Reply-To: <5ac14eb276d81223d2d677593600f37c0fd1ce48.camel@mengyan1223.wang>
On Sun, May 09, 2021 at 12:54:08AM +0800, Xi Ruoyao via Gcc-help wrote:
> On Sat, 2021-05-08 at 20:07 +0800, 172060045@hdu.edu.cn wrote:
> > Recently I noticed that gcc -O2 didn't turn on vectorization
> > optimization,
> > which it turns on in clang -O2.
> >
> > Does GCC think it involves the trade-off of space speed, or other
> > considerations?
-O2 is for optimisations that (almost) never degrade code quality. -O3
is for those that are only beneficial "on average".
> It's just a decision I think. The "original" reason may be that
> vectorization can make code *slower* on some cases.
Yup. Vectorisation always causes hugely different code.
> There was some discussion about enabling -ftree-loop-vectorization at -
> O2 for x86, but that was too late (for GCC 9):
AFAIK the current plan is to enable vectorisation at -O2 with a more
conservative cost model. This will be a generic change, for all
architectures, and hopefully will arrive in GCC 12.
Segher
next prev parent reply other threads:[~2021-05-09 13:45 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-08 12:07 172060045
2021-05-08 16:54 ` Xi Ruoyao
2021-05-09 13:44 ` Segher Boessenkool [this message]
2021-05-10 8:21 ` Richard Biener
2021-05-10 9:11 ` Jan Hubicka
2021-05-10 9:27 ` Segher Boessenkool
2021-05-10 12:10 ` Jan Hubicka
2021-05-10 9:24 ` Segher Boessenkool
2021-05-17 16:03 ` Jan Hubicka
2021-05-17 18:56 ` Richard Sandiford
2021-08-04 8:21 ` Hongtao Liu
2021-08-04 8:22 ` Richard Sandiford
2021-08-04 8:31 ` Richard Biener
2021-08-04 9:10 ` Richard Sandiford
2021-08-04 9:56 ` Segher Boessenkool
2021-08-04 10:22 ` Richard Sandiford
2021-08-04 21:18 ` Segher Boessenkool
2021-08-04 9:12 ` Hongtao Liu
2021-08-11 17:14 ` Jan Hubicka
2021-08-14 14:22 ` Jan Hubicka
2021-08-16 8:03 ` Richard Biener
2021-08-16 3:22 ` Kewen.Lin
2021-08-16 6:00 ` Hongtao Liu
2021-08-16 6:09 ` Hongtao Liu
2021-08-24 2:21 ` Hongtao Liu
2021-08-04 8:36 ` Hongtao Liu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210509134436.GX10366@gate.crashing.org \
--to=segher@kernel.crashing.org \
--cc=172060045@hdu.edu.cn \
--cc=bin.cheng@linux.alibaba.com \
--cc=gcc-help@gcc.gnu.org \
--cc=hubicka@ucw.cz \
--cc=ook@ucw.cz \
--cc=rguenther@suse.de \
--cc=xry111@mengyan1223.wang \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).