public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "hubicka at kam dot mff.cuni.cz" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/101908] [12 regression] cray regression with -O2 -ftree-slp-vectorize compared to -O2
Date: Thu, 28 Oct 2021 13:09:30 +0000	[thread overview]
Message-ID: <bug-101908-4-NHkaWA8AEE@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-101908-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101908

--- Comment #10 from hubicka at kam dot mff.cuni.cz ---
>        |     b = 2.0 * ray.dir.x * (ray.orig.x - sph->pos.x) +                
> #
>        |       movupd   (%rdi),%xmm5                                          
> #
>        |     2.0 * ray.dir.y * (ray.orig.y - sph->pos.y) +                    
> #
>        |     2.0 * ray.dir.z * (ray.orig.z - sph->pos.z);                     
> #
>   0.02 |       movsd    0x10(%rdi),%xmm9                                      
> #
>   0.01 |       movupd   0xb8(%rsp),%xmm13                                     
> #
>  37.67 |       movupd   0xa0(%rsp),%xmm15                                
> 
> so we pass struct ray on the stack(?) and perform SSE loads from it but
> the argument passing does
> 
>   0.88 |       movups %xmm2,(%rsp)                                            
> #
>   0.22 |       movups %xmm3,0x10(%rsp)                                        
> #
>  43.81 |       movups %xmm4,0x20(%rsp)                                        
> #
>   0.66 |       call   ray_sphere                   

Adding Martin to CC.  I think we could teach ipa-sra to, with -flto,
turn the structure either to scalar arguments or to be passed by
reference which would allow us to hoist its initialization out of the
loop body.

Honza

  parent reply	other threads:[~2021-10-28 13:09 UTC|newest]

Thread overview: 50+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-14 14:27 [Bug middle-end/101908] New: " hubicka at gcc dot gnu.org
2021-08-16  8:33 ` [Bug middle-end/101908] " crazylht at gmail dot com
2021-08-16  9:05 ` [Bug tree-optimization/101908] " rguenth at gcc dot gnu.org
2021-08-16  9:06 ` rguenth at gcc dot gnu.org
2021-08-25  7:47 ` crazylht at gmail dot com
2021-10-28 12:26 ` [Bug tree-optimization/101908] [12 regression] " hubicka at gcc dot gnu.org
2021-10-28 12:39 ` hubicka at gcc dot gnu.org
2021-10-28 13:02 ` rguenth at gcc dot gnu.org
2021-10-28 13:05 ` rguenth at gcc dot gnu.org
2021-10-28 13:06 ` hubicka at kam dot mff.cuni.cz
2021-10-28 13:09 ` hubicka at kam dot mff.cuni.cz [this message]
2021-10-28 13:09 ` rguenth at gcc dot gnu.org
2021-10-28 13:12 ` hubicka at kam dot mff.cuni.cz
2021-10-28 13:13 ` rguenth at gcc dot gnu.org
2021-10-28 13:15 ` rguenth at gcc dot gnu.org
2021-10-28 13:21 ` rguenth at gcc dot gnu.org
2021-10-29 13:58 ` hubicka at kam dot mff.cuni.cz
2022-01-20 11:12 ` rguenth at gcc dot gnu.org
2022-01-20 11:26 ` [Bug target/101908] " rguenth at gcc dot gnu.org
2022-01-20 11:52 ` rguenth at gcc dot gnu.org
2022-02-24  9:54 ` rguenth at gcc dot gnu.org
2022-02-25  3:57 ` crazylht at gmail dot com
2022-02-25  7:33 ` rguenth at gcc dot gnu.org
2022-02-25  8:26 ` lili.cui at intel dot com
2022-02-25  8:31 ` lili.cui at intel dot com
2022-02-25 15:27 ` hjl.tools at gmail dot com
2022-02-28  1:29 ` crazylht at gmail dot com
2022-02-28  1:30 ` crazylht at gmail dot com
2022-02-28  5:13 ` lili.cui at intel dot com
2022-03-01  9:33 ` crazylht at gmail dot com
2022-03-10 13:47 ` crazylht at gmail dot com
2022-03-10 13:54 ` crazylht at gmail dot com
2022-03-10 13:55 ` crazylht at gmail dot com
2022-03-11  7:11 ` crazylht at gmail dot com
2022-03-11  8:32 ` rguenth at gcc dot gnu.org
2022-03-11  8:48 ` crazylht at gmail dot com
2022-03-11 10:41 ` rguenth at gcc dot gnu.org
2022-03-11 13:14 ` crazylht at gmail dot com
2022-03-11 13:27 ` rguenther at suse dot de
2022-03-14  9:05 ` crazylht at gmail dot com
2022-03-14  9:16 ` rguenther at suse dot de
2022-03-15  1:52 ` crazylht at gmail dot com
2022-03-15  7:14 ` rguenther at suse dot de
2022-03-29  3:40 ` crazylht at gmail dot com
2022-03-29  4:01 ` crazylht at gmail dot com
2022-03-29  6:47 ` rguenther at suse dot de
2022-03-29  8:27 ` crazylht at gmail dot com
2022-03-29 10:00 ` rguenther at suse dot de
2022-04-05  5:17 ` cvs-commit at gcc dot gnu.org
2022-04-05  6:24 ` rguenth at gcc dot gnu.org

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=bug-101908-4-NHkaWA8AEE@http.gcc.gnu.org/bugzilla/ \
    --to=gcc-bugzilla@gcc.gnu.org \
    --cc=gcc-bugs@gcc.gnu.org \
    /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).