From: Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
To: gcc Patches <gcc-patches@gcc.gnu.org>,
Kyrill Tkachov <kyrylo.tkachov@arm.com>
Subject: [ARM] PR98435: Missed optimization in expanding vector constructor
Date: Fri, 4 Jun 2021 12:55:49 +0530 [thread overview]
Message-ID: <CAAgBjMmiorDvtoZxu5aFL8ZND1uOgfPx+oPBycZ2Fd-gk2LPWQ@mail.gmail.com> (raw)
Hi,
As mentioned in PR, for the following test-case:
#include <arm_neon.h>
bfloat16x4_t f1 (bfloat16_t a)
{
return vdup_n_bf16 (a);
}
bfloat16x4_t f2 (bfloat16_t a)
{
return (bfloat16x4_t) {a, a, a, a};
}
Compiling with arm-linux-gnueabi -O3 -mfpu=neon -mfloat-abi=softfp
-march=armv8.2-a+bf16+fp16 results in f2 not being vectorized:
f1:
vdup.16 d16, r0
vmov r0, r1, d16 @ v4bf
bx lr
f2:
mov r3, r0 @ __bf16
adr r1, .L4
ldrd r0, [r1]
mov r2, r3 @ __bf16
mov ip, r3 @ __bf16
bfi r1, r2, #0, #16
bfi r0, ip, #0, #16
bfi r1, r3, #16, #16
bfi r0, r2, #16, #16
bx lr
This seems to happen because vec_init pattern in neon.md has VDQ mode
iterator, which doesn't include V4BF. In attached patch, I changed
mode
to VDQX which seems to work for the test-case, and the compiler now generates:
f2:
vdup.16 d16, r0
vmov r0, r1, d16 @ v4bf
bx lr
However, the pattern is also gated on TARGET_HAVE_MVE and I am not
sure if either VDQ or VDQX are correct modes for MVE since MVE has
only 128-bit vectors ?
Thanks,
Prathamesh
next reply other threads:[~2021-06-04 7:26 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-04 7:25 Prathamesh Kulkarni [this message]
2021-06-04 7:45 ` Christophe Lyon
2021-06-09 10:28 ` Prathamesh Kulkarni
2021-06-14 8:01 ` Prathamesh Kulkarni
2021-06-21 8:34 ` Prathamesh Kulkarni
2021-06-24 16:31 ` Kyrylo Tkachov
2021-06-28 8:37 ` Prathamesh Kulkarni
2021-06-28 8:40 ` Kyrylo Tkachov
2021-06-28 9:17 ` Christophe LYON
2021-06-29 10:46 ` Prathamesh Kulkarni
2021-06-30 15:21 ` Christophe LYON
2021-07-01 10:56 ` Prathamesh Kulkarni
2021-07-06 7:05 ` Prathamesh Kulkarni
2021-07-06 8:03 ` Kyrylo Tkachov
2021-07-06 9:25 ` Prathamesh Kulkarni
2021-07-06 9:28 ` Kyrylo Tkachov
2021-07-06 10:16 ` Christophe Lyon
2021-08-03 9:29 ` Christophe Lyon
2021-08-03 10:56 ` Prathamesh Kulkarni
2021-08-03 15:22 ` Christophe Lyon
2021-08-05 12:27 ` Prathamesh Kulkarni
2021-08-05 12:34 ` Christophe Lyon
2021-08-06 8:59 ` Prathamesh Kulkarni
2021-08-06 9:19 ` Christophe Lyon
2021-08-06 9:50 ` Prathamesh Kulkarni
2021-08-06 12:01 ` Christophe Lyon
2021-08-09 5:07 ` Prathamesh Kulkarni
2021-08-09 16:19 ` Christophe Lyon
2021-08-13 7:04 ` Prathamesh Kulkarni
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=CAAgBjMmiorDvtoZxu5aFL8ZND1uOgfPx+oPBycZ2Fd-gk2LPWQ@mail.gmail.com \
--to=prathamesh.kulkarni@linaro.org \
--cc=gcc-patches@gcc.gnu.org \
--cc=kyrylo.tkachov@arm.com \
/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).