From: "Kewen.Lin" <linkw@linux.ibm.com>
To: 钟居哲 <juzhe.zhong@rivai.ai>
Cc: gcc-patches <gcc-patches@gcc.gnu.org>,
Jeff Law <jeffreyalaw@gmail.com>, rdapp <rdapp@linux.ibm.com>,
"richard.sandiford" <richard.sandiford@arm.com>,
rguenther <rguenther@suse.de>
Subject: Re: [PATCH] VECT: Add WHILE_LEN pattern for decrement IV support for auto-vectorization
Date: Thu, 13 Apr 2023 15:29:35 +0800 [thread overview]
Message-ID: <0fb64ee1-1bf6-2a5f-4214-7567af5facbd@linux.ibm.com> (raw)
In-Reply-To: <8D9731A1540E082A+202304122122129793085@rivai.ai>
Hi Juzhe,
on 2023/4/12 21:22, 钟居哲 wrote:
> Thanks Kewen.
>
> It seems that this proposal WHILE_LEN can help s390 when using --param vect-partial-vector-usage=2 compile option.
>
Yeah, IMHO, the previous sequence vs. the proposed sequence are like:
int
foo (int *__restrict a, int *__restrict b, int n)
{
if (n <= 0)
return 0;
int iv = 0;
int len = MIN (n, 16);
int sum = 0;
do
{
sum += a[len] + b[len];
iv += 16;
int n1 = MIN (n, iv); // line A
int n2 = n - n1;
len = MIN (n2, 16);
}
while (n > iv);
return sum;
}
vs.
int
foo (int *__restrict a, int *__restrict b, int n)
{
if (n <= 0)
return 0;
int len;
int sum = 0;
do
{
len = MIN (n, 16);
sum += a[len] + b[len];
n -= len;
}
while (n > 0);
return sum;
}
it at least saves one MIN (at line A) and one length preparation in the
last iteration (it's useless since loop ends). But I think the concern
that this proposed IV isn't recognized as simple iv may stay. I tried
to compile the above source files on Power, the former can adopt doloop
optimization but the latter fails to.
> Would you mind apply this patch && support WHILE_LEN in s390 backend and test it to see the overal benefits for s390
> as well as the correctness of this sequence ?
Sure, if all of you think this approach and this revision is good enough to go forward for this kind of evaluation,
I'm happy to give it a shot, but only for rs6000. ;-) I noticed that there are some discussions on withdrawing this
WHILE_LEN by using MIN_EXPR instead, I'll stay tuned.
btw, now we only adopt vector with length on the epilogues rather than the main vectorized loops, because of the
non-trivial extra costs for length preparation than just using the normal vector load/store (all lanes), so we don't
care about the performance with --param vect-partial-vector-usage=2 much. Even if this new proposal can optimize
the length preparation for --param vect-partial-vector-usage=2, the extra costs for length preparation is still
unavoidable (MIN, shifting, one more GPR used), we would still stay with default --param vect-partial-vector-usage=1
(which can't benefit from this new proposal).
BR,
Kewen
next prev parent reply other threads:[~2023-04-13 7:59 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-07 1:47 juzhe.zhong
2023-04-07 3:23 ` Li, Pan2
2023-04-11 12:12 ` juzhe.zhong
2023-04-11 12:44 ` Richard Sandiford
2023-04-12 7:00 ` Richard Biener
2023-04-12 8:00 ` juzhe.zhong
2023-04-12 8:42 ` Richard Biener
2023-04-12 9:15 ` juzhe.zhong
2023-04-12 9:29 ` Richard Biener
2023-04-12 9:42 ` Robin Dapp
2023-04-12 11:17 ` Richard Sandiford
2023-04-12 11:37 ` juzhe.zhong
2023-04-12 12:24 ` Richard Sandiford
2023-04-12 14:18 ` 钟居哲
2023-04-13 6:47 ` Richard Biener
2023-04-13 9:54 ` juzhe.zhong
2023-04-18 9:32 ` Richard Sandiford
2023-04-12 12:56 ` Kewen.Lin
2023-04-12 13:22 ` 钟居哲
2023-04-13 7:29 ` Kewen.Lin [this message]
2023-04-13 13:44 ` 钟居哲
2023-04-14 2:54 ` Kewen.Lin
2023-04-14 3:09 ` juzhe.zhong
2023-04-14 5:40 ` Kewen.Lin
2023-04-14 3:39 ` juzhe.zhong
2023-04-14 6:31 ` Kewen.Lin
2023-04-14 6:39 ` juzhe.zhong
2023-04-14 7:41 ` Kewen.Lin
2023-04-14 6:52 ` Richard Biener
2023-04-12 11:42 ` Richard Biener
[not found] ` <2023041217154958074655@rivai.ai>
2023-04-12 9:20 ` juzhe.zhong
2023-04-19 21:53 ` 钟居哲
2023-04-20 8:52 ` Richard Sandiford
2023-04-20 8:57 ` juzhe.zhong
2023-04-20 9:11 ` Richard Sandiford
2023-04-20 9:19 ` juzhe.zhong
2023-04-20 9:22 ` Richard Sandiford
2023-04-20 9:50 ` Richard Biener
2023-04-20 9:54 ` Richard Sandiford
2023-04-20 10:38 ` juzhe.zhong
2023-04-20 12:05 ` Richard Biener
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=0fb64ee1-1bf6-2a5f-4214-7567af5facbd@linux.ibm.com \
--to=linkw@linux.ibm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jeffreyalaw@gmail.com \
--cc=juzhe.zhong@rivai.ai \
--cc=rdapp@linux.ibm.com \
--cc=rguenther@suse.de \
--cc=richard.sandiford@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).