From: Richard Sandiford <richard.sandiford@arm.com>
To: Tamar Christina <tamar.christina@arm.com>
Cc: gcc-patches@gcc.gnu.org, nd@arm.com, Richard.Earnshaw@arm.com,
Marcus.Shawcroft@arm.com, Kyrylo.Tkachov@arm.com
Subject: Re: [PATCH 3/4]AArch64: correct dot-product RTL patterns for aarch64.
Date: Thu, 15 Jul 2021 20:44:37 +0100 [thread overview]
Message-ID: <mpt1r7z9xay.fsf@arm.com> (raw)
In-Reply-To: <20210715164019.GA5409@arm.com> (Tamar Christina's message of "Thu, 15 Jul 2021 17:40:23 +0100")
Tamar Christina <tamar.christina@arm.com> writes:
> Hi All,
>
> The previous fix for this problem was wrong due to a subtle difference between
> where NEON expects the RMW values and where intrinsics expects them.
>
> The insn pattern is modeled after the intrinsics and so needs an expand for
> the vectorizer optab to switch the RTL.
>
> However operand[3] is not expected to be written to so the current pattern is
> bogus.
>
> Instead we use the expand to shuffle around the RTL.
>
> The vectorizer expects operands[3] and operands[0] to be
> the same but the aarch64 intrinsics expanders expect operands[0] and
> operands[1] to be the same.
>
> Bootstrapped Regtested on aarch64-none-linux-gnu and no issues.
>
> Ok for master? and active branches after some stew?
>
> Thanks,
> Tamar
>
> gcc/ChangeLog:
>
> * config/aarch64/aarch64-simd.md (<sur>dot_prod<vsi2qi>): Correct
> RTL.
>
> --- inline copy of patch --
> diff --git a/gcc/config/aarch64/aarch64-simd.md b/gcc/config/aarch64/aarch64-simd.md
> index 7397f1ec5ca0cb9e3cdd5c46772f604e640666e4..51789f954affd9fa88e2bc1bcc3dacf64ccb5bde 100644
> --- a/gcc/config/aarch64/aarch64-simd.md
> +++ b/gcc/config/aarch64/aarch64-simd.md
> @@ -635,18 +635,12 @@ (define_insn "aarch64_usdot<vsi2qi>"
> ;; and so the vectorizer provides r, in which the result has to be accumulated.
> (define_expand "<sur>dot_prod<vsi2qi>"
> [(set (match_operand:VS 0 "register_operand")
> - (plus:VS (unspec:VS [(match_operand:<VSI2QI> 1 "register_operand")
> + (plus:VS (match_operand:VS 3 "register_operand")
> + (unspec:VS [(match_operand:<VSI2QI> 1 "register_operand")
> (match_operand:<VSI2QI> 2 "register_operand")]
> - DOTPROD)
> - (match_operand:VS 3 "register_operand")))]
> + DOTPROD)))]
> "TARGET_DOTPROD"
The canonical plus: operand order was the original one, so I think
it would be better to keep this rtl as-is and instead change
aarch64_<sur>dot<vsi2qi> to:
(plus:VS (unspec:VS [(match_operand:<VSI2QI> 2 "register_operand" "w")
(match_operand:<VSI2QI> 3 "register_operand" "w")]
DOTPROD)
(match_operand:VS 1 "register_operand" "0"))
Same idea for aarch64_<sur>dot_lane<vsi2qi> and
aarch64_<sur>dot_laneq<vsi2qi>.
Sorry to be awkward…
Thanks,
Richard
> -{
> - emit_insn (
> - gen_aarch64_<sur>dot<vsi2qi> (operands[3], operands[3], operands[1],
> - operands[2]));
> - emit_insn (gen_rtx_SET (operands[0], operands[3]));
> - DONE;
> -})
> +)
>
> ;; Auto-vectorizer pattern for usdot. The operand[3] and operand[0] are the
> ;; RMW parameters that when it comes to the vectorizer.
next prev parent reply other threads:[~2021-07-15 19:44 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-15 16:39 [PATCH 1/4][committed] testsuite: Fix testisms in scalar tests PR101457 Tamar Christina
2021-07-15 16:39 ` [PATCH 2/4]AArch64: correct usdot vectorizer and intrinsics optabs Tamar Christina
2021-07-15 19:34 ` Richard Sandiford
2021-07-20 12:34 ` Tamar Christina
2021-07-20 16:15 ` Richard Sandiford
2021-07-22 11:50 ` Tamar Christina
2021-07-22 18:09 ` Richard Sandiford
2021-07-15 16:40 ` [PATCH 3/4]AArch64: correct dot-product RTL patterns for aarch64 Tamar Christina
2021-07-15 19:44 ` Richard Sandiford [this message]
2021-07-22 11:51 ` Tamar Christina
2021-07-22 18:11 ` Richard Sandiford
2021-07-23 8:14 ` Tamar Christina
2021-07-26 13:56 ` Richard Sandiford
2021-07-15 16:40 ` [PATCH 4/4][AArch32]: correct dot-product RTL patterns Tamar Christina
2021-07-16 2:20 ` [PATCH 1/4][committed] testsuite: Fix testisms in scalar tests PR101457 H.J. Lu
2021-07-16 8:42 ` Tamar Christina
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=mpt1r7z9xay.fsf@arm.com \
--to=richard.sandiford@arm.com \
--cc=Kyrylo.Tkachov@arm.com \
--cc=Marcus.Shawcroft@arm.com \
--cc=Richard.Earnshaw@arm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=nd@arm.com \
--cc=tamar.christina@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).