From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <SRS0=YkbG=CZ=gmail.com=crazylht@sourceware.org>
Received: from mail-yb1-xb36.google.com (mail-yb1-xb36.google.com [IPv6:2607:f8b0:4864:20::b36])
	by sourceware.org (Postfix) with ESMTPS id D78E83858C60
	for <gcc-patches@gcc.gnu.org>; Fri,  7 Jul 2023 06:44:01 +0000 (GMT)
DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org D78E83858C60
Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=gmail.com
Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=gmail.com
Received: by mail-yb1-xb36.google.com with SMTP id 3f1490d57ef6-bc379e4c1cbso1756597276.2
        for <gcc-patches@gcc.gnu.org>; Thu, 06 Jul 2023 23:44:01 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=gmail.com; s=20221208; t=1688712240; x=1691304240;
        h=content-transfer-encoding:cc:to:subject:message-id:date:from
         :in-reply-to:references:mime-version:from:to:cc:subject:date
         :message-id:reply-to;
        bh=UCMAzj2M+o7Lx1Eq676Dp33udDLIGanRpqGvkFtbcDk=;
        b=UE60s+bUFsVLQuuR7Zw3jqWhrgrPsPV9a9D+saSURrXdoGilKfSIVH52+A285V423i
         7/x0aZg7PV/uxXbYl7Dwrg0qBop06/TvEAxJeU+K6Ct1Q53/ZymcyBxc/rr1Lts2o5GV
         cexTFCNCInTZPhl6fdawA8PmeQByE3iSWvJgzN3fVs35pOdQ2FCP+HKm9igYZovXkv80
         F8HSesgfp7nerOMhjbFlx3otRir/R883M9eujHNlPxR6jcOX8yqLOPzcsek/Dgn69hmz
         O31jEidPGyoDVN7Q2LCMr1NT8/Td8W3R+PdIeneqVTBl7IrPJZoAcnU1URlpbvIuy/jy
         I1fg==
X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
        d=1e100.net; s=20221208; t=1688712240; x=1691304240;
        h=content-transfer-encoding:cc:to:subject:message-id:date:from
         :in-reply-to:references:mime-version:x-gm-message-state:from:to:cc
         :subject:date:message-id:reply-to;
        bh=UCMAzj2M+o7Lx1Eq676Dp33udDLIGanRpqGvkFtbcDk=;
        b=KKnJm1YOhDEZei1VUh6//9A9PzeIZhim3nM24AQikt/Uq4FN1S85R4y526PPeJU78Q
         wyRDPz7j/GoqxvyDj4V8CSP9/KWINIPIe5Yn9pBdZsCr67EDqJ1diASVmJs/mTtptYmu
         hNltaataqRtQIfj2bqJCnHs0a4VbbsMQ52bl52kVy4EtZ4xwpY5yQmzr7JalPUmg85wJ
         Ye2mkspo8qsDJLyftmsa3XMr+9Z3a2BLx8gY0p+kogdYVW01CW62Z4bjFHkDQsWj2LIL
         /RW43mHv8FqWJuMJcKtmNefNfvoiuuCLeNfo92c9+wR3iKt6JYO34MGIltag/aTJdDmF
         T7XQ==
X-Gm-Message-State: ABy/qLa1q20oD3wF2R9jjNRt7uuz+XgCarP0msyCuSAiE6TOo4cl8wwU
	0fxBev4AOtrcVgHzmA+QlXWb9mxeoIU3GMPs1WQ=
X-Google-Smtp-Source: APBJJlH+iqeAMrqoEZolR9BPrZKcSZVaT5+k2A8yeLEUpyB5tsYMZeENwwQFgi3Mx/u14AgIhnX/A1nPy36oKolgt1s=
X-Received: by 2002:a25:1e55:0:b0:c4f:c0eb:451b with SMTP id
 e82-20020a251e55000000b00c4fc0eb451bmr4868982ybe.25.1688712240410; Thu, 06
 Jul 2023 23:44:00 -0700 (PDT)
MIME-Version: 1.0
References: <20230704025035.1074040-1-hongtao.liu@intel.com> <de09b46b951df9cd8a475af56ebc125f@ispras.ru>
In-Reply-To: <de09b46b951df9cd8a475af56ebc125f@ispras.ru>
From: Hongtao Liu <crazylht@gmail.com>
Date: Fri, 7 Jul 2023 14:50:54 +0800
Message-ID: <CAMZc-bwUk=MiuUCCQtj1dagWP1P7tW0SsArgSqwS3Zp8RjcG3g@mail.gmail.com>
Subject: Re: [PATCH] Break false dependence for vpternlog by inserting vpxor.
To: simonaytes.yan@ispras.ru
Cc: liuhongt <hongtao.liu@intel.com>, gcc-patches@gcc.gnu.org, hjl.tools@gmail.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Spam-Status: No, score=-1.6 required=5.0 tests=BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,FREEMAIL_FROM,RCVD_IN_DNSWL_NONE,SPF_HELO_NONE,SPF_PASS,TXREP,T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6
X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org
List-Id: <gcc-patches.gcc.gnu.org>

On Thu, Jul 6, 2023 at 11:46=E2=80=AFPM <simonaytes.yan@ispras.ru> wrote:
>
> > +; False dependency happens on destination register which is not really
> > +; used when moving all ones to vector register
> > +(define_split
> > +  [(set (match_operand:VMOVE 0 "register_operand")
> > +     (match_operand:VMOVE 1 "int_float_vector_all_ones_operand"))]
> > +  "TARGET_AVX512F && reload_completed
> > +  && (<MODE_SIZE> =3D=3D 64 || EXT_REX_SSE_REG_P (operands[0]))"
> > +  [(set (match_dup 0) (match_dup 2))
> > +   (parallel
> > +     [(set (match_dup 0) (match_dup 1))
> > +      (unspec [(match_dup 0)] UNSPEC_INSN_FALSE_DEP)])]
> > +  "operands[2] =3D CONST0_RTX (<MODE>mode);")
>
> I think we shouldnt emit PXOR when optimizing for size. So should change
> define_split:
> define_split
>    [(set (match_operand:VMOVE 0 "register_operand")
>         (match_operand:VMOVE 1 "int_float_vector_all_ones_operand"))]
>    "TARGET_AVX512F && reload_completed
>    && (<MODE_SIZE> =3D=3D 64 || EXT_REX_SSE_REG_P (operands[0]))
>    && optimize_insn_for_speed_p ()"
>    [(set (match_dup 0) (match_dup 2))
>     (parallel
>       [(set (match_dup 0) (match_dup 1))
>        (unspec [(match_dup 0)] UNSPEC_INSN_FALSE_DEP)])]
>    "operands[2] =3D CONST0_RTX (<MODE>mode);")
Yes, will do. I'm still working on breaking the false depence for
pternlog in newly added pattern *iornot<mode>3,*xnor<mode>3 and
*<nlogic><mode>3.
Will repost the patch when it's done.



--=20
BR,
Hongtao