From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-oa1-x33.google.com (mail-oa1-x33.google.com [IPv6:2001:4860:4864:20::33]) by sourceware.org (Postfix) with ESMTPS id 5CFB23858D35 for ; Sat, 21 Oct 2023 20:14:50 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 5CFB23858D35 Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=gmail.com ARC-Filter: OpenARC Filter v1.0.0 sourceware.org 5CFB23858D35 Authentication-Results: server2.sourceware.org; arc=none smtp.remote-ip=2001:4860:4864:20::33 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1697919292; cv=none; b=kJAhHNw4yIKOAjnvyunxD1J1vghZ5P5E7FBh5SdMQHkigUXMZwLhRrhukCvFMA2+9OJXarBmYHeY+JkWyvubNdTVy1NWAi94okvvE7UfPboO+h3dPc1OAgZc78vMvwVuCmB2ErMq2QKMTkbxRmriK1BMqCHs4TFWh7OCM9wNwV8= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1697919292; c=relaxed/simple; bh=clG0u34BcOzdxCHtp82gPRUp/dsKo95WEorzaURlNiQ=; h=DKIM-Signature:MIME-Version:From:Date:Message-ID:Subject:To; b=wi8RmU7dpmy2235Ae3K04Ihhift5ydrOQkHuaslHr4mGvCnxvgHU3UBl8g1N3I59BvndNPGKvBt8RxnLXpElXCV87HcXECHpMUA6sNSRdEG+4qCBrR3EnrN9OaM2rUrafoOC+Zpzkv+RjUBWuAHM3ivV+daJfjdMk9ZurE3aNjU= ARC-Authentication-Results: i=1; server2.sourceware.org Received: by mail-oa1-x33.google.com with SMTP id 586e51a60fabf-1e9b6f39f9eso1528184fac.2 for ; Sat, 21 Oct 2023 13:14:50 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1697919289; x=1698524089; darn=gcc.gnu.org; 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=IfLkQ51d86+3VTYTwyopdzsK/wZhLt8XTndmStfuqzM=; b=gDwU0myK0dbKcJoiNwDHLd8QuEpRZ5kPXQiJc/6ButhxMK6G0ZWtVTO/W7Nl62in8n a+lrIvFQKH756oUBNyq79Xh5LpVwG8p435doFUU5b1oCbuQwXeNKSPAMC3JZi3gjcD9r LBRd4PWPXgb8SJrz+hgI0Fy3hhdQZO45PN31brThREujDZFgVsDkHcxfq2Dli+dzMmW/ DgJrBLIrkUZa3m/9QfThNGWWLY0VlZ4kChyt0g+74WXm5QfnZW5kmWYX98LXjh3eZ9cQ RXdQ2fOwTCIqeNs7190I1qtVWwX1nbqvNorglJHH+MZHf4Nr+uLdapDvTcV+2hbeZeOc fD2w== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1697919289; x=1698524089; 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=IfLkQ51d86+3VTYTwyopdzsK/wZhLt8XTndmStfuqzM=; b=d3+TB28eVamLFhPhYv85yYbVy0ohESHk2gFVJgC5n9VtjA9eA/CigFR2jyciDnaBXH XPKAskp3LyfV/nNkwSVK6p4EfOQTzWGaslLizbxPZetq1C6crVmxaMzL9xVgCdknAV7L dnobThGXiwjMVHAlxlTEBM851Lwtp3jbKHO2xtM7oGqNJhPMsMCPZ+9MupnCzuq97k0/ H/iRyRHz3FlqAaphdB20U+a2QMCrTK3ftKiQnjnwESvuYYi9KG23ErJcH6g+54whRc+4 KG9ZW0FHcbm1qlCl6eircF4NIUX6dSZvyJCququipxer5MOhRMJXkYrRfbgRMprNPLlL UrKA== X-Gm-Message-State: AOJu0Yx9k2HqxEBRBg5lEusM1+wbtYrI0pAsMgy1vZI9p+GD5QaHNjRY lR1GIsOPSEykqrlsfrd3eOvtmgJ1DNVKxwt1KPk= X-Google-Smtp-Source: AGHT+IGmegqt206Sp+1keDPv19ESaEDmS8WOoS/VJcdNUf+xoEwHt7czVTwQ0b3ZHD5ExDATZK4djjQTdAieJC0pVac= X-Received: by 2002:a05:6358:894:b0:132:f294:77fe with SMTP id m20-20020a056358089400b00132f29477femr6685632rwj.2.1697919289048; Sat, 21 Oct 2023 13:14:49 -0700 (PDT) MIME-Version: 1.0 References: <1625969489-14344-1-git-send-email-apinski@marvell.com> In-Reply-To: From: Andrew Pinski Date: Sat, 21 Oct 2023 13:14:37 -0700 Message-ID: Subject: Re: [PATCH] move the (a-b) CMP 0 ? (a-b) : (b-a) optimization from fold_cond_expr_with_comparison to match To: Richard Biener Cc: GCC Patches Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable X-Spam-Status: No, score=-7.8 required=5.0 tests=BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,FREEMAIL_FROM,GIT_PATCH_0,RCVD_IN_DNSWL_NONE,SPF_HELO_NONE,SPF_PASS,TXREP 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: On Thu, Oct 19, 2023 at 10:13=E2=80=AFPM Andrew Pinski = wrote: > > On Mon, Jul 12, 2021 at 4:47=E2=80=AFAM Richard Biener via Gcc-patches > wrote: > > > > On Sun, Jul 11, 2021 at 4:12 AM apinski--- via Gcc-patches > > wrote: > > > > > > From: Andrew Pinski > > > > > > This patch moves the (a-b) CMP 0 ? (a-b) : (b-a) optimization > > > from fold_cond_expr_with_comparison to match. > > > > So I searched and I guess these transforms are produced from > > > > /* If we have A op 0 ? A : -A, consider applying the following > > transformations: > > > > A =3D=3D 0? A : -A same as -A > > A !=3D 0? A : -A same as A > > A >=3D 0? A : -A same as abs (A) > > A > 0? A : -A same as abs (A) > > A <=3D 0? A : -A same as -abs (A) > > A < 0? A : -A same as -abs (A) > > > > None of these transformations work for modes with signed > > zeros. If A is +/-0, the first two transformations will > > change the sign of the result (from +0 to -0, or vice > > versa). The last four will fix the sign of the result, > > even though the original expressions could be positive or > > negative, depending on the sign of A. > > > > Note that all these transformations are correct if A is > > NaN, since the two alternatives (A and -A) are also NaNs. */ > > if (!HONOR_SIGNED_ZEROS (type) > > && (FLOAT_TYPE_P (TREE_TYPE (arg01)) > > ? real_zerop (arg01) > > : integer_zerop (arg01)) > > && ((TREE_CODE (arg2) =3D=3D NEGATE_EXPR > > && operand_equal_p (TREE_OPERAND (arg2, 0), arg1, 0)) > > /* In the case that A is of the form X-Y, '-A' (arg2) may > > have already been folded to Y-X, check for that. */ > > || (TREE_CODE (arg1) =3D=3D MINUS_EXPR > > && TREE_CODE (arg2) =3D=3D MINUS_EXPR > > && operand_equal_p (TREE_OPERAND (arg1, 0), > > TREE_OPERAND (arg2, 1), 0) > > && operand_equal_p (TREE_OPERAND (arg1, 1), > > TREE_OPERAND (arg2, 0), 0)))) > > ... > > > > I wonder at which point we can remove the code from fold-const.c? > > I have to double check if after an updated patch, if that code does > anything that match does not do. > I will do that before I submit an updated patch. I looked and the main thing left is solving the stripping of sign nops that happen at the beginning of fold_cond_expr_with_comparison. I did solve part of that with the recent r14-4662-gc7609acb8a8210188d21b2cd7 but not with this new patterns; I will solve that in a separate patch. Thanks, Andrew Pinski > > > > > Some comments inline below. > > > > > OK? Bootstrapped and tested on x86_64-linux-gnu. > > > > > > gcc/ChangeLog: > > > > > > * match.pd ((A-B) CMP 0 ? (A-B) : (B - A)): > > > New patterns. > > > > > > gcc/testsuite/ChangeLog: > > > > > > * gcc.dg/tree-ssa/phi-opt-25.c: New test. > > > --- > > > gcc/match.pd | 48 ++++++++++++++++++++= -- > > > gcc/testsuite/gcc.dg/tree-ssa/phi-opt-25.c | 45 ++++++++++++++++++++ > > > 2 files changed, 90 insertions(+), 3 deletions(-) > > > create mode 100644 gcc/testsuite/gcc.dg/tree-ssa/phi-opt-25.c > > > > > > diff --git a/gcc/match.pd b/gcc/match.pd > > > index 30680d488ab..aa88381fdcb 100644 > > > --- a/gcc/match.pd > > > +++ b/gcc/match.pd > > > @@ -4040,9 +4040,7 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT) > > > (cnd (logical_inverted_value truth_valued_p@0) @1 @2) > > > (cnd @0 @2 @1))) > > > > > > -/* abs/negative simplifications moved from fold_cond_expr_with_compa= rison, > > > - Need to handle (A - B) case as fold_cond_expr_with_comparison doe= s. > > > - Need to handle UN* comparisons. > > > +/* abs/negative simplifications moved from fold_cond_expr_with_compa= rison. > > > > > > None of these transformations work for modes with signed > > > zeros. If A is +/-0, the first two transformations will > > > @@ -4098,6 +4096,50 @@ DEFINE_INT_AND_FLOAT_ROUND_FN (RINT) > > > (convert (negate (absu:utype @0)))) > > > (negate (abs @0))))) > > > ) > > > + > > > + /* (A - B) =3D=3D 0 ? (A - B) : (B - A) same as (B - A) */ > > > + (for cmp (eq uneq) > > > + (simplify > > > + (cnd (cmp (minus@0 @1 @2) zerop) @0 (minus@3 @2 @1)) > > > + (if (!HONOR_SIGNED_ZEROS (type)) > > > + @3)) > > > + (simplify > > > + (cnd (cmp (minus@0 @1 @2) zerop) integer_zerop (minus@3 @2 @1)) > > > > So that makes me think why integer_zerop? 'type' should then be > > integer and thus never HONOR_SIGNED_ZEROS. > > yes that should be done. > > > > > Don't we also need the inverted condition case for completeness? > > Yes we should. Though for phiopt we don't. > > > > > > > + (if (!HONOR_SIGNED_ZEROS (type)) > > > + @3)) > > > + (simplify > > > + (cnd (cmp @1 @2) integer_zerop (minus@3 @2 @1)) > > > > I think this needs to be (cmp:c @1 @2) > > This is now actually handled already by r14-3606-g3d86e7f4a8ae so I remov= ed it. > I will submit a new patch in the next few days for this too. > > Thanks, > Andrew Pinski > > > > > > + (if (!HONOR_SIGNED_ZEROS (type)) > > > + @3)) > > > + ) > > > + /* (A - B) !=3D 0 ? (A - B) : (B - A) same as (A - B) */ > > > + (for cmp (ne ltgt) > > > + (simplify > > > + (cnd (cmp (minus@0 @1 @2) zerop) @0 (minus @2 @1)) > > > + (if (!HONOR_SIGNED_ZEROS (type)) > > > + @0)) > > > + ) > > > + /* (A - B) >=3D/> 0 ? (A - B) : (B - A) same as abs (A - B) */ > > > + (for cmp (ge gt) > > > + (simplify > > > + (cnd (cmp (minus@0 @1 @2) zerop) @0 (minus @2 @1)) > > > + (if (!HONOR_SIGNED_ZEROS (type) > > > + && !TYPE_UNSIGNED (type)) > > > + (abs @0)))) > > > + /* (A - B) <=3D/< 0 ? (A - B) : (B - A) same as -abs (A - B) */ > > > + (for cmp (le lt) > > > + (simplify > > > + (cnd (cmp (minus@0 @1 @2) zerop) @0 (minus @2 @1)) > > > + (if (!HONOR_SIGNED_ZEROS (type) > > > + && !TYPE_UNSIGNED (type)) > > > + (if (ANY_INTEGRAL_TYPE_P (type) > > > + && !TYPE_OVERFLOW_WRAPS (type)) > > > + (with { > > > + tree utype =3D unsigned_type_for (type); > > > + } > > > + (convert (negate (absu:utype @0)))) > > > + (negate (abs @0))))) > > > + ) > > > ) > > > > > > /* -(type)!A -> (type)A - 1. */ > > > diff --git a/gcc/testsuite/gcc.dg/tree-ssa/phi-opt-25.c b/gcc/testsui= te/gcc.dg/tree-ssa/phi-opt-25.c > > > new file mode 100644 > > > index 00000000000..0f0e3170f8d > > > --- /dev/null > > > +++ b/gcc/testsuite/gcc.dg/tree-ssa/phi-opt-25.c > > > @@ -0,0 +1,45 @@ > > > +/* { dg-options "-O2 -fno-signed-zeros -fdump-tree-phiopt" } */ > > > +int minus1(int a, int b) > > > +{ > > > + int c =3D a - b; > > > + if (c =3D=3D 0) c =3D b - a; > > > + return c; > > > +} > > > +int minus2(int a, int b) > > > +{ > > > + int c =3D a - b; > > > + if (c !=3D 0) c =3D b - a; > > > + return c; > > > +} > > > +int minus3(int a, int b) > > > +{ > > > + int c =3D a - b; > > > + if (c =3D=3D 0) c =3D 0; > > > + else c =3D b - a; > > > + return c; > > > +} > > > +int minus4(int a, int b) > > > +{ > > > + int c; > > > + if (a =3D=3D b) c =3D 0; > > > + else > > > + c =3D b - a; > > > + return c; > > > +} > > > +int abs0(int a, int b) > > > +{ > > > + int c =3D a - b; > > > + if (c <=3D 0) c =3D b - a; > > > + return c; > > > +} > > > +int negabs(int a, int b) > > > +{ > > > + int c =3D a - b; > > > + if (c >=3D 0) c =3D b - a; > > > + return c; > > > +} > > > + > > > +/* The above should be optimized at phiopt1 except for negabs which = has to wait > > > + until phiopt2 as -abs is not acceptable in early phiopt. */ > > > +/* { dg-final { scan-tree-dump-times "if" 1 "phiopt1" } } */ > > > +/* { dg-final { scan-tree-dump-not "if" "phiopt2" } } */ > > > -- > > > 2.27.0 > > >