From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 28491 invoked by alias); 21 Oct 2011 12:19:49 -0000 Received: (qmail 28479 invoked by uid 22791); 21 Oct 2011 12:19:47 -0000 X-SWARE-Spam-Status: No, hits=-1.9 required=5.0 tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_ENVFROM_END_DIGIT,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW X-Spam-Check-By: sourceware.org Received: from mail-vw0-f47.google.com (HELO mail-vw0-f47.google.com) (209.85.212.47) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Fri, 21 Oct 2011 12:19:31 +0000 Received: by vwe42 with SMTP id 42so3187154vwe.20 for ; Fri, 21 Oct 2011 05:19:30 -0700 (PDT) MIME-Version: 1.0 Received: by 10.52.22.9 with SMTP id z9mr13916736vde.59.1319199570776; Fri, 21 Oct 2011 05:19:30 -0700 (PDT) Received: by 10.220.181.131 with HTTP; Fri, 21 Oct 2011 05:19:30 -0700 (PDT) In-Reply-To: References: <7d649d85-73fa-4c37-bd6e-d3527500196b@zmail14.collab.prod.int.phx2.redhat.com> Date: Fri, 21 Oct 2011 12:40:00 -0000 Message-ID: Subject: Re: [patch tree-optimization]: allow branch-cost optimization for truth-and/or on mode-expanded simple boolean-operands From: Kai Tietz To: Richard Guenther Cc: Kai Tietz , gcc-patches@gcc.gnu.org Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable X-IsSubscribed: yes Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2011-10/txt/msg01961.txt.bz2 2011/10/21 Richard Guenther : > On Thu, Oct 20, 2011 at 3:08 PM, Kai Tietz wrote: >> Hello, >> >> this patch re-enables the branch-cost optimization on simple boolean-typ= ed operands, which are casted to a wider integral type. =A0This happens due= casts from >> boolean-types are preserved, but FE might expands simple-expression to w= ider mode. >> >> I added two tests for already working branch-cost optimization for IA-ar= chitecture and >> two for explicit checking for boolean-type. >> >> ChangeLog >> >> 2011-10-20 =A0Kai Tietz =A0 >> >> =A0 =A0 =A0 =A0* fold-const.c (simple_operand_p_2): Handle integral >> =A0 =A0 =A0 =A0casts from boolean-operands. >> >> 2011-10-20 =A0Kai Tietz =A0 >> >> =A0 =A0 =A0 =A0* gcc.target/i386/branch-cost1.c: New test. >> =A0 =A0 =A0 =A0* gcc.target/i386/branch-cost2.c: New test. >> =A0 =A0 =A0 =A0* gcc.target/i386/branch-cost3.c: New test. >> =A0 =A0 =A0 =A0* gcc.target/i386/branch-cost4.c: New test. >> >> Bootstrapped and regression tested on x86_64-unknown-linux-gnu for all l= anguages including Ada and Obj-C++. =A0Ok for apply? >> >> Regards, >> Kai >> >> Index: gcc-head/gcc/testsuite/gcc.target/i386/branch-cost2.c >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> --- /dev/null >> +++ gcc-head/gcc/testsuite/gcc.target/i386/branch-cost2.c >> @@ -0,0 +1,16 @@ >> +/* { dg-do compile } */ >> +/* { dg-options "-O2 -fdump-tree-gimple -mbranch-cost=3D2" } */ >> + >> +extern int doo (void); >> + >> +int >> +foo (int a, int b) >> +{ >> + =A0if (a && b) >> + =A0 return doo (); >> + =A0return 0; >> +} >> + >> +/* { dg-final { scan-tree-dump-times "if " 1 "gimple" } } */ >> +/* { dg-final { scan-tree-dump-times " & " 1 "gimple" } } */ >> +/* { dg-final { cleanup-tree-dump "gimple" } } */ >> Index: gcc-head/gcc/testsuite/gcc.target/i386/branch-cost3.c >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> --- /dev/null >> +++ gcc-head/gcc/testsuite/gcc.target/i386/branch-cost3.c >> @@ -0,0 +1,16 @@ >> +/* { dg-do compile } */ >> +/* { dg-options "-O2 -fdump-tree-gimple -mbranch-cost=3D2" } */ >> + >> +extern int doo (void); >> + >> +int >> +foo (_Bool a, _Bool b) >> +{ >> + =A0if (a && b) >> + =A0 return doo (); >> + =A0return 0; >> +} >> + >> +/* { dg-final { scan-tree-dump-times "if " 1 "gimple" } } */ >> +/* { dg-final { scan-tree-dump-times " & " 1 "gimple" } } */ >> +/* { dg-final { cleanup-tree-dump "gimple" } } */ >> Index: gcc-head/gcc/testsuite/gcc.target/i386/branch-cost4.c >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> --- /dev/null >> +++ gcc-head/gcc/testsuite/gcc.target/i386/branch-cost4.c >> @@ -0,0 +1,16 @@ >> +/* { dg-do compile } */ >> +/* { dg-options "-O2 -fdump-tree-gimple -mbranch-cost=3D0" } */ >> + >> +extern int doo (void); >> + >> +int >> +foo (_Bool a, _Bool b) >> +{ >> + =A0if (a && b) >> + =A0 return doo (); >> + =A0return 0; >> +} >> + >> +/* { dg-final { scan-tree-dump-times "if " 2 "gimple" } } */ >> +/* { dg-final { scan-tree-dump-not " & " "gimple" } } */ >> +/* { dg-final { cleanup-tree-dump "gimple" } } */ >> Index: gcc-head/gcc/fold-const.c >> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D >> --- gcc-head.orig/gcc/fold-const.c >> +++ gcc-head/gcc/fold-const.c >> @@ -3706,6 +3706,19 @@ simple_operand_p_2 (tree exp) >> =A0 /* Strip any conversions that don't change the machine mode. =A0*/ >> =A0 STRIP_NOPS (exp); >> >> + =A0/* Handle integral widening casts from boolean-typed >> + =A0 =A0 expressions as simple. =A0This happens due casts from >> + =A0 =A0 boolean-types are preserved, but FE might expands >> + =A0 =A0 simple-expression to wider mode. =A0*/ >> + =A0if (INTEGRAL_TYPE_P (TREE_TYPE (exp)) >> + =A0 =A0 =A0&& CONVERT_EXPR_P (exp) >> + =A0 =A0 =A0&& TREE_CODE (TREE_TYPE (TREE_OPERAND (exp, 0))) >> + =A0 =A0 =A0 =A0=3D=3D BOOLEAN_TYPE) >> + =A0 =A0{ >> + =A0 =A0 =A0exp =3D TREE_OPERAND (exp, 0); >> + =A0 =A0 =A0STRIP_NOPS (exp); >> + =A0 =A0} >> + > > Huh, well. =A0I think the above is just too special and you instead should > replace the existing STRIP_NOPS by > > while (CONVERT_EXPR_P (exp)) > =A0exp =3D TREE_OPERAND (exp, 0); > > with a comment that conversions are considered simple. > > Ok with that change, if it bootstraps & tests ok. > > Richard. Ok, bootstrapped and regression-tested on x86_64-unknown-linux-gnu and applied to trunk with modifying as you suggested. One question I have about handling of TRUTH-binaries in general in fold-const.c. Why aren't we enforcing already here in fold_binary for those operations, that operands get boolean-type? I see here some advantages of C-AST folding. I've tested it and saw that even later in SSA-passes we get slightly better results on that. Regards, Kai