From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-485073-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 81725 invoked by alias); 30 Apr 2015 11:28:47 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Received: (qmail 81653 invoked by uid 48); 30 Apr 2015 11:28:44 -0000
From: "alalaw01 at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/35226] Induction with multiplication are not vectorized
Date: Thu, 30 Apr 2015 11:28:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 6.0
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: enhancement
X-Bugzilla-Who: alalaw01 at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution:
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_status cf_reconfirmed_on cc version short_desc everconfirmed
Message-ID: <bug-35226-4-LyLoGaqOnY@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-35226-4@http.gcc.gnu.org/bugzilla/>
References: <bug-35226-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2015-04/txt/msg02625.txt.bz2

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=35226

alalaw01 at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-04-30
                 CC|                            |alalaw01 at gcc dot gnu.org
            Version|4.3.0                       |6.0
            Summary|Reduction and induction     |Induction with
                   |with multiplication are not |multiplication are not
                   |vectorized                  |vectorized
     Ever confirmed|0                           |1

--- Comment #1 from alalaw01 at gcc dot gnu.org ---
Multiplication reductions are supported, certainly in gcc 4.9, I think longer.
However, the following induction does not vectorize on gcc 6 development branch
(x86_64, -O3, with or without -mavx or -msse2):

int a[24];

int
main (int argc, char **argv)
{
  int p = 1;
  for (int i = 0; i < 24; i++, p*=2)
    a[i] *= p;
}

-fdump-tree-vect-details suggests the multiplication is recognized as a
reduction but not as an induction:

test_induc.c:7:3: note: Analyze phi: p_13 = PHI <p_9(4), 1(2)>

test_induc.c:7:3: note: reduction used in loop.
test_induc.c:7:3: note: Unknown def-use cycle pattern.
test_induc.c:7:3: note: === vect_pattern_recog ===
test_induc.c:7:3: note: vect_is_simple_use: operand _5
test_induc.c:7:3: note: def_stmt: _5 = a[i_14];
test_induc.c:7:3: note: type of def: 3.
test_induc.c:7:3: note: vect_is_simple_use: operand p_13
test_induc.c:7:3: note: def_stmt: p_13 = PHI <p_9(4), 1(2)>
test_induc.c:7:3: note: Unsupported pattern.
...
test_induc.c:7:3: note: def_stmt: p_13 = PHI <p_9(4), 1(2)>
test_induc.c:7:3: note: Unsupported pattern.
test_induc.c:7:3: note: not vectorized: unsupported use in stmt.
test_induc.c:7:3: note: unexpected pattern.
test_induc.c:4:1: note: vectorized 0 loops in function.
...

  <bb 3>:
  # p_13 = PHI <p_9(4), 1(2)>
...
  p_9 = p_13 * 2;