[Bug tree-optimization/43567] linear loop transform

[Bug tree-optimization/43567] linear loop transform

[tjvries at xs4all dot nl]

------- Comment #2 from tjvries at xs4all dot nl  2010-06-25 19:16 -------
Created an attachment (id=21007)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21007&action=view)
slightly more minimal testcase

reproduced on trunk revision 161295


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

[Bug tree-optimization/43567] linear loop transform

[tjvries at xs4all dot nl]

------- Comment #3 from tjvries at xs4all dot nl  2010-06-25 20:06 -------
Created an attachment (id=21008)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21008&action=view)
partially redoing the fix for bug 20612

The problem is in this piece of code in lambda_loopnest_gcc_loopnest:
...
      /* Create the new iv.  */

      standard_iv_increment_position (temp, &bsi, &insert_after);
      create_iv (newlowerbound,
                 build_int_cst (type, LL_STEP (newloop)),
                 ivvar, temp, &bsi, insert_after, &ivvar,
                 NULL);

      /* Unfortunately, the incremented ivvar that create_iv inserted may not
         dominate the block containing the exit condition.
         So we simply create our own incremented iv to use in the new exit
         test,  and let redundancy elimination sort it out.  */
      inc_stmt = gimple_build_assign_with_ops (PLUS_EXPR, SSA_NAME_VAR (ivvar),
                                               ivvar,
                                               build_int_cst (type, LL_STEP
(newloop)));

      ivvarinced = make_ssa_name (SSA_NAME_VAR (ivvar), inc_stmt);
      gimple_assign_set_lhs (inc_stmt, ivvarinced);
      bsi = gsi_for_stmt (exitcond);
      gsi_insert_before (&bsi, inc_stmt, GSI_SAME_STMT);
...

In case the loop header is copied (-ftree-ch), the second increment is in the
same block as the first increment:
...
  lnivtmp.6_13 = lnivtmp.6_20 + 1;
  lnivtmp.6_1 = lnivtmp.6_20 + 1;
  if (lletmp.8_21 >= lnivtmp.6_1)
...
indeed redundancy elimination sorts this out.

However, if the loop header is not copied:
...<bb 5>:
  i_11 = i_1 + 1;
  lnivtmp.2_19 = lnivtmp.2_5 + 1;

<bb 6>:
  # i_1 = PHI <0(2), i_11(5)>
  # lnivtmp.2_5 = PHI <0(2), lnivtmp.2_19(5)>
  lletmp.4_16 = n_4(D) + -1;
  lnivtmp.2_17 = lnivtmp.2_5 + 1;
  if (lletmp.4_16 >= lnivtmp.2_17)
...
the second increment is an extra increment, on top of the first one.

The patches fixes this, I'm not sure how minimal or efficient, but I did a
x86_64-unknown-linux-gnu bootstrap build and ran testsuites (gcc, objc,
gfortran, g++, libgomp, libstdc++, libjava, libmudflap, libffi) for r161295
and r161295+patch, with identical results.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

[Bug tree-optimization/43567] linear loop transform

[tjvries at xs4all dot nl]

------- Comment #4 from tjvries at xs4all dot nl  2010-07-11 06:10 -------
submitted patch for review:
http://gcc.gnu.org/ml/gcc-patches/2010-07/msg00879.html


-- 

tjvries at xs4all dot nl changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tjvries at xs4all dot nl


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

[Bug tree-optimization/43567] linear loop transform

[dnovillo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

--- Comment #10 from Diego Novillo <dnovillo at gcc dot gnu.org> 2011-02-02 17:48:04 UTC ---
Author: dnovillo
Date: Wed Feb  2 17:48:00 2011
New Revision: 169591

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=169591
Log:
Add testcase for PR43567.

2011-01-25  Sebastian Pop  <sebastian.pop@amd.com>

    PR tree-optimization/43567
    * gcc.dg/graphite/pr43567.c: New.

Added:
    branches/google/integration/gcc/testsuite/gcc.dg/graphite/pr43567.c
Modified:
    branches/google/integration/gcc/ChangeLog.graphite
    branches/google/integration/gcc/testsuite/ChangeLog

[Bug tree-optimization/43567] linear loop transform

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

Sebastian Pop <spop at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #9 from Sebastian Pop <spop at gcc dot gnu.org> 2011-01-25 21:28:23 UTC ---
Fixed.

[Bug tree-optimization/43567] linear loop transform

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

--- Comment #8 from Sebastian Pop <spop at gcc dot gnu.org> 2011-01-25 21:24:41 UTC ---
Author: spop
Date: Tue Jan 25 21:24:35 2011
New Revision: 169252

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=169252
Log:
Add testcase for PR43567.

2011-01-25  Sebastian Pop  <sebastian.pop@amd.com>

    PR tree-optimization/43567
    * gcc.dg/graphite/pr43567.c: New.

Added:
    trunk/gcc/testsuite/gcc.dg/graphite/pr43567.c
Modified:
    trunk/gcc/ChangeLog.graphite
    trunk/gcc/testsuite/ChangeLog

[Bug tree-optimization/43567] linear loop transform

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

Sebastian Pop <spop at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2011.01.18 20:57:45
         AssignedTo|unassigned at gcc dot       |spop at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #7 from Sebastian Pop <spop at gcc dot gnu.org> 2011-01-18 20:57:45 UTC ---
Fixed on the graphite branch.

[Bug tree-optimization/43567] linear loop transform

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

--- Comment #6 from Sebastian Pop <spop at gcc dot gnu.org> 2011-01-18 20:54:23 UTC ---
Author: spop
Date: Tue Jan 18 20:54:18 2011
New Revision: 168962

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=168962
Log:
Add testcase for PR43567.

2011-01-18  Sebastian Pop  <sebastian.pop@amd.com>

    PR tree-optimization/43567
    * gcc.dg/graphite/pr43567.c: New.

Added:
    branches/graphite/gcc/testsuite/gcc.dg/graphite/pr43567.c
Modified:
    branches/graphite/gcc/ChangeLog.graphite

[Bug tree-optimization/43567] linear loop transform

[spop at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43567

--- Comment #5 from Sebastian Pop <spop at gcc dot gnu.org> 2011-01-18 20:47:10 UTC ---
*** Bug 43568 has been marked as a duplicate of this bug. ***