[Bug tree-optimization/17165] New: [lno] pass by reference inhibits vectorization

[Bug tree-optimization/17165] New: [lno] pass by reference inhibits vectorization

[stefaandr at hotmail dot com]

The following code doesn't get vectorized:

typedef int T; T a[4], c[4];

T & input(int x) { return a[x]; };
T & result(int x) { return c[x]; };

int main() {
        for (int x=0; x<4; ++x)
                result(x) = input(x);
};

which is just c[x] = a[x], but with wrapper functions.  

It gives:
Access function of ptr: {(T &)&a, +, 4}_1
loop at test23.cpp:4: not vectorized: non constant init/step for pointer access.
loop at test23.cpp:4: bad data references.

The step is constant, so I guess &a is considered "non-constant".  And why does
it say &a (incorrect in my opinion), instead of &a[0]?

Reading specs from
/esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/specs
Configured with: /esat/alexandria1/stefaan/src/gcc/lno/configure
--prefix=/esat/olympia/install --program-suffix=-lno --enable-languages=c,c++ :
(reconfigured) /esat/alexandria1/stefaan/src/gcc/lno/configure
--prefix=/esat/olympia/install --program-suffix=-lno
--with-gcc-version-trigger=/esat/alexandria1/stefaan/src/gcc/lno/gcc/version.c
--enable-languages=c,c++ --no-create --no-recursion : (reconfigured)
/esat/alexandria1/stefaan/src/gcc/lno/configure --prefix=/esat/olympia/install
--program-suffix=-lno
--with-gcc-version-trigger=/esat/alexandria1/stefaan/src/gcc/lno/gcc/version.c
--enable-languages=c,c++ --no-create --no-recursion
Thread model: posix
gcc version 3.5-tree-ssa-lno 20040824 (merged 20040718)
 /esat/olympia/install/libexec/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/cc1plus
-quiet -v -D_GNU_SOURCE test23.cpp -quiet -dumpbase test23.cpp -march=pentium4
-auxbase-strip test23.o -O9 -version -ftree-vectorize -fdump-tree-vect-stats
-fdump-tree-vect-details -funroll-all-loops -finline-limit=1000000
-finline-functions -o /tmp/ccsL4r8X.s
ignoring nonexistent directory "/usr/local/include"
ignoring nonexistent directory
"/esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/../../../../i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/../../../../include/c++/3.5-tree-ssa-lno
 /esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/../../../../include/c++/3.5-tree-ssa-lno/i686-pc-linux-gnu
 /esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/../../../../include/c++/3.5-tree-ssa-lno/backward
 /esat/olympia/install/include
 /esat/olympia/install/lib/gcc/i686-pc-linux-gnu/3.5-tree-ssa-lno/include
 /usr/include
End of search list.
GNU C++ version 3.5-tree-ssa-lno 20040824 (merged 20040718) (i686-pc-linux-gnu)
        compiled by GNU C version 3.4.1.
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
 as -V -Qy -o test23.o /tmp/ccsL4r8X.s
GNU assembler version 2.15 (i686-pc-linux-gnu) using BFD version 2.15

-- 
           Summary: [lno] pass by reference inhibits vectorization
           Product: gcc
           Version: lno
            Status: UNCONFIRMED
          Severity: minor
          Priority: P3
         Component: tree-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: stefaandr at hotmail dot com
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-24 15:14 -------
Confirmed on the mainline (since the vectorizer was merged):
Access function of ptr: {(T &)&a, +, 4}_1
step: 4,  init: (T &)&a
loop at pr17165.cc:4: not vectorized: non constant init/step for pointer access.
loop at pr17165.cc:4: bad data references.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|minor                       |enhancement
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |missed-optimization
   Last reconfirmed|0000-00-00 00:00:00         |2004-08-24 15:14:45
               date|                            |
            Summary|[lno] pass by reference     |pass by reference inhibits
                   |inhibits vectorization      |vectorization
            Version|lno                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-31 23:43 -------
Well now the problem is slightly difference:
Access function of ptr: T.5_11
loop at pr17165.cc:4: not vectorized: pointer access is not simple.
loop at pr17165.cc:4: bad data references.

  T.4_7 = &c[x_26];
  T.5_11 = &a[x_26];
  T.6_13 = *T.5_11;
  *T.4_7 = T.6_13;

And this would be related to PR 14814 and some other PRs too.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |14814


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-01 00:47 -------
Here is another example without inlining and references which show that this problem is really the 
same as the problem in PR 14814:
typedef int T; T a[4], c[4];
int main() {
        for (int x=0; x<4; ++x)
         {
               T *i = &a[x];
               T *r = &c[x];
               *r = *i;
         }
}

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-03 09:34 -------
Mine, I have a patch/working on getting this fixed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-11 01:30 -------
Basically fold should convert (T<D1029> *) (x.0<D1040>_5 * 4) + &a<D1030> into 
&a<D1030>[x.0<D1040>_5] and then we can do the C case but the C++ case is easier and gets done 
when I finish up my tree combiner see 15459.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |15459


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

-- 
Bug 17165 depends on bug 14814, which changed state.

Bug 14814 Summary: no folding back to ARRAY_REF
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14814

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-09-15 04:55 -------
While 14814 is fixed, this one is no.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[dorit at il dot ibm dot com]

------- Additional Comments From dorit at il dot ibm dot com  2004-11-07 14:50 -------
(1) The first (original) testcase does not get vectorized because 
analyze_scalar_evolution is unable to analyze the access-function of the 
pointers in the loop. For the pointer D.1780_5 it returns "D.1780_5" instead 
of "(c,+,4B)_1" (and similarly for D.1781_9: we have "D.1781_9", we 
want "(a,+,4B)_1"):

Access function of ptr: D.1781_9
loop at pr17165.cc:17: not vectorized: pointer access is not simple.
loop at pr17165.cc:17: not vectorized: unhandled data ref: D.1782_11 = *D.1781_9
loop at pr17165.cc:17: bad data references.

Don't know if we can expect analyze_scalar_evolution to return a meaningful 
access_function when the code is represented this way:

  # ivtmp.7_22 = PHI <4(0), ivtmp.7_10(3)>;
  # x_24 = PHI <0(0), x_12(3)>;
<L0>:;
  D.1780_5 = &c[x_24];
  D.1781_9 = &a[x_24];
  D.1782_11 = *D.1781_9;
  *D.1780_5 = D.1782_11;
  x_12 = x_24 + 1;
  ivtmp.7_10 = ivtmp.7_22 - 1;
  if (ivtmp.7_10 != 0) goto <L7>; else goto <L4>;

I don't see which of the tree combiner PRs listed in PR15459 would help us 
here. ?

(2) The second testcase does not get vectorized because the alias information 
tells the vectorizer that the pointers 'r' and 'i' may alias:

loop at pr17165.c:6: not vectorized: can't determine dependence between: *i_6 
and *r_10

This loop will be vectorized once we have runtime alias/dependence testing in 
the vectorizer, or when points-to analysis is able to tell that 'a' and 'c' 
don't alias:

  # ivtmp.2_9 = PHI <4(0), ivtmp.2_8(3)>;
  # c_22 = PHI <c_15(0), c_18(3)>;
  # a_23 = PHI <a_16(0), a_17(3)>;
  # x_24 = PHI <0(0), x_12(3)>;
<L0>:;
  x.0_3 = (unsigned int) x_24;
  D.1328_4 = x.0_3 * 4;
  D.1329_5 = (T *) D.1328_4;
  i_6 = D.1329_5 + &a;
  r_10 = D.1329_5 + &c;
  #   VUSE <a_23>;
  #   VUSE <c_22>;
  D.1330_11 = *i_6;
  #   a_17 = V_MAY_DEF <a_23>;
  #   c_18 = V_MAY_DEF <c_22>;
  *r_10 = D.1330_11;
  x_12 = x_24 + 1;
  ivtmp.2_8 = ivtmp.2_9 - 1;
  if (ivtmp.2_8 != 0) goto <L5>; else goto <L2>;

<L5>:;
  goto <bb 1> (<L0>);


========> Here's the alias dump:


Referenced variables in main1: 10

Variable: x, UID 0, int

Variable: x.0, UID 1, unsigned int

Variable: D.1328, UID 2, unsigned int

Variable: D.1329, UID 3, T *

Variable: i, UID 4, T *, type memory tag: TMT.1

Variable: a, UID 5, T[3], is an alias tag, is addressable, is global, call 
clobbered, default def: a_16

Variable: r, UID 6, T *, type memory tag: TMT.1

Variable: c, UID 7, T[3], is an alias tag, is addressable, is global, call 
clobbered, default def: c_15

Variable: D.1330, UID 8, int

Variable: TMT.1, UID 9, T, is addressable, is global, call clobbered, may 
aliases: { a c }



Pointed-to sets for pointers in main1

D.1329_5, points-to anything
i_6, is dereferenced, points-to anything, points-to vars: { a }
r_10, is dereferenced, points-to anything, points-to vars: { c }


Flow-insensitive alias information for main1

Aliased symbols

a, UID 5, T[3], is an alias tag, is addressable, is global, call clobbered, 
default def: a_16
c, UID 7, T[3], is an alias tag, is addressable, is global, call clobbered, 
default def: c_15
TMT.1, UID 9, T, is addressable, is global, call clobbered, may aliases: { a c }

Dereferenced pointers

i, UID 4, T *, type memory tag: TMT.1
r, UID 6, T *, type memory tag: TMT.1

Type memory tags

TMT.1, UID 9, T, is addressable, is global, call clobbered, may aliases: { a c }


Flow-sensitive alias information for main1

SSA_NAME pointers



Name memory tags



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-17 18:10 -------
Fixed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[jsm28 at gcc dot gnu dot org]

-- 
Bug 17165 depends on bug 14814, which changed state.

Bug 14814 Summary: no folding back to ARRAY_REF
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14814

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu dot org]

-- 
Bug 17165 depends on bug 14814, which changed state.

Bug 14814 Summary: no folding back to ARRAY_REF
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14814

           What    |Old Value                   |New Value
----------------------------------------------------------------------------
             Status|REOPENED                    |RESOLVED
         Resolution|                            |FIXED

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165

[Bug tree-optimization/17165] pass by reference inhibits vectorization

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=17165
Bug 17165 depends on bug 15459, which changed state.

Bug 15459 Summary: [meta-bug] there should be a tree combiner like the rtl one
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=15459

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