[Bug libstdc++/54869] New: ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[Bug libstdc++/54869] New: ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

             Bug #: 54869
           Summary: ext/random/simd_fast_mersenne_twister_engine/cons/defa
                    ult.cc FAILs
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: ro@gcc.gnu.org
                CC: drepper.fsp@gmail.com, ebotcazou@gcc.gnu.org,
                    paolo.carlini@oracle.com
              Host: sparc*-*-solaris2*
            Target: sparc*-*-solaris2*
             Build: sparc*-*-solaris2*


The new ext/random/simd_fast_mersenne_twister_engine/cons/default.cc testcase
FAILs on Solaris/SPARC (both 32 and 64-bit):

1318548553 vs 1318548553
1985957974 vs 1985957974
1367744196 vs 1367744196
3463392791 vs 3463392791
2780736231 vs 2780736231
3894488561 vs 3894488561
3157036262 vs 3157036262
3491812767 vs 3491812767
1724574180 vs 1724574180
3645035493 vs 3645035493
15030273820155882740 vs 15510024334182072935
7424847061300251637 vs 5793753331747412752
12274779367645387890 vs 16198353238554625740
2232540708403540888 vs 2233208824926016498
17502480069811850555 vs 3566091399820823780
4216703503856069688 vs 16608268514591292798
166208635478247899 vs 10684941689666043359
4604613833690380892 vs 12463424292910456802
5930322781620806096 vs 5902567440240131366
8093226496393420735 vs 7228030834036501150
Assertion failed: success, file
/vol/gcc/src/hg/trunk/local/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/default.cc,
line 44, function run_test

  Rainer

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-09 09:49:08 UTC ---
Guys, I don't have Sparc-Solaris machines and didn't design the testcase. To be
honest, I didn't even realize so far that this simd_* variant was meant for
other targets besides x86_64. Ulrich, please help a bit the target
maintainers...

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|sparc*-*-solaris2*          |sparc*-*-solaris2*
                   |                            |powerpc*-*-*
                   |                            |hppa-unknown-linux-gnu
                   |                            |s390-ibm-linux-gnu
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-10-09
               Host|sparc*-*-solaris2*          |sparc*-*-solaris2*
                   |                            |powerpc*-*-*
     Ever Confirmed|0                           |1
              Build|sparc*-*-solaris2*          |sparc*-*-solaris2*
                   |                            |powerpc*-*-*

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-10-09 10:14:45 UTC ---
Also fails for at least powerpc*-*-* (see
http://gcc.gnu.org/ml/gcc-testresults/2012-10/msg00888.html ),
hppa-unknown-linux-gnu (
http://gcc.gnu.org/ml/gcc-testresults/2012-10/msg00525.html ), and
s390-ibm-linux-gnu (
http://gcc.gnu.org/ml/gcc-testresults/2012-10/msg00602.html ). Endianness
issue?

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #3 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-09 10:20:45 UTC ---
Can also be something even more trivial, because I bet we didn't exercise much
the codepath for __SSE2__ undefined before committing the code.

Actually - and this is a question for Ulrich - is the code actually useful when
__SSE2__ is undefined?

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[drepper.fsp at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #4 from Ulrich Drepper <drepper.fsp at gmail dot com> 2012-10-09 11:23:41 UTC ---
(In reply to comment #0)
> The new ext/random/simd_fast_mersenne_twister_engine/cons/default.cc testcase
> FAILs on Solaris/SPARC (both 32 and 64-bit):

That's expected.  I mentioned when I posted the patches that the implementation
is for little endian machines.  I don't have access to any big endian machines
and therefore didn't even try to make it work.

It might be sufficient, at end of _M_gen_rand, to swap the order of the four
32-bit words in a 128-bit word.  I never tested this, someone else will have to
do this.

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #5 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-09 11:28:45 UTC ---
Ok, thanks, I missed that remark of yours. Well, at present nothing prevents
people with big endian machines to try to use the code, and we unconditionally
run the testsuite. That's not Ok.

Let's keep the PR open for a while in case somebody wants to try your
suggestion, otherwise, as far as I'm concerned, for 4.8.0, I will just disable
the code and its tests on big endian.

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #6 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-10-09 11:35:19 UTC ---
> ... I don't have access to any big endian machines
> and therefore didn't even try to make it work.

AFAICT there are big endian machines in the CompileFarm (see
http://gcc.gnu.org/wiki/CompileFarm ) and any maintainer has access to it.

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #7 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2012-10-09 11:39:17 UTC ---
As  Nathan Froyd said at
http://gcc.gnu.org/ml/gcc-patches/2012-10/msg00772.html

"Please try to consider what's best for all the people who use GCC, not just
the cases you happen to be working with every day."

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #8 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-09 11:44:09 UTC ---
Boy, nobody cites me like that, at least, not while I'm still alive.

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #9 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> 2012-10-09 14:50:31 UTC ---
Author: paolo
Date: Tue Oct  9 14:50:19 2012
New Revision: 192256

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=192256
Log:
2012-10-09  Paolo Carlini  <paolo.carlini@oracle.com>

    PR libstdc++/54869
    * include/ext/random (simd_fast_mersenne_twister_engine): Provide
    only for little endian targets.
    * include/ext/random.tcc: Likewise.
    * config/cpu/i486/opt/ext/opt_random.h: Likewise.
    * testsuite/lib/libstdc++.exp (check_v3_target_little_endian): Add.
    * testsuite/lib/dg-options.exp (dg-require-little-endian): Add.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    operators/equal.cc: Use the latter.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    operators/serialize.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    operators/inequal.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    cons/copy.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    cons/seed1.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    cons/seed2.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    cons/default.cc: Likewise.
    * testsuite/ext/random/simd_fast_mersenne_twister_engine/
    cons/seed_seq.cc: Likewise.

Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/config/cpu/i486/opt/ext/opt_random.h
    trunk/libstdc++-v3/include/ext/random
    trunk/libstdc++-v3/include/ext/random.tcc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/copy.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/default.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/seed1.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/seed2.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/cons/seed_seq.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/operators/equal.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/operators/inequal.cc
   
trunk/libstdc++-v3/testsuite/ext/random/simd_fast_mersenne_twister_engine/operators/serialize.cc
    trunk/libstdc++-v3/testsuite/lib/dg-options.exp
    trunk/libstdc++-v3/testsuite/lib/libstdc++.exp

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

Paolo Carlini <paolo.carlini at oracle dot com> changed:

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

--- Comment #10 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-09 14:53:14 UTC ---
If the problem is endianness should be fixed now. Providing an implementation
for big endian targets is enhancement.

[Bug libstdc++/54869] ext/random/simd_fast_mersenne_twister_engine/cons/default.cc FAILs

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54869

--- Comment #11 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2012-10-09 15:05:31 UTC ---
> --- Comment #4 from Ulrich Drepper <drepper.fsp at gmail dot com>
> 2012-10-09 11:23:41 UTC ---
> (In reply to comment #0)
>> The new ext/random/simd_fast_mersenne_twister_engine/cons/default.cc testcase
>> FAILs on Solaris/SPARC (both 32 and 64-bit):
>
> That's expected.  I mentioned when I posted the patches that the implementation
> is for little endian machines.  I don't have access to any big endian machines
> and therefore didn't even try to make it work.
>
> It might be sufficient, at end of _M_gen_rand, to swap the order of the four
> 32-bit words in a 128-bit word.  I never tested this, someone else will have to
> do this.

I'm not familiar with the code at all.  Could you provide a tentative
patch for me to try?

Thanks.
        Rainer