[mvapich-discuss] rpl_malloc error

Jonathan Perkins perkinjo at cse.ohio-state.edu
Thu Oct 20 12:27:20 EDT 2011 

The osu_benchmarks are configured and compiled after the library is
created in the lib directory.  The osu_benchmarks use -I../include and
-L../lib to find the necessary files of the yet to be installed
mvapich2 library.

Can you try a fresh directory extracted from the tarball and then the
following commands?

export ac_cv_func_malloc_0_nonnull=yes ac_cv_func_realloc_0_nonnull=yes
./configure <your configure arguments>
make
make install

The reason I'm choosing to export at the top level is because I want
to make sure it is in the environment when the make step is run since
that is when the osu benchmarks are actually configured and built.

On Thu, Oct 20, 2011 at 11:34 AM, Dan Kokron <daniel.kokron at nasa.gov> wrote:
> I added the following at the top of my osu_benchmarks/configure
>
> ac_cv_func_malloc_0_nonnull=yes
> export ac_cv_func_malloc_0_nonnull
> ac_cv_func_realloc_0_nonnull=yes
> export ac_cv_func_realloc_0_nonnull
>
> then at the top level
> gmake clean
> configure ....
>
> malloc is still getting redefined in the osu_benchmarks/Makefile
>
> grep -n rpl_malloc osu_benchmarks/Makefile
> 169:DEFS = -DPACKAGE_NAME=\"OSU-Micro-Benchmarks\" -DPACKAGE_TARNAME=
> \"osu-micro-benchmarks\" -DPACKAGE_VERSION=\"3.4\" -DPACKAGE_STRING=
> \"OSU-Micro-Benchmarks\ 3.4\" -DPACKAGE_BUGREPORT=
> \"mvapich-discuss at cse.ohio-state.edu\" -DPACKAGE_URL=\"\" -DPACKAGE=
> \"osu-micro-benchmarks\" -DVERSION=\"3.4\" -DSTDC_HEADERS=1
> -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1
> -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
> -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1
> -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_SYS_TIME_H=1
> -DHAVE_UNISTD_H=1 -DHAVE_STDLIB_H=1 -DHAVE_MALLOC=0 -Dmalloc=rpl_malloc
> -DHAVE_GETPAGESIZE=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_MEMSET=1 -DHAVE_SQRT=1
> -DFIELD_WIDTH=20 -DFLOAT_PRECISION=2
>
> Here is some output from the config.log in the benchmark directory.  Why
> is configure looking for -lmpich at this stage?
>
> configure:2782: checking whether the C compiler works
> configure:2804: ../bin/mpicc  -I ../include -L ../lib -Wl,-rpath=/u/dkokron/play/mvapich2-1.7/install/12.1.0.233/lib conftest.c  >&5
> ld: cannot find -lmpich
> configure:2808: $? = 1
> configure:2846: result: no
> configure: failed program was:
> | /* confdefs.h */
> | #define PACKAGE_NAME "OSU-Micro-Benchmarks"
> | #define PACKAGE_TARNAME "osu-micro-benchmarks"
> | #define PACKAGE_VERSION "3.4"
> | #define PACKAGE_STRING "OSU-Micro-Benchmarks 3.4"
> | #define PACKAGE_BUGREPORT "mvapich-discuss at cse.ohio-state.edu"
> | #define PACKAGE_URL ""
> | #define PACKAGE "osu-micro-benchmarks"
> | #define VERSION "3.4"
> | /* end confdefs.h.  */
> |
> | int
> | main ()
> | {
> |
> |   ;
> |   return 0;
> | }
> configure:2851: error: in `/u/dkokron/play/mvapich2-1.7/osu_benchmarks':
> configure:2853: error: C compiler cannot create executables
>
> On Wed, 2011-10-19 at 17:47 -0500, Jonathan Perkins wrote:
>> Thanks, this is actually pretty useful.  It still doesn't explain why
>> I've been unable to reproduce this issue in my environment but I can
>> at least create a patch to resolve this issue.  At this point I'm
>> thinking about just removing AC_FUNC_MALLOC.
>>
>> Can you try the following?
>>
>> export ac_cv_func_malloc_0_nonnull=yes ac_cv_func_realloc_0_nonnull=yes
>> ./configure <your configure arguments>
>> make
>> make install
>>
>> This should work around the issue while I work on a more permanent
>> solution to this.
>>
>> On Wed, Oct 19, 2011 at 6:29 PM, Dan Kokron <daniel.kokron at nasa.gov> wrote:
>> > I was able to build the 1.7 release under the intel-11.0.083 compiler
>> > without any issues.
>> >
>> > I did a 'module purge' before loading the 12.1 environment.  My env is
>> > pretty barren after a purge.
>> >
>> >
>> > I found this reference to the same issue.
>> > http://www.cygwin.com/ml/automake/2003-05/msg00023.html
>> >
>> > I did a search for AC_FUNC_MALLOC in all of the configure.ac under the
>> > 1.7 release.  The only place it's present is
>> > osu_benchmarks/configure.ac:23:AC_FUNC_MALLOC
>> >
>> > Looking through the config.log for osu_benchmarks (see attached)
>> >
>> > grep -n rpl_malloc config.log
>> > 397:DEFS='-DPACKAGE_NAME=\"OSU-Micro-Benchmarks\" -DPACKAGE_TARNAME=
>> > \"osu-micro-benchmarks\" -DPACKAGE_VERSION=\"3.4\" -DPACKAGE_STRING=
>> > \"OSU-Micro-Benchmarks\ 3.4\" -DPACKAGE_BUGREPORT=
>> > \"mvapich-discuss at cse.ohio-state.edu\" -DPACKAGE_URL=\"\" -DPACKAGE=
>> > \"osu-micro-benchmarks\" -DVERSION=\"3.4\" -DSTDC_HEADERS=1
>> > -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1
>> > -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
>> > -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1
>> > -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_SYS_TIME_H=1
>> > -DHAVE_UNISTD_H=1 -DHAVE_STDLIB_H=1 -DHAVE_MALLOC=0 -Dmalloc=rpl_malloc
>> > -DHAVE_GETPAGESIZE=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_MEMSET=1 -DHAVE_SQRT=1
>> > -DFIELD_WIDTH=20 -DFLOAT_PRECISION=2'
>> > 498:#define malloc rpl_malloc
>> >
>> > hope this helps.
>> >
>> > Dan
>> >
>> > On Wed, 2011-10-19 at 16:36 -0500, Jonathan Perkins wrote:
>> >> Hi Dan.  Does this problem occur with all versions of the intel
>> >> compiler or only with intel 12?  I haven't been able to reproduce this
>> >> issue personally but I think it may be related to having both a
>> >> previous version of the intel compilers in my path in addition to
>> >> intel 12.  Does this case apply to you?
>> >>
>> >> We're still looking into this but it may be helpful if you can forward
>> >> your config.log file from the osu_benchmarks directory as well.
>> >> Thanks.
>> >>
>> >> On Wed, Oct 19, 2011 at 1:50 PM, Dan Kokron <daniel.kokron at nasa.gov> wrote:
>> >> > I am seeing the same 'rpl_malloc' error using mvapich2-1.7
>> >> > (http://mvapich.cse.ohio-state.edu/download/mvapich2/mvapich2-1.7.tgz)
>> >> >
>> >> > under
>> >> > Intel(R) 64, Version 12.1.0.233 Build 20110811
>> >> >
>> >> > My config command is
>> >> > ./configure CC=icc CXX=icpc F77=ifort FC=ifort CFLAGS=-fpic
>> >> > CXXFLAGS=-fpic FFLAGS=-fpic FCFLAGS=-fpic
>> >> > --prefix=/u/dkokron/play/mvapich2-1.7/install/12.1.0.233 --enable-f77
>> >> > --enable-fc --enable-cxx --enable-romio --with-file-system=lustre
>> >> > --enable-threads=default --with-device=ch3:mrail --with-rdma=gen2
>> >> > --with-hwloc -disable-multi-aliases -enable-xrc=yes -enable-hybrid
>> >> >
>> >> >
>> >> > ../bin/mpicc -DPACKAGE_NAME=\"OSU-Micro-Benchmarks\" -DPACKAGE_TARNAME=
>> >> > \"osu-micro-benchmarks\" -DPACKAGE_VERSION=\"3.4\" -DPACKAGE_STRING=
>> >> > \"OSU-Micro-Benchmarks\ 3.4\" -DPACKAGE_BUGREPORT=
>> >> > \"mvapich-discuss at cse.ohio-state.edu\" -DPACKAGE_URL=\"\" -DPACKAGE=
>> >> > \"osu-micro-benchmarks\" -DVERSION=\"3.4\" -DSTDC_HEADERS=1
>> >> > -DHAVE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1
>> >> > -DHAVE_STRING_H=1 -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1
>> >> > -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNISTD_H=1
>> >> > -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_SYS_TIME_H=1
>> >> > -DHAVE_UNISTD_H=1 -DHAVE_STDLIB_H=1 -DHAVE_MALLOC=0 -Dmalloc=rpl_malloc
>> >> > -DHAVE_GETPAGESIZE=1 -DHAVE_GETTIMEOFDAY=1 -DHAVE_MEMSET=1 -DHAVE_SQRT=1
>> >> > -DFIELD_WIDTH=20 -DFLOAT_PRECISION=2 -I.   -I ../include  -g -O2 -c
>> >> > osu_alltoall.c
>> >> > ../bin/mpicc  -g -O2  -L ../lib
>> >> > -Wl,-rpath=/u/dkokron/play/mvapich2-1.7/install/12.1.0.233/lib -o
>> >> > osu_alltoall osu_alltoall.o
>> >> > osu_alltoall.o: In function `main':
>> >> > /home1/dkokron/play/mvapich2-1.7/osu_benchmarks/osu_alltoall.c:82:
>> >> > undefined reference to `rpl_malloc'
>> >> > /home1/dkokron/play/mvapich2-1.7/osu_benchmarks/osu_alltoall.c:91:
>> >> > undefined reference to `rpl_malloc'
>> >> >
>> >> > Just to be sure.
>> >> >
>> >> > bin/mpicc -V
>> >> > Intel(R) C Intel(R) 64 Compiler XE for applications running on Intel(R)
>> >> > 64, Version 12.1.0.233 Build 20110811
>> >> > Copyright (C) 1985-2011 Intel Corporation.  All rights reserved.
>> >> >
>> >> >
>> >> > On Fri, 2011-10-14 at 14:10 -0500, Jonathan Perkins wrote:
>> >> >> It looks like you may be hitting a different error for 1.7rc2 compared
>> >> >> to 1.6.  Can you send me the config.log from both the top level
>> >> >> directory and the osu_benchmarks sub directory.
>> >> >>
>> >> >> Also can you provide the exact steps needed to reproduce this?  I've
>> >> >> quickly tried both 1.6 and 1.7rc2 with the intel12 compilers and I was
>> >> >> unable to see this problem.
>> >> >>
>> >> >> On Fri, Oct 14, 2011 at 1:05 PM, Albino A. Aveleda <bino at coc.ufrj.br> wrote:
>> >> >> > Hi All,
>> >> >> >
>> >> >> > I am using SLES11SP1 and below you have more details about the error.
>> >> >> >
>> >> >> > Intel(R) C Intel(R) 64 Compiler XE for applications running on Intel(R) 64, Version 12.1.0.233 Build 20110811
>> >> >> > Copyright (C) 1985-2011 Intel Corporation.  All rights reserved.
>> >> >> > configure:2761: $? = 0
>> >> >> > configure:2750: ../bin/mpicc -qversion >&5
>> >> >> > icc: command line warning #10006: ignoring unknown option '-qversion'
>> >> >> > /usr/lib64/gcc/x86_64-suse-linux/4.3/../../../../lib64/crt1.o: In function `_start':
>> >> >> > /usr/src/packages/BUILD/glibc-2.11.1/csu/../sysdeps/x86_64/elf/start.S:109: undefined reference to `main'
>> >> >> > configure:2761: $? = 1
>> >> >> > configure:2781: checking whether the C compiler works
>> >> >> > configure:2803: ../bin/mpicc  -I ../include -L ../lib -Wl,-rpath=/sw/mpi/mvapich2/1.7rc2-intel/lib conftest.c  >&5
>> >> >> > configure:2807: $? = 0
>> >> >> > configure:2855: result: yes
>> >> >> > configure:2858: checking for C compiler default output file name
>> >> >> > configure:2860: result: a.out
>> >> >> > configure:2866: checking for suffix of executables
>> >> >> > configure:2873: ../bin/mpicc -o conftest  -I ../include -L ../lib -Wl,-rpath=/sw/mpi/mvapich2/1.7rc2-intel/lib conftest.c  >&5
>> >> >> > configure:2877: $? = 0
>> >> >> > configure:2899: result:
>> >> >> > configure:2921: checking whether we are cross compiling
>> >> >> > configure:2929: ../bin/mpicc -o conftest  -I ../include -L ../lib -Wl,-rpath=/sw/mpi/mvapich2/1.7rc2-intel/lib conftest.c  >&5
>> >> >> > configure:2933: $? = 0
>> >> >> > configure:2940: ./conftest
>> >> >> > /sw/src/mvapich2/mvapich2-1.7rc2/osu_benchmarks/configure: line 2942: 31101 Segmentation fault      ./conftest$ac_cv_exeext
>> >> >> > configure:2944: $? = 139
>> >> >> > configure:2951: error: in `/sw/src/mvapich2/mvapich2-1.7rc2/osu_benchmarks':
>> >> >> > configure:2953: error: cannot run C compiled programs.
>> >> >> >
>> >> >> > Regards,
>> >> >> >
>> >> >> > ----- Original Message -----
>> >> >> > From: "Albino A. Aveleda" <bino at coc.ufrj.br>
>> >> >> > To: mvapich-discuss at cse.ohio-state.edu
>> >> >> > Sent: Friday, October 14, 2011 9:30:18 AM
>> >> >> > Subject: [mvapich-discuss] rpl_malloc error
>> >> >> >
>> >> >> > Hi All,
>> >> >> >
>> >> >> > I compiled the mvapich2-1.6 with gcc witout problem, but
>> >> >> > when I try to compile with intel compiler version 12 I
>> >> >> > receive this error.
>> >> >> >
>> >> >> > make[2]: Entering directory `/sw/src/mvapich2/mvapich2-1.6/osu_benchmarks'
>> >> >> > ../bin/mpicc  -g -O2  -L ../lib -Wl,-rpath=/sw/mpi/mvapich2/1.6-intel/lib -o osu_alltoall osu_alltoall.o
>> >> >> > osu_alltoall.o: In function `main':
>> >> >> > /sw/src/mvapich2/mvapich2-1.6/osu_benchmarks/osu_alltoall.c:69: undefined reference to `rpl_malloc'
>> >> >> > /sw/src/mvapich2/mvapich2-1.6/osu_benchmarks/osu_alltoall.c:78: undefined reference to `rpl_malloc'
>> >> >> > make[2]: *** [osu_alltoall] Error 1
>> >> >> > make[2]: Leaving directory `/sw/src/mvapich2/mvapich2-1.6/osu_benchmarks'
>> >> >> > make[1]: *** [all-postamble] Error 2
>> >> >> > make[1]: Leaving directory `/sw/src/mvapich2/mvapich2-1.6'
>> >> >> > make: *** [all-redirect] Error 2
>> >> >> >
>> >> >> >
>> >> >> > How do I fix this?
>> >> >> >
>> >> >> > Regards,
>> >> >> > Bibo
>> >> >> > _______________________________________________
>> >> >> > mvapich-discuss mailing list
>> >> >> > mvapich-discuss at cse.ohio-state.edu
>> >> >> > http://mail.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>> >> >> >
>> >> >> > --
>> >> >> >
>> >> >> > __________________________________________________
>> >> >> > Albino A. Aveleda                 bino at coc.ufrj.br
>> >> >> > System Engineer                   +55 21 2562-8080
>> >> >> > PEC-COPPE/UFRJ                    +55 21 2562-8465
>> >> >> > Federal University of Rio de Janeiro (UFRJ)
>> >> >> > _______________________________________________
>> >> >> > mvapich-discuss mailing list
>> >> >> > mvapich-discuss at cse.ohio-state.edu
>> >> >> > http://mail.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>> >> >> >
>> >> >> >
>> >> >>
>> >> >>
>> >> >>
>> >> >> --
>> >> >> Jonathan Perkins
>> >> >> http://www.cse.ohio-state.edu/~perkinjo
>> >> >>
>> >> >> _______________________________________________
>> >> >> mvapich-discuss mailing list
>> >> >> mvapich-discuss at cse.ohio-state.edu
>> >> >> http://mail.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>> >> > --
>> >> > Dan Kokron
>> >> > Global Modeling and Assimilation Office, Code 610.1
>> >> > NASA Goddard Space Flight Center
>> >> > Greenbelt, MD 20771 USA
>> >> > Daniel.S.Kokron at nasa.gov
>> >> > Phone: (301) 614-5192
>> >> > Fax:   (301) 614-5304
>> >> >
>> >> >
>> >>
>> >>
>> >>
>> > --
>> > Dan Kokron
>> > Global Modeling and Assimilation Office, Code 610.1
>> > NASA Goddard Space Flight Center
>> > Greenbelt, MD 20771 USA
>> > Daniel.S.Kokron at nasa.gov
>> > Phone: (301) 614-5192
>> > Fax:   (301) 614-5304
>> >
>>
>>
>>
> --
> Dan Kokron
> Global Modeling and Assimilation Office, Code 610.1
> NASA Goddard Space Flight Center
> Greenbelt, MD 20771 USA
> Daniel.S.Kokron at nasa.gov
> Phone: (301) 614-5192
> Fax:   (301) 614-5304
>
>



-- 
Jonathan Perkins
http://www.cse.ohio-state.edu/~perkinjo

More information about the mvapich-discuss mailing list