search

pmodels / mpich

Official MPICH Repository
http://www.mpich.org
Other
536 stars 280 forks source link

mpich2-1.5b1 build failure on opensolaris #1639

Closed mpichbot closed 7 years ago

mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-06-11 15:59:09 -0500

I get the following error on opensolaris - when building mpich2-1.5b1 [and also with the latest nightly snapshot]. Is this supporsed to work with solaris compilers?


  CC     hydra_pmi_proxy-pmip_pmi_v2.o
  CC     hydra_pmi_proxy-common.o
  CC     hydra_pmi_proxy-pmi_v2_common.o
  CCLD   hydra_pmi_proxy
  CC     mpiexec.o
  CC     utils.o
  CCLD   mpiexec.hydra
Making all in .
  CC     src/mpi/attr/attrutil.lo
"./src/include/mpiimpl.h", line 3230: zero-sized struct/union
cc: acomp failed for src/mpi/attr/attrutil.c
*** Error code 1
The following command caused the error:
echo "  CC    " src/mpi/attr/attrutil.lo; \
source='src/mpi/attr/attrutil.c' object='src/mpi/attr/attrutil.lo' libtool=yes \
DEPDIR=.deps depmode=none /bin/sh ./confdb/depcomp \
/bin/sh ./libtool --silent --tag=CC    --mode=compile cc -DHAVE_CONFIG_H  -I. -I./src/include  -I./src/include  -I./src/mpi/datatype -I./src/mpi/spawn  -Isrc/util/logging/common -I./src/util/wrappers  -I./src/util/wrappers -I./src/binding/cxx  -I/export/home/petsc/soft/build/mpich2-1.5b1/src/binding/f77 -Isrc/binding/f90 -I./src/mpid/ch3/include    -I./src/mpid/ch3/include   -I./src/mpid/ch3/channels/nemesis/include  -I./src/mpid/ch3/channels/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/utils/monitor   -I./src/mpid/common/datatype  -I./src/mpid/common/datatype -I./src/mpid/common/sched    -I./src/mpid/common/thread  -I/export/home/petsc/soft/build/mpich2-1.5b1/src/mpi/comm   -I./src/pmi/simple  -I/export/home/petsc/soft/build/mpich2-1.5b1/src/mpl/include -I/export/home/petsc/soft/build/mpich2-1.5b1/src/mpl/include -I/export/home/petsc/soft/build/mpich2-1.5b1/src/openpa/src -I/export/home/petsc/soft/build/mpich2-1.5b1/src/openpa/src -I/export/home/petsc/soft/build/mpich2-1.5b1/src/mpi/romio/include    -c -o src/mpi/attr/attrutil.lo src/mpi/attr/attrutil.c
make: Fatal error: Command failed for target `src/mpi/attr/attrutil.lo'
Current working directory /export/home/petsc/soft/build/mpich2-1.5b1
*** Error code 1
The following command caused the error:
fail= failcom='exit 1'; \
for f in x $MAKEFLAGS; do \
  case $f in \
    -=* | --[!k]*);; \
    -k*) failcom='fail=yes';; \
  esac; \
done; \
dot_seen=no; \
target=`echo all-recursive | sed s/-recursive//`; \
list='src/mpl src/openpa src/mpe2 src/mpi/romio  src/pm/hydra  . examples'; for subdir in $list; do \
  echo "Making $target in $subdir"; \
  if test "$subdir" = "."; then \
    dot_seen=yes; \
    local_target="$target-am"; \
  else \
    local_target="$target"; \
  fi; \
  (CDPATH="${ZSH_VERSION+.}:" && cd $subdir && make  $local_target) \
  || eval $failcom; \
done; \
if test "$dot_seen" = "no"; then \
  make  "$target-am" || exit 1; \
fi; test -z "$fail"
make: Fatal error: Command failed for target `all-recursive'
Current working directory /export/home/petsc/soft/build/mpich2-1.5b1
*** Error code 1
make: Fatal error: Command failed for target `all'
petsc@n-gage:~/soft/build/mpich2-1.5b1$
mpichbot commented 7 years ago

Originally by goodell on 2012-06-11 16:13:47 -0500

try with [8da041ce16df820783cd0e3887b1e412d91c3f54]

mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-06-12 11:26:23 -0500

Replying to goodell:

try with [8da041ce16df820783cd0e3887b1e412d91c3f54]

Now get different errors.

  $ ./configure --prefix=/export/home/petsc/petsc-dev/arch-mpich-nemesis CC=cc CFLAGS= -O  --disable-cxx FC=f90 FCFLAGS= -O   F77=f90 FFLAGS= -O   --with-device=ch3:nemesis --without-mpe --with-pm=hydra --enable-g=meminit --enable-fast

Making all in .
  CC     src/mpi/attr/attrutil.lo
  CC     src/mpi/attr/dup_fn.lo
  CC     src/mpi/coll/allred_group.lo
  CC     src/mpi/coll/barrier_group.lo
  CC     src/mpi/coll/helper_fns.lo
  CC     src/mpi/coll/opsum.lo
*** Error code 1
The following command caused the error:
echo "  CC    " src/mpi/coll/opsum.lo; \
source='src/mpi/coll/opsum.c' object='src/mpi/coll/opsum.lo' libtool=yes \
DEPDIR=.deps depmode=none /bin/sh ./confdb/depcomp \
/bin/sh ./libtool --silent --tag=CC    --mode=compile cc -DHAVE_CONFIG_H  -I. -I./src/include  -I./src/include  -I./src/mpi/datatype -I./src/mpi/spawn  -Isrc/util/logging/common -I./src/util/wrappers  -I./src/util/wrappers   -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/binding/f77 -Isrc/binding/f90 -I./src/mpid/ch3/include    -I./src/mpid/ch3/include   -I./src/mpid/ch3/channels/nemesis/include  -I./src/mpid/ch3/channels/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/include  -I./src/mpid/ch3/channels/nemesis/nemesis/utils/monitor   -I./src/mpid/common/datatype  -I./src/mpid/common/datatype -I./src/mpid/common/sched    -I./src/mpid/common/thread  -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/mpi/comm   -I./src/pmi/simple  -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/mpl/include -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/mpl/include -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/openpa/src -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/openpa/src -I/export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]/src/mpi/romio/include   -O    -DNDEBUG -DNVALGRIND -c -o src/mpi/coll/opsum.lo src/mpi/coll/opsum.c
*** Error code 1
The following command caused the error:
fail= failcom='exit 1'; \
for f in x $MAKEFLAGS; do \
  case $f in \
    -=* | --[!k]*);; \
    -k*) failcom='fail=yes';; \
  esac; \
done; \
dot_seen=no; \
target=`echo all-recursive | sed s/-recursive//`; \
list='src/mpl src/openpa  src/mpi/romio  src/pm/hydra  . examples'; for subdir in $list; do \
  echo "Making $target in $subdir"; \
  if test "$subdir" = "."; then \
    dot_seen=yes; \
    local_target="$target-am"; \
  else \
    local_target="$target"; \
  fi; \
  (CDPATH="${ZSH_VERSION+.}:" && cd $subdir && make  $local_target) \
  || eval $failcom; \
done; \
if test "$dot_seen" = "no"; then \
  make  "$target-am" || exit 1; \
fi; test -z "$fail"
*** Error code 1
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
"adio/common/ad_fstype.c", line 309: warning: statement not reached
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
make: Warning: Ignoring DistributedMake -j option
"src/mpi/coll/oputil.h", line 236: syntax error before or at: re
"src/mpi/coll/oputil.h", line 246: syntax error before or at: typedef
"src/mpi/coll/oputil.h", line 252: syntax error before or at: typedef
"src/mpi/coll/opsum.c", line 26: member cannot be function: MPIR_SUM
"src/mpi/coll/opsum.c", line 26: warning: syntax requires ";" after last struct/union member
"src/mpi/coll/opsum.c", line 26: syntax error before or at: {
"src/mpi/coll/opsum.c", line 27: warning: syntax requires ";" after last struct/union member
"src/mpi/coll/opsum.c", line 27: syntax error before or at: =
"src/mpi/coll/opsum.c", line 33: warning: syntax requires ";" after last struct/union member
"src/mpi/coll/opsum.c", line 33: zero-sized struct/union
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000407
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000203
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000204
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000406
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000408
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000809
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000819
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000118
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000102
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000137
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000238
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c000439
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c00083a
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c00013b
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c00023c
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c00043d
"src/mpi/coll/opsum.c", line 33: syntax error before or at: 0x4c00083e
"src/mpi/coll/opsum.c", line 34: syntax error before or at: 1275069467
"src/mpi/coll/opsum.c", line 34: syntax error before or at: 0x4c000443
"src/mpi/coll/opsum.c", line 34: syntax error before or at: 0x4c000844
"src/mpi/coll/opsum.c", line 35: syntax error before or at: 0x4c00040a
"src/mpi/coll/opsum.c", line 35: syntax error before or at: 0x4c00080b
"src/mpi/coll/opsum.c", line 35: syntax error before or at: 1275069468
"src/mpi/coll/opsum.c", line 35: syntax error before or at: 1275070495
"src/mpi/coll/opsum.c", line 35: syntax error before or at: 0x4c000c0c
"src/mpi/coll/opsum.c", line 37: syntax error before or at: 0x4c000101
"src/mpi/coll/opsum.c", line 38: syntax error before or at: 1275068698
"src/mpi/coll/opsum.c", line 38: syntax error before or at: 0x4c00012d
"src/mpi/coll/opsum.c", line 38: syntax error before or at: 0x4c00022f
"src/mpi/coll/opsum.c", line 38: syntax error before or at: 0x4c000430
"src/mpi/coll/opsum.c", line 38: syntax error before or at: 0x4c000831
"src/mpi/coll/opsum.c", line 39: syntax error before or at: 0x4c000427
"src/mpi/coll/opsum.c", line 39: syntax error before or at: 0x4c000829
"src/mpi/coll/opsum.c", line 56: syntax error before or at: 1275070494
"src/mpi/coll/opsum.c", line 56: warning: syntax error:  empty declaration
"src/mpi/coll/opsum.c", line 56: syntax error before or at: 0x4c000840
"src/mpi/coll/opsum.c", line 56: syntax error before or at: 0x4c001041
"src/mpi/coll/opsum.c", line 56: syntax error before or at: 0x4c001842
"src/mpi/coll/opsum.c", line 57: syntax error before or at: 1275072546
"src/mpi/coll/opsum.c", line 57: warning: syntax error:  empty declaration
"src/mpi/coll/opsum.c", line 57: syntax error before or at: 0x4c000828
"src/mpi/coll/opsum.c", line 57: warning: syntax error:  empty declaration
"src/mpi/coll/opsum.c", line 57: syntax error before or at: 0x4c00102a
"src/mpi/coll/opsum.c", line 57: warning: syntax error:  empty declaration
"src/mpi/coll/opsum.c", line 78: type specifier can not be used as array size expression qualifier
"src/mpi/coll/opsum.c", line 78: warning: no explicit type given
"src/mpi/coll/opsum.c", line 79: warning: old-style declaration or incorrect type for: MPIR_SUM_check_dtype
"src/mpi/coll/opsum.c", line 100: cannot recover from previous errors
cc: acomp failed for src/mpi/coll/opsum.c
make: Fatal error: Command failed for target `src/mpi/coll/opsum.lo'
Current working directory /export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]
make: Fatal error: Command failed for target `all-recursive'
Current working directory /export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[8da041ce16df820783cd0e3887b1e412d91c3f54]
make: Fatal error: Command failed for target `all'
*******************************************************************************
mpichbot commented 7 years ago

Originally by goodell on 2012-06-12 12:19:20 -0500

Can you attach your "$(topbuilddir)/src/include/mpichconf.h" to the ticket?

mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-06-12 13:15:04 -0500

Attachment added: mpichconf.h (36.5 KiB) $(topbuilddir)/src/include/mpichconf.h

mpichbot commented 7 years ago

Originally by goodell on 2012-06-12 13:30:23 -0500

Looks like we had a typo in a configure fallback case that is probably rarely exercised. I've fixed that bug in [6dc9922f29ca8c95060d351ccea693fb15095f8e], but I can't guarantee everything will work fine still. It's rather suspicious that configure is unable to locate a C type that matches the Fortran REAL type. If you want to try going further and pretend that the variable was not typo'ed without a full re-./autogen.sh, then modify line 793 of your mpichconf.h to read:

#define MPIR_FC_REAL_CTYPE float
mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-06-13 14:09:41 -0500

Trying out [9948] tarball - and I get the following:

./configure --prefix=/export/home/petsc/petsc-dev/arch-mpich-nemesis CC=cc CFLAGS=" -O"  --disable-cxx FC=f90 FCFLAGS=" -O "  F77=f90 FFLAGS=" -O "  --with-device=ch3:nemesis --without-mpe --with-pm=hydra --enable-g=meminit --enable-fast
  CC     src/binding/f77/lib_libfmpich_la-statusf2c.lo
  CC     src/binding/f77/lib_libfmpich_la-statusc2f.lo
  CCLD   lib/libfmpich.la
rm -f src/binding/f90/mpifnoext.h
sed -e 's/^C/\!/g' -e '/EXTERNAL/d' \
        -e '/REAL\*8/d' \
        -e '/DOUBLE PRECISION/d' \
        -e '/MPI_WTICK/d' src/binding/f77/mpif.h > src/binding/f90/mpifnoext.h
   MOD   src/binding/f90/mpi_constants.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_constants.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_sizeofs.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_sizeofs.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_base.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_constants.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_base.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
  FCLD   lib/libmpichf90.la
  CC     src/env/mpich2version.o
  CCLD   src/env/mpich2version
cp -p src/env/mpicc.bash src/env/mpicc
cp -p src/env/mpif77.bash src/env/mpif77
cp -p src/env/mpif90.bash src/env/mpif90
   MOD   src/binding/f90/mpi_constants.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_sizeofs.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_base.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_constants.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi_base.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
   MOD   src/binding/f90/mpi.mod-stamp
f90: Invalid input file name src/binding/f90, no output generated for this file.
Making all in examples
  CC     cpi.o
  CCLD   cpi
petsc@n-gage:~/petsc-dev/externalpackages/mpich2-trunk-[9948]$

'make install' gives errors

{{{ if [ ! -e /export/home/petsc/petsc-dev/arch-mpich-nemesis/share/doc/mpich2 ] ; then mkdir -p /export/home/petsc/petsc -dev/arch-mpich-nemesis/share/doc/mpich2 ; fi if [ -s ./doc/userguide/user.pdf ] ; then /usr/bin/ginstall -c -m 644 ./doc/userguide/user.pdf /export/home/petsc/pet sc-dev/arch-mpich-nemesis/share/doc/mpich2/user.pdf ; fi if [ -s ./doc/installguide/install.pdf ] ; then /usr/bin/ginstall -c -m 644 ./doc/installguide/install.pdf /export/ho me/petsc/petsc-dev/arch-mpich-nemesis/share/doc/mpich2/install.pdf ; fi if [ -s ./doc/smpd/smpd_pmi.pdf ] ; then /usr/bin/ginstall -c -m 644 ./doc/smpd/smpd_pmi.pdf /export/home/petsc/petsc -dev/arch-mpich-nemesis/share/doc/mpich2/smpd_pmi.pdf ; fi if [ -s ./doc/logging/logging.pdf ] ; then /usr/bin/ginstall -c -m 644 ./doc/logging/logging.pdf /export/home/petsc/p etsc-dev/arch-mpich-nemesis/share/doc/mpich2/logging.pdf ; fi MOD src/binding/f90/mpi_constants.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. MOD src/binding/f90/mpi_sizeofs.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. MOD src/binding/f90/mpi_base.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. MOD src/binding/f90/mpi_constants.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. MOD src/binding/f90/mpi_base.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. MOD src/binding/f90/mpi.mod-stamp f90: Invalid input file name src/binding/f90, no output generated for this file. test -z "/export/home/petsc/petsc-dev/arch-mpich-nemesis/include" || mkdir -p "/export/home/petsc/petsc-dev/arch-mpic h-nemesis/include" /usr/bin/ginstall -c -m 644 ./src/binding/f90/mpi.mod ./src/binding/f90/mpi_sizeofs.mod ./src/binding/f90/mpi_consta nts.mod ./src/binding/f90/mpi_base.mod src/include/mpix.h '/export/home/petsc/petsc-dev/arch-mpich-nemesis/include' /usr/bin/ginstall: cannot stat ./src/binding/f90/mpi.mod': No such file or directory /usr/bin/ginstall: cannot stat./src/binding/f90/mpi_sizeofs.mod': No such file or directory /usr/bin/ginstall: cannot stat ./src/binding/f90/mpi_constants.mod': No such file or directory /usr/bin/ginstall: cannot stat./src/binding/f90/mpi_base.mod': No such file or directory * Error code 1 The following command caused the error: list='src/binding/f90/mpi.mod src/binding/f90/mpi_sizeofs.mod src/binding/f90/mpi_constants.mod src/binding/f90/mp ibase.mod src/include/mpix.h'; test -n "/export/home/petsc/petsc-dev/arch-mpich-nemesis/include" || list=; \ for p in $list; do \ if test -f "$p"; then d=; else d="./"; fi; \ echo "$d$p"; \ done | sed '$!N;$!N;$!N;$!N;$!N;$!N;$!N;s/\n/ /g' | sed '$!N;$!N;$!N;$!N;s/\n/ /g' | \ while read files; do \ echo " /usr/bin/ginstall -c -m 644 $files '/export/home/petsc/petsc-dev/arch-mpich-nemesis/include'"; \ /usr/bin/ginstall -c -m 644 $files "/export/home/petsc/petsc-dev/arch-mpich-nemesis/include" || exit $?; \ done make: Fatal error: Command failed for target install-includeHEADERS' Current working directory /export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[9948] *** Error code 1 The following command caused the error: make install-exec-am install-data-am make: Fatal error: Command failed for targetinstall-am' Current working directory /export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[9948] ** Error code 1 The following command caused the error: fail= failcom='exit 1'; \ for f in x $MAKEFLAGS; do \ case $f in \ -=\ | --[!k]);; \ -k_) failcom='fail=yes';; \ esac; \ done; \ dot_seen=no; \ target=echo install-recursive | sed s/-recursive//; \ list='src/mpl src/openpa src/mpi/romio src/pm/hydra . examples'; for subdir in $list; do \ echo "Making $target in $subdir"; \ if test "$subdir" = "."; then \ dot_seen=yes; \ local_target="$target-am"; \ else \ local_target="$target"; \ fi; \ (CDPATH="${ZSH_VERSION+.}:" && cd $subdir && make $local_target) \ || eval $failcom; \ done; \ if test "$dot_seen" = "no"; then \ make "$target-am" || exit 1; \ fi; test -z "$fail" make: Fatal error: Command failed for target `install-recursive' Current working directory /export/home/petsc/petsc-dev/externalpackages/mpich2-trunk-[9948] * Error code 1 make: Fatal error: Command failed for target`install' petsc@n-gage:~/petsc-dev/externalpackages/mpich2-trunk-[9948]$

mpichbot commented 7 years ago

Originally by goodell on 2012-06-13 14:33:15 -0500

Unfortunately I don't have the time to troubleshoot/fix this any further right now. Clearly some f90 build system code isn't compatible with something about the Solaris f90 compiler.

I'm happy to take patches if you end up debugging the problem yourself. I just can't flush a lot of time getting my opensolaris env going again right now.

mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-06-13 14:39:18 -0500

Replying to goodell:

Unfortunately I don't have the time to troubleshoot/fix this any further right now. Clearly some f90 build system code isn't compatible with something about the Solaris f90 compiler.

I'm happy to take patches if you end up debugging the problem yourself. I just can't flush a lot of time getting my opensolaris env going again right now.

I can provide access to this test box - if it helps. [Anthony has access to it] I haven't tried it - but presumably its reproducable on niagara1 aswell?

Satish

mpichbot commented 7 years ago

Originally by goodell on 2012-06-14 10:35:07 -0500

Some more Solaris problems from mpich-discuss@:

http://lists.mcs.anl.gov/pipermail/mpich-discuss/2012-June/012625.html

1.5b1:

Making all in src/mpi/romio

....

 CCLD   libromio.la
ar: cannot open .libs/libromio.a
       No such file or directory
ar: .libs/libromio.a not found
*** Error code 1
The following command caused the error:
echo "  CCLD  " libromio.la;/bin/bash ./libtool --silent --tag=CC    --mode=link /grid/common/pkgs/forte/v12.0p2/SUNWspro/bin/cc  -m64 -KPIC -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -DHAVE_ROMIOCONF_H -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64    -o libromio.la   mpi-io/close.lo mpi-io/delete.lo  mpi-io/file_c2f.lo mpi-io/file_f2c.lo mpi-io/fsync.lo  mpi-io/get_amode.lo mpi-io/get_atom.lo mpi-io/get_bytoff.lo  mpi-io/get_extent.lo mpi-io/get_group.lo mpi-io/get_info.lo  mpi-io/get_posn.lo mpi-io/get_posn_sh.lo mpi-io/get_size.lo mpi-io/get_view.lo mpi-io/iread.lo mpi-io/iread_at.lo  mpi-io/iread_sh.lo mpi-io/iwrite.lo mpi-io/iwrite_at.lo  mpi-io/iwrite_sh.lo mpi-io/open.lo mpi-io/prealloc.lo  mpi-io/rd_atallb.lo mpi-io/rd_atalle.lo mpi-io/read.lo  mpi-io/read_all.lo mpi-io/read_allb.lo mpi-io/read_alle.lo  mpi-io/read_at.lo mpi-io/read_atall.lo mpi-io/read_ord.lo  mpi-io/read_ordb.lo mpi-io/read_orde.lo mpi-io/read_sh.lo  mpi-io/register_datarep.lo mpi-io/seek.lo mp
i-io/seek_sh.lo  mpi-io/set_atom.lo mpi-io/set_info.lo mpi-io/set_size.lo  mpi-io/set_view.lo mpi-io/wr_atallb.lo mpi-io/wr_atalle.lo  mpi-io/write.lo mpi-io/write_all.lo mpi-io/write_allb.lo  mpi-io/write_alle.lo mpi-io/write_at.lo mpi-io/write_atall.lo  mpi-io/write_ord.lo mpi-io/write_ordb.lo mpi-io/write_orde.lo  mpi-io/write_sh.lo    mpi-io/glue/mpich2/mpio_file.lo  mpi-io/glue/mpich2/mpio_err.lo  mpi-io/mpich2_fileutil.lo mpi-io/mpir-mpioinit.lo  mpi-io/mpiu_greq.lo          adio/ad_nfs/ad_nfs_read.lo  adio/ad_nfs/ad_nfs_open.lo adio/ad_nfs/ad_nfs_write.lo  adio/ad_nfs/ad_nfs_done.lo  adio/ad_nfs/ad_nfs_fcntl.lo  adio/ad_nfs/ad_nfs_iread.lo adio/ad_nfs/ad_nfs_iwrite.lo  adio/ad_nfs/ad_nfs_wait.lo  adio/ad_nfs/ad_nfs_setsh.lo  adio/ad_nfs/ad_nfs_getsh.lo  adio/ad_nfs/ad_nfs.lo adio/ad_nfs/ad_nfs_resize.lo  adio/ad_nfs/ad_nfs_features.lo          adio/ad_testfs/ad_testfs_close.lo  adio/ad_testfs/ad_testfs_read.lo adio/ad_testfs/ad_testfs_rdcoll.lo  adio/ad_testfs/ad_
testfs_wrcoll.lo  adio/ad_testfs/ad_testfs_open.lo  adio/ad_testfs/ad_testfs_write.lo  adio/ad_testfs/ad_testfs_done.lo adio/ad_testfs/ad_testfs_fcntl.lo  adio/ad_testfs/ad_testfs_iread.lo  adio/ad_testfs/ad_testfs_iwrite.lo  adio/ad_testfs/ad_testfs_wait.lo adio/ad_testfs/ad_testfs_flush.lo  adio/ad_testfs/ad_testfs_seek.lo  adio/ad_testfs/ad_testfs_resize.lo  adio/ad_testfs/ad_testfs_hints.lo adio/ad_testfs/ad_testfs_delete.lo  adio/ad_testfs/ad_testfs.lo adio/ad_ufs/ad_ufs.lo  adio/ad_ufs/ad_ufs_open.lo    adio/common/ad_aggregate.lo  adio/common/ad_aggregate_new.lo adio/common/ad_close.lo  adio/common/ad_coll_build_req_new.lo adio/common/ad_coll_exch_new.lo adio/common/ad_darray.lo  adio/common/ad_delete.lo adio/common/ad_done.lo  adio/common/ad_done_fake.lo adio/common/ad_end.lo  adio/common/ad_fcntl.lo adio/common/ad_features.lo  adio/common/ad_flush.lo adio/common/ad_fstype.lo adio/common/ad_get_sh_fp.lo adio/common/ad_hints.lo  adio/common/ad_init.lo adio/common
/ad_io_coll.lo  adio/common/ad_iopen.lo adio/common/ad_iread.lo  adio/common/ad_iread_fake.lo adio/common/ad_iwrite.lo adio/common/ad_iwrite_fake.lo adio/common/ad_open.lo  adio/common/ad_opencoll.lo adio/common/ad_opencoll_failsafe.lo adio/common/ad_opencoll_scalable.lo adio/common/ad_prealloc.lo  adio/common/ad_read.lo adio/common/ad_read_coll.lo adio/common/ad_read_str.lo adio/common/ad_read_str_naive.lo  adio/common/ad_resize.lo adio/common/ad_seek.lo  adio/common/ad_set_sh_fp.lo adio/common/ad_set_view.lo  adio/common/ad_subarray.lo adio/common/ad_wait.lo  adio/common/ad_wait_fake.lo adio/common/ad_write.lo adio/common/ad_write_coll.lo adio/common/ad_write_nolock.lo  adio/common/ad_write_str.lo adio/common/ad_write_str_naive.lo adio/common/adi_close.lo adio/common/byte_offset.lo  adio/common/cb_config_list.lo adio/common/eof_offset.lo  adio/common/error.lo adio/common/flatten.lo  adio/common/get_fp_posn.lo adio/common/greq_fns.lo  adio/common/heap-sort.lo adio/comm
on/iscontig.lo  adio/common/lock.lo adio/common/malloc.lo  adio/common/shfp_fname.lo adio/common/status_setb.lo  adio/common/strfns.lo adio/common/system_hints.lo   -lrt -lnsl -lsocket -lpthread
make: Fatal error: Command failed for target `libromio.la'

Configure options in both cases:

       --with-pm=hydra --with-device=ch3:nemesis                               \
       --disable-chekpointing                                                  \
       --disable-f77 --disable-fc                                              \
       --enable-fast=O4                                                        \
       CC=/grid/common/pkgs/forte/v12.0p2/SUNWspro/bin/cc CFLAGS="-m64"        \
       CXX=/grid/common/pkgs/forte/v12.0p2/SUNWspro/bin/CC CXXFLAGS="-m64"
mpichbot commented 7 years ago

Originally by goodell on 2012-06-14 20:48:27 -0500

More info from Eugene:

I tried to compile 1.5b1  with --disable-romio and there is error in src/pm/hydra/utils/sock/sock.c:

"./utils/sock/sock.c", line 546: undefined symbol: remote_access

remote_access is defined under HAVE_GETIFADDRS condition, but used under HAVE_INET_NTOP.

Issue with cpi eems gone in [9948].

Eugene.
mpichbot commented 7 years ago

Originally by goodell on 2012-06-14 20:49:05 -0500

CC'ing Eugene

mpichbot commented 7 years ago

Originally by zelenko@cadence.com on 2012-06-15 12:32:25 -0500

sock.c issue is trivial to fix by changing from

#!c
#if defined(HAVE_GETIFADDRS)
    struct ifaddrs *ifaddr, *ifa;
    int remote_access;
#endif /* HAVE_GETIFADDRS */

to

#!c
#if defined(HAVE_GETIFADDRS)
    struct ifaddrs *ifaddr, *ifa;
#endif /* HAVE_GETIFADDRS */
#if defined HAVE_INET_NTOP
    int remote_access;
#endif /* HAVE_INET_NTOP */

By there are also hwloc linking issues after:

  CCLD   hydra_persist
Undefined                       first referenced
 symbol                             in file
picl_initialize                     /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_first_prop                 /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_walk_tree_by_class             /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_next_prop                  /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_propval                    /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_shutdown                       /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_root                       /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_next_by_col                /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_next_by_row                /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)
picl_get_propinfo                   /home/zelenko/Installs/mpich2-[9948]/source/src/pm/hydra/tools/topo/hwloc/hwloc/src/.libs/libhwloc_embedded.a(topology-solaris-chiptype.o)

Looks like hwloc shipped with MPICH2 contains only picl.h without relevant implementation.

mpichbot commented 7 years ago

Originally by goodell on 2012-06-15 17:16:44 -0500

Thanks for the sock.c suggestion. I'll be sure to incorporate it when I sit down to get our Solaris support back into shape.

For the hwloc issue, can you add -lpicl to your LIBS at configure time? I think this is some sort of Solaris system service that hwloc is contacting.

mpichbot commented 7 years ago

Originally by zelenko@cadence.com on 2012-06-15 18:00:59 -0500

You are correct libpicl.so is part of OS and you suggestion works. However picl.h is also present in /usr/include and it is bigger then file in hwloc.

I think will be good idea to rely on OS supplied files and remove copy from hwloc. May be configure should check for both include and library files if it's optional Solaris package?

I also noticed inconsistency between make output in mpe2 and rest of MPICH2. In MPE2 full compilation line shown, when rest of code made with just CC.

mpichbot commented 7 years ago

Originally by zelenko@cadence.com on 2012-06-18 16:13:26 -0500

I experimented with standalone hwloc 1.4.2 build. In standalone mode dependency libs list includes libpicl, in embedded mode - not.

mpichbot commented 7 years ago

Originally by goodell on 2012-08-09 16:05:04 -0500

Satish's issues should be fixed by [63612417e0359e85136c122afea48967d0021006], [97314c97148208941749918c243782416688798d], and [cac8caee238116315f32cc69db2ddd6acebdab47]. A default build will now yield a working MPICH2 installation.

I haven't addressed Eugene's concerns yet, although I'm not encountering them on our Niagara machine running Solaris 10 (I think that's the version).

mpichbot commented 7 years ago

Originally by goodell on 2012-08-09 16:13:13 -0500

It looks like Eugene's sock.c suggestion was already incorporated by Pavan in [1c19029f6332fa64ef8a9bfb3a52801e063c4348]. I suspect the libpicl issue was sorted out by the upgrade to hwloc-1.5rc2.

Eugene, can you re-test with the latest SVN or one of tonight's nightly tarballs and let us know if you are still having problems? I'm going to resolve this for now, feel free to reopen if problems remain.

mpichbot commented 7 years ago

Originally by balay@mcs.anl.gov on 2012-08-10 11:04:21 -0500

Replying to goodell:

Satish's issues should be fixed by [63612417e0359e85136c122afea48967d0021006], [97314c97148208941749918c243782416688798d], and [cac8caee238116315f32cc69db2ddd6acebdab47]. A default build will now yield a working MPICH2 installation.

Ok I've tried building mpich2-trunk-[cac8caee238116315f32cc69db2ddd6acebdab47].tar.gz with petsc configure - and it builds fine.

Thanks!

mpichbot commented 7 years ago

Originally by zelenko@cadence.com on 2012-08-10 13:02:41 -0500

Hi, Dave!

I tried [cac8caee238116315f32cc69db2ddd6acebdab47] and it compiled fine for me on Solaris 10.

Couple of suggestions:

src/pm/hydra/tools/topo/hwloc/hwloc/tests/ports/include/picl.h is still there as well as bunch of other files which should be taken from OS.

Also list of Sun compiler warnings:

"adio/common/ad_fstype.c", line 309: warning: statement not reached
"src/util/mem/handlemem.c", line 114: warning: null dimension: indirect
"src/util/mem/handlemem.c", line 175: warning: null dimension: indirect
"src/mpid/ch3/channels/nemesis/src/ch3i_comm.c", line 35: warning: statement not reached
"src/mpid/ch3/channels/nemesis/src/ch3i_comm.c", line 98: warning: statement not reached
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 124: warning: assignment type mismatch:
    pointer to void "=" pointer to function(int) returning void
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 125: warning: operands have incompatible pointer types: op "=="
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 126: warning: operands have incompatible pointer types: op "!="
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 126: warning: operands have incompatible pointer types: op "!="
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 128: warning: argument #2 is incompatible with prototype:
    prototype: pointer to function(int) returning void : "/usr/include/iso/signal_iso.h", line 48
    argument : pointer to void
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 128: warning: assignment type mismatch:
    pointer to void "=" pointer to function(int) returning void
"src/mpid/ch3/channels/nemesis/netmod/tcp/tcp_init.c", line 129: warning: operands have incompatible pointer types: op "=="
"src/mpid/common/datatype/dataloop/segment.c", line 876: warning: statement not reached
"src/mpi/coll/iallgather.c", line 726: warning: statement not reached
"src/mpi/coll/iallgatherv.c", line 819: warning: statement not reached
"src/mpi/coll/iallreduce.c", line 790: warning: statement not reached
"src/mpi/coll/ialltoall.c", line 656: warning: statement not reached
"src/mpi/coll/ialltoallv.c", line 382: warning: statement not reached
"src/mpi/coll/ialltoallw.c", line 376: warning: statement not reached
"src/mpi/coll/ibarrier.c", line 285: warning: statement not reached
"src/mpi/coll/ibcast.c", line 983: warning: statement not reached
"src/mpi/coll/iexscan.c", line 322: warning: statement not reached
"src/mpi/coll/igather.c", line 697: warning: statement not reached
"src/mpi/coll/igatherv.c", line 306: warning: statement not reached
"src/mpi/coll/ired_scat.c", line 1140: warning: statement not reached
"src/mpi/coll/ired_scat_block.c", line 1049: warning: statement not reached
"src/mpi/coll/ireduce.c", line 900: warning: statement not reached
"src/mpi/coll/iscan.c", line 456: warning: statement not reached
"src/mpi/coll/iscatter.c", line 724: warning: statement not reached
"src/mpi/coll/iscatterv.c", line 318: warning: statement not reached
"src/binding/cxx/initcxx.cxx", line 507: Warning (Anachronism): Formal argument copy_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*).
"src/binding/cxx/initcxx.cxx", line 507: Warning (Anachronism): Formal argument delete_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*).
"src/binding/cxx/initcxx.cxx", line 571: Warning (Anachronism): Formal argument copy_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*).
"src/binding/cxx/initcxx.cxx", line 571: Warning (Anachronism): Formal argument delete_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*).
"src/binding/cxx/initcxx.cxx", line 635: Warning (Anachronism): Formal argument copy_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*).
"src/binding/cxx/initcxx.cxx", line 635: Warning (Anachronism): Formal argument delete_proxy of type extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*) in call to MPI::MPIR_Keyval_set_proxy(int, extern "C" int(*)(extern "C" int(*)(int,int,void*,void*,void*,int*),int,int,void*,MPI::MPIR_AttrType,void*,void**,int*), extern "C" int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*)) is being passed int(*)(extern "C" int(*)(int,int,void*,void*),int,int,MPI::MPIR_AttrType,void*,void*).
6 Warning(s) detected.

Eugene.