November 2013 - devel - mailman3.cels.anl.gov

Re: [mpich-devel] /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:116:56: error: expected identifier static const MPI_Datatype mpich_mpi_char MPICH_ATTR_TYPE_T...
by Lisandro Dalcin 30 Nov '13

30 Nov '13

On 30 November 2013 00:43, Jeff Hammond <jeff.science(a)gmail.com> wrote: > A fellow GA user is having trouble with GA 5.2 on Mavericks with MPICH > 3.0.4 We have tried many things, including switching from Clang to > GCC and all sorts of different configure options. There appears to be > some issue with GCC on Mac (goto X) that makes linking MPICH break. > > I cannot reproduce this error because I neither have Intel Fortran on > any Mac nor have I upgraded to Mavericks. > Can you ask your fellow to install like this?: $ brew install mpich2 --enable-shared A friend of mine has similar issues when using static libraries, however I remember this was also related to Python distutils from the Anaconda Python distribution insisting on using an old OS X deployment target. > First, is there a Mavericks Mac in MCS to which I can SSH in order to > test the non-ifort part of this? Even better if said Mac has ifort... > > Second, can someone help me resolve the MPICH_ATTR_TYPE_TAG issues? > Ask your fellow to make sure to have the latest Xcode, "brew doctor" will tell. I've been having this issue a few month ago, it was related to a know clang issue that eventually got fixed in latest Xcode. -- Lisandro Dalcin --------------- CIMEC (UNL/CONICET) Predio CONICET-Santa Fe Colectora RN 168 Km 472, Paraje El Pozo 3000 Santa Fe, Argentina Tel: +54-342-4511594 (ext 1016) Tel/Fax: +54-342-4511169

1 0

Re: [mpich-devel] Unable to determine the size of MPI_BSEND_OVERHEAD
by Jeff Hammond 29 Nov '13

29 Nov '13

I needed to run autogen.sh because the configure stuff changed since my last update. In theory, configure could verify that it's file create is newer than the last git commit, but perhaps I should just accept responsibility for doing only self-consistent builds. Jeff On Fri, Nov 29, 2013 at 4:56 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > This one is real. It's just busted with every C compiler I have on my > Mac. I have no issue configuring with the 3.1b1 tarball version. > > Jeff > > On Fri, Nov 29, 2013 at 4:30 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >> Please hold. It appears that, one again, my Mac is unfit for >> development due to stupid ****ing gremlins in the package manager, >> which is now Homebrew, per Dave's recommendation. >> >> Jeff >> >> On Fri, Nov 29, 2013 at 4:13 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >>> It's broken with CC=clang as well. >>> >>> I'm at the head of master as of an hour ago: >>> >>> Jeffs-MacBook-Pro:build-peter jhammond$ git log >>> commit dcf00b350be1f1d33c039d1d6995d01fa743e1ba >>> Author: Ken Raffenetti <raffenet(a)mcs.anl.gov> >>> Date: Fri Nov 22 13:23:38 2013 -0600 >>> >>> Jeff >>> >>> On Fri, Nov 29, 2013 at 4:11 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >>>> Jeffs-MacBook-Pro:build-peter jhammond$ ../configure CC=cc >>>> --disable-f77 F77=false --disable-fc FC=false --disable-cxx CXX=false >>>> >>>> Jeffs-MacBook-Pro:build-peter jhammond$ cc -v >>>> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >>>> Target: x86_64-apple-darwin11.4.2 >>>> Thread model: posix >>>> >>>> Jeffs-MacBook-Pro:build-peter jhammond$ CC -v >>>> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >>>> Target: x86_64-apple-darwin11.4.2 >>>> Thread model: posix >>>> >>>> The associated config.log is attached. >>>> >>>> Jeff >>>> >>>> -- >>>> Jeff Hammond >>>> jeff.science(a)gmail.com >>> >>> >>> >>> -- >>> Jeff Hammond >>> jeff.science(a)gmail.com >> >> >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com > > > > -- > Jeff Hammond > jeff.science(a)gmail.com -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpich-devel] Unable to determine the size of MPI_BSEND_OVERHEAD
by Jeff Hammond 29 Nov '13

29 Nov '13

This one is real. It's just busted with every C compiler I have on my Mac. I have no issue configuring with the 3.1b1 tarball version. Jeff On Fri, Nov 29, 2013 at 4:30 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Please hold. It appears that, one again, my Mac is unfit for > development due to stupid ****ing gremlins in the package manager, > which is now Homebrew, per Dave's recommendation. > > Jeff > > On Fri, Nov 29, 2013 at 4:13 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >> It's broken with CC=clang as well. >> >> I'm at the head of master as of an hour ago: >> >> Jeffs-MacBook-Pro:build-peter jhammond$ git log >> commit dcf00b350be1f1d33c039d1d6995d01fa743e1ba >> Author: Ken Raffenetti <raffenet(a)mcs.anl.gov> >> Date: Fri Nov 22 13:23:38 2013 -0600 >> >> Jeff >> >> On Fri, Nov 29, 2013 at 4:11 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >>> Jeffs-MacBook-Pro:build-peter jhammond$ ../configure CC=cc >>> --disable-f77 F77=false --disable-fc FC=false --disable-cxx CXX=false >>> >>> Jeffs-MacBook-Pro:build-peter jhammond$ cc -v >>> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >>> Target: x86_64-apple-darwin11.4.2 >>> Thread model: posix >>> >>> Jeffs-MacBook-Pro:build-peter jhammond$ CC -v >>> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >>> Target: x86_64-apple-darwin11.4.2 >>> Thread model: posix >>> >>> The associated config.log is attached. >>> >>> Jeff >>> >>> -- >>> Jeff Hammond >>> jeff.science(a)gmail.com >> >> >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com > > > > -- > Jeff Hammond > jeff.science(a)gmail.com -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpich-devel] Unable to determine the size of MPI_BSEND_OVERHEAD
by Jeff Hammond 29 Nov '13

29 Nov '13

Please hold. It appears that, one again, my Mac is unfit for development due to stupid ****ing gremlins in the package manager, which is now Homebrew, per Dave's recommendation. Jeff On Fri, Nov 29, 2013 at 4:13 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > It's broken with CC=clang as well. > > I'm at the head of master as of an hour ago: > > Jeffs-MacBook-Pro:build-peter jhammond$ git log > commit dcf00b350be1f1d33c039d1d6995d01fa743e1ba > Author: Ken Raffenetti <raffenet(a)mcs.anl.gov> > Date: Fri Nov 22 13:23:38 2013 -0600 > > Jeff > > On Fri, Nov 29, 2013 at 4:11 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: >> Jeffs-MacBook-Pro:build-peter jhammond$ ../configure CC=cc >> --disable-f77 F77=false --disable-fc FC=false --disable-cxx CXX=false >> >> Jeffs-MacBook-Pro:build-peter jhammond$ cc -v >> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >> Target: x86_64-apple-darwin11.4.2 >> Thread model: posix >> >> Jeffs-MacBook-Pro:build-peter jhammond$ CC -v >> Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) >> Target: x86_64-apple-darwin11.4.2 >> Thread model: posix >> >> The associated config.log is attached. >> >> Jeff >> >> -- >> Jeff Hammond >> jeff.science(a)gmail.com > > > > -- > Jeff Hammond > jeff.science(a)gmail.com -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpich-devel] Unable to determine the size of MPI_BSEND_OVERHEAD
by Jeff Hammond 29 Nov '13

29 Nov '13

It's broken with CC=clang as well. I'm at the head of master as of an hour ago: Jeffs-MacBook-Pro:build-peter jhammond$ git log commit dcf00b350be1f1d33c039d1d6995d01fa743e1ba Author: Ken Raffenetti <raffenet(a)mcs.anl.gov> Date: Fri Nov 22 13:23:38 2013 -0600 Jeff On Fri, Nov 29, 2013 at 4:11 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Jeffs-MacBook-Pro:build-peter jhammond$ ../configure CC=cc > --disable-f77 F77=false --disable-fc FC=false --disable-cxx CXX=false > > Jeffs-MacBook-Pro:build-peter jhammond$ cc -v > Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) > Target: x86_64-apple-darwin11.4.2 > Thread model: posix > > Jeffs-MacBook-Pro:build-peter jhammond$ CC -v > Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) > Target: x86_64-apple-darwin11.4.2 > Thread model: posix > > The associated config.log is attached. > > Jeff > > -- > Jeff Hammond > jeff.science(a)gmail.com -- Jeff Hammond jeff.science(a)gmail.com

1 0

Unable to determine the size of MPI_BSEND_OVERHEAD
by Jeff Hammond 29 Nov '13

29 Nov '13

Jeffs-MacBook-Pro:build-peter jhammond$ ../configure CC=cc --disable-f77 F77=false --disable-fc FC=false --disable-cxx CXX=false Jeffs-MacBook-Pro:build-peter jhammond$ cc -v Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) Target: x86_64-apple-darwin11.4.2 Thread model: posix Jeffs-MacBook-Pro:build-peter jhammond$ CC -v Apple LLVM version 4.2 (clang-425.0.28) (based on LLVM 3.2svn) Target: x86_64-apple-darwin11.4.2 Thread model: posix The associated config.log is attached. Jeff -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpich-devel] /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:116:56: error: expected identifier static const MPI_Datatype mpich_mpi_char MPICH_ATTR_TYPE_T...
by Dmitri Gribenko 29 Nov '13

29 Nov '13

On Fri, Nov 29, 2013 at 1:43 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote: > Second, can someone help me resolve the MPICH_ATTR_TYPE_TAG issues? It looks like there is #define MPI somewhere that interferes with the attribute. Dmitri -- main(i,j){for(i=2;;i++){for(j=2;j<i;j++){if(!(i%j)){j=0;break;}}if (j){printf("%d\n",i);}}} /*Dmitri Gribenko <gribozavr(a)gmail.com>*/

1 0

/usr/local/Cellar/mpich2/3.0.4/include/mpi.h:116:56: error: expected identifier static const MPI_Datatype mpich_mpi_char MPICH_ATTR_TYPE_T...
by Jeff Hammond 29 Nov '13

29 Nov '13

A fellow GA user is having trouble with GA 5.2 on Mavericks with MPICH 3.0.4 We have tried many things, including switching from Clang to GCC and all sorts of different configure options. There appears to be some issue with GCC on Mac (goto X) that makes linking MPICH break. I cannot reproduce this error because I neither have Intel Fortran on any Mac nor have I upgraded to Mavericks. First, is there a Mavericks Mac in MCS to which I can SSH in order to test the non-ifort part of this? Even better if said Mac has ifort... Second, can someone help me resolve the MPICH_ATTR_TYPE_TAG issues? If I cannot resolve this soon, I am going to have no choice but to resort to recommending OtherMPI. Thanks, Jeff X: Undefined symbols for architecture x86_64: "___emutls_get_address", referenced from: _MPIU_Thread_CS_exit_lockname_recursive_impl_.isra.1.constprop.2 in libpmp ich.a(lib_libpmpich_la-finalize.o) _MPIU_Thread_CS_enter_lockname_recursive_impl_.isra.0.constprop.3 in libpm pich.a(lib_libpmpich_la-finalize.o) which I interpret as missing -legcc_eh or similar. ---------- Forwarded message ---------- From: Peter Knowles <peterjknowles(a)gmail.com> Date: Thu, Nov 28, 2013 at 1:59 PM Subject: [hpctools] Compliation of ga-5-2 under OSX 10.9 To: hpctools(a)googlegroups.com I am frustrated that I do not seem to be able to build GA on Mavericks. Does anyone have a hint that would help me get going, please? Details below; I've also tried quite hard to do it a number of other different ways. Thanks! Peter FC=ifort brew install --default-fortran-flags=-O3 mpich # brings in mpich2-3.0.4 ./configure >>>> produces: configure: MPI_LIBS= configure: MPI_LDFLAGS= configure: MPI_CPPFLAGS= configure: COMEX_NETWORK=MPI_TS configure: COMEX_NETWORK_LDFLAGS= configure: COMEX_NETWORK_LIBS= configure: COMEX_NETWORK_CPPFLAGS= configure: CC=mpicc configure: unwrapped CC=cc configure: CFLAGS=-g -O2 configure: CPP=mpicc -E configure: CPPFLAGS= configure: LDFLAGS= configure: LIBS= configure: FLIBS= configure: AR=ar configure: AR_FLAGS=cru configure: DEFS=-DHAVE_CONFIG_H configure: SHELL=/bin/sh configure: MPIEXEC=/usr/local/bin/mpirun -n %NP% configure: NPROCS=4 configure: configure: configure: ************************************************************** configure: Global Arrays (GA) configured as follows: configure: ************************************************************** configure: configure: TARGET=MACX64 configure: MSG_COMMS=MPI configure: GA_MP_LIBS= configure: GA_MP_LDFLAGS= configure: GA_MP_CPPFLAGS= configure: ARMCI_NETWORK=SOCKETS configure: ARMCI_NETWORK_LDFLAGS= configure: ARMCI_NETWORK_LIBS= configure: ARMCI_NETWORK_CPPFLAGS= configure: F77=mpif90 configure: unwrapped F77=ifort configure: FFLAGS= configure: FFLAG_INT=-integer-size 64 configure: GA_FOPT=-O3 -prefetch -w -cm configure: CC=mpicc configure: unwrapped CC=cc configure: CFLAGS= configure: GA_COPT= configure: CPP=mpicc -E configure: CPPFLAGS= configure: LDFLAGS= configure: LIBS= configure: FLIBS= -L/usr/bin/ifort-14.0-base/compiler/lib -L/usr/lib /usr/bin/ifort-14.0-base/compiler/lib/libifport.a /usr/bin/ifort-14.0-base/compiler/lib/libifcore.a /usr/bin/ifort-14.0-base/compiler/lib/libimf.a /usr/bin/ifort-14.0-base/compiler/lib/libsvml.a /usr/bin/ifort-14.0-base/compiler/lib/libipgo.a /usr/bin/ifort-14.0-base/compiler/lib/libirc.a -lpthread /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/../lib/clang/5.0/lib/darwin/libclang_rt.osx.a configure: BLAS_LDFLAGS= configure: BLAS_LIBS=-framework Accelerate configure: BLAS_CPPFLAGS= configure: AR=ar configure: AR_FLAGS=cru configure: CCAS=mpicc configure: CCAS_FLAGS= configure: DEFS=-DHAVE_CONFIG_H configure: SHELL=/bin/sh configure: MPIEXEC=/usr/local/bin/mpirun -n %NP% configure: NPROCS=4 configure: <<<< make /Applications/Xcode.app/Contents/Developer/usr/bin/make all-recursive Making all in armci /Applications/Xcode.app/Contents/Developer/usr/bin/make all-am depbase=`echo src/collectives/message.lo | sed 's|[^/]*$|.deps/&|;s|\.lo$||'`;\ /bin/sh ./libtool --tag=CC --mode=compile mpicc -DHAVE_CONFIG_H -I. -I./src/devices/sockets -DSHMEM -DMMAP -I./src/include -MT src/collectives/message.lo -MD -MP -MF $depbase.Tpo -c -o src/collectives/message.lo src/collectives/message.c &&\ mv -f $depbase.Tpo $depbase.Plo libtool: compile: mpicc -DHAVE_CONFIG_H -I. -I./src/devices/sockets -DSHMEM -DMMAP -I./src/include -MT src/collectives/message.lo -MD -MP -MF src/collectives/.deps/message.Tpo -c src/collectives/message.c -o src/collectives/message.o In file included from src/collectives/message.c:52: /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:116:56: error: expected identifier static const MPI_Datatype mpich_mpi_char MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:117:56: error: expected identifier static const MPI_Datatype mpich_mpi_signed_char MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:118:56: error: expected identifier static const MPI_Datatype mpich_mpi_unsigned_char MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:120:56: error: expected identifier static const MPI_Datatype mpich_mpi_wchar MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:121:56: error: expected identifier static const MPI_Datatype mpich_mpi_short MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:122:56: error: expected identifier static const MPI_Datatype mpich_mpi_unsigned_short MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:123:56: error: expected identifier static const MPI_Datatype mpich_mpi_int MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:124:56: error: expected identifier static const MPI_Datatype mpich_mpi_unsigned MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:125:56: error: expected identifier static const MPI_Datatype mpich_mpi_long MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:126:56: error: expected identifier static const MPI_Datatype mpich_mpi_unsigned_long MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:127:56: error: expected identifier static const MPI_Datatype mpich_mpi_float MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:128:56: error: expected identifier static const MPI_Datatype mpich_mpi_double MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:130:56: error: expected identifier static const MPI_Datatype mpich_mpi_long_double MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:132:56: error: expected identifier static const MPI_Datatype mpich_mpi_long_long_int MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:133:56: error: expected identifier static const MPI_Datatype mpich_mpi_unsigned_long_long MPICH_ATTR_TYPE_T... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:24:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG' ...__attribute__((type_tag_for_datatype(MPI,type))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:165:53: error: expected identifier static const MPI_Datatype mpich_mpi_float_int MPICH_ATTR_TYPE_TAG_... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:25:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG_LAYOUT_COMPATIBLE' ...__attribute__((type_tag_for_datatype(MPI,type,layout_compatible))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:166:53: error: expected identifier static const MPI_Datatype mpich_mpi_double_int MPICH_ATTR_TYPE_TAG_... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:25:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG_LAYOUT_COMPATIBLE' ...__attribute__((type_tag_for_datatype(MPI,type,layout_compatible))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:167:53: error: expected identifier static const MPI_Datatype mpich_mpi_long_int MPICH_ATTR_TYPE_TAG_... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:25:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG_LAYOUT_COMPATIBLE' ...__attribute__((type_tag_for_datatype(MPI,type,layout_compatible))) ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:168:53: error: expected identifier static const MPI_Datatype mpich_mpi_short_int MPICH_ATTR_TYPE_TAG_... ^ /usr/local/Cellar/mpich2/3.0.4/include/mpi.h:25:109: note: expanded from macro 'MPICH_ATTR_TYPE_TAG_LAYOUT_COMPATIBLE' ...__attribute__((type_tag_for_datatype(MPI,type,layout_compatible))) ^ fatal error: too many errors emitted, stopping now [-ferror-limit=] 20 errors generated. make[3]: *** [src/collectives/message.lo] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 -- You received this message because you are subscribed to the Google Groups "hpctools" group. To unsubscribe from this group and stop receiving emails from it, send an email to hpctools+unsubscribe(a)googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out. -- Jeff Hammond jeff.science(a)gmail.com

1 0

Re: [mpich-devel] Access to MPICH GIT repo
by Pavan Balaji 10 Nov '13

10 Nov '13

Ashwin, This was probably declined because no one in the group knew you. :-) I’ve added your new keys. —- Pavan On Nov 10, 2013, at 7:02 PM, Ashwin Aji <aaji(a)cs.vt.edu> wrote: > Hi Pavan, > > (I had emailed the devel@mpich ID but to no response, so am trying you directly instead) > > I wanted to use MPI-ACC from a few other machines other than HokieSpeed. So, can you please permit the following public keys to access the GIT repo [email protected]:dev/mpich-gpu-merge? > > ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEA1uiZ/2aUmxIHyqhgyBSUE6c1vRh1kd3D17iG/XOOY3KgRc6ldD1BlHgZQymWn9U8+spBFeoA4UfDUMLNVwhnMRq2urLteLEaa3fT7RhKbZikqRH/TuHhpZgqt159Ef+VfivF1i/cTwMiQnHuAyh+rnBXWyKyNZDkrcIr5ETDJkQwPzYwbLdIRLaGsQum+IwvbhqX71UQeP/CZ4sWlPfE+9l+ZDoIBVu60Lmv5E78G6MoGMxtUDGgWxt+di6gmK+iAohZtPmOuzpj67jp/xmH5kr/CKC4Aa+INE3HsSAb4jW5PieazR3Pl8FNUBQdIps1TErqPj/QNlf4RGFmbUIbEQ== aaji(a)fire.cs.vt.edu > > ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCewrgUnBgkvCaS5AFvopLEPopihwqwYsJgqEdqoEfCBv4D1eQ7gCfzBTqiHev8UvzYAMCmW+idLLxi9WyHdB/J2G+pUMTw4lWoBTsEB76+rrdF3kq5da36z2tKC7UCgpe7gKK46/N04zcmcXklkuukCaeb7FK5bL06cBmymw8kSc0JdkRvEEM1UiIcgGSLf+iN/qkgwE4yIQW182lXxLNxIC0Jo2g6KF9QodmlzyqLw5pO35rFVcp41nJ5lv17vWUFTA0xC1nVB0gfKJgR9XwVeLh2wEz6w9v2fxFVFi4C819QX+Zd2Y/ZmyJ8rodTpn3QbGLm2hGk4aYq+BRLiI5j aaji@charon > > Thanks, > Ashwin -- Pavan Balaji http://www.mcs.anl.gov/~balaji

1 0

Re: [mpich-devel] Bunch of warnings compiling MPICH
by Junchao Zhang 01 Nov '13

01 Nov '13

It is due to MPIU_Assert(addr != NULL) when addr is taken from a static var. Let me fix that. Thanks. --Junchao Zhang On Fri, Nov 1, 2013 at 5:15 PM, Antonio J. Peña <apenya(a)mcs.anl.gov> wrote: > > Hi folks, > > Since my last git fetch + update, I'm getting the warnings listed below. > It's > mentioning MPIR_CVAR in function MPIR_T_cvar_init, what seems to point to > Junchao. I'm using gcc 4.7.2. Anyone else experiencing this issue? > > Toni > > > src/util/cvar/mpich_cvars.c: In function 'MPIR_T_cvar_init': > src/util/cvar/mpich_cvars.c:278:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLTOALL_SHORT_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:296:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLTOALL_MEDIUM_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:314:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLTOALL_THROTTLE' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:332:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_REDSCAT_COMMUTATIVE_LONG_MSG_SIZE' will never be NULL > [-Waddress] > src/util/cvar/mpich_cvars.c:350:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_BCAST_MIN_PROCS' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:368:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_BCAST_SHORT_MSG_SIZE' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:386:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_BCAST_LONG_MSG_SIZE' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:404:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLGATHER_SHORT_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:422:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLGATHER_LONG_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:440:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_REDUCE_SHORT_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:458:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ALLREDUCE_SHORT_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:476:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_GATHER_VSMALL_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:494:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_GATHER_INTER_SHORT_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:512:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_GATHERV_INTER_SSEND_MIN_PROCS' will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:530:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_SCATTER_INTER_SHORT_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:548:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_ALLGATHERV_PIPELINE_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:566:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_SMP_COLLECTIVES' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:584:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_SMP_BARRIER' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:602:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_SMP_BCAST' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:620:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_MAX_SMP_BCAST_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:638:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_SMP_REDUCE' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:656:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_MAX_SMP_REDUCE_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:674:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_SMP_ALLREDUCE' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:692:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_MAX_SMP_ALLREDUCE_MSG_SIZE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:710:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_COMM_SPLIT_USE_QSORT' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:728:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_DEBUG_HOLD' will never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:746:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_PROCTABLE_SIZE' will > never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:764:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_PROCTABLE_PRINT' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:782:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ENABLE_COLL_FT_RET' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:800:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ABORT_ON_LEAKED_HANDLES' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:818:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CTXID_EAGER_SIZE' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:836:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ASYNC_PROGRESS' will > never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:853:20: warning: assignment discards 'const' > qualifier from pointer target type [enabled by default] > src/util/cvar/mpich_cvars.c:854:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_DEFAULT_THREAD_LEVEL' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:884:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_MEMDUMP' will never be > NULL > [-Waddress] > src/util/cvar/mpich_cvars.c:902:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_ERROR_CHECKING' will > never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:920:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_PRINT_ERROR_STACK' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:938:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CHOP_ERROR_STACK' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:956:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_RMA_ACC_IMMED' will > never > be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:974:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_CH3_RMA_NREQUEST_THRESHOLD' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:992:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_CH3_RMA_NREQUEST_NEW_THRESHOLD' will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1010:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_RMA_LOCK_IMMED' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1028:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_CH3_RMA_MERGE_LOCK_OP_UNLOCK' > will never be NULL [-Waddress] > FCLD lib/libmpichf90.la > src/util/cvar/mpich_cvars.c:1046:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_NOLOCAL' will never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1070:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_ODD_EVEN_CLIQUES' will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1094:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_INTERFACE_HOSTNAME' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1130:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_EAGER_MAX_MSG_SIZE' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1151:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_CH3_PORT_RANGE' will > never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1181:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_NEMESIS_POLLS_BEFORE_YIELD' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1199:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NEMESIS_LMT_DMA_THRESHOLD' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1216:20: warning: assignment discards 'const' > qualifier from pointer target type [enabled by default] > src/util/cvar/mpich_cvars.c:1217:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NEMESIS_NETMOD' will > never be > NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1247:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NEMESIS_SHM_EAGER_MAX_SZ' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1265:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_NEMESIS_SHM_READY_EAGER_MAX_SZ' will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1283:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NEMESIS_ENABLE_CKPOINT' > will > never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1301:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NEMESIS_TCP_NETWORK_IFACE' > will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1337:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_NEMESIS_TCP_HOST_LOOKUP_RETRIES' will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1355:5: warning: the comparison will always > evaluate as 'true' for the address of > 'MPIR_CVAR_NEMESIS_PORTALS_COMM_OVERRIDES' will never be NULL [-Waddress] > src/util/cvar/mpich_cvars.c:1373:5: warning: the comparison will always > evaluate as 'true' for the address of 'MPIR_CVAR_NAMESERV_FILE_PUBDIR' will > never be NULL [-Waddress] > -- > Antonio J. Peña > Postdoctoral Appointee > Mathematics and Computer Science Division > Argonne National Laboratory > 9700 South Cass Avenue, Bldg. 240, Of. 3148 > Argonne, IL 60439-4847 > (+1) 630-252-7928 > apenya(a)mcs.anl.gov > www.mcs.anl.gov/~apenya >

1 0