October 2017 - discuss - mailman3.cels.anl.gov

mpich and hwloc 2
by Samuel Thibault 27 Oct '17

27 Oct '17

Hello, hwloc 2 is to be released sooner or later, and it introduces a few API changes. The build log for version 3.3~a2 of mpich shows: tools/topo/hwloc/topo_hwloc.c:428:40: error: 'HWLOC_MEMBIND_REPLICATE' undeclared (first use in this function); did you mean 'HWLOC_MEMBIND_MIGRATE'? HYDT_topo_hwloc_info.membind = HWLOC_MEMBIND_REPLICATE; ^~~~~~~~~~~~~~~~~~~~~~~ HWLOC_MEMBIND_MIGRATE There might also be other issues. To test against the development tree for hwloc 2, one can use git clone http://github.com/open-mpi/hwloc.git Samuel _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] GCC Compiler Warnings in mpicxx.h when compiling with -Wcast-qual
by Jeff Hammond 22 Oct '17

22 Oct '17

MPI-3 added const in a bunch of places while also deleting the C++ bindings so these issues are not surprising. They’re also innocuous and should be ignored. If they bother you too much, they are very easy to fix and I’m sure a patch would be appreciated. Jeff On Sun, Oct 22, 2017 at 7:39 AM Vincent Bode <vincent(a)jd-bode.de> wrote: > Dear MPICH Discuss, > > > > When compiling a program that includes the header mpi.h with the GCC > Compiler option –Wcast-qual several compiler warnings occur in mpicxx.h: > > > > In file included from /usr/include/mpich/mpi.h:2231:0, > > /usr/include/mpich/mpicxx.h: In member function ‘virtual void > MPI::Datatype::Set_attr(int, const void*)’: > > /usr/include/mpich/mpicxx.h:335:90: warning: cast from type ‘const > void*’ to type ‘void*’ casts away qualifiers [-Wcast-qual] > > MPIX_CALLWORLD( MPI_Type_set_attr( (MPI_Datatype) > the_real_datatype, v2, (void *)v3 )); > > > ^ > > /usr/include/mpich/mpicxx.h:87:17: note: in definition of macro > ‘MPIX_CALLWORLD’ > > { int err = fnc ; if (err) MPIR_Call_world_errhand( err ); } > > ^ > > > > There appear to be 17 compiler warnings of this type: > > > > mpicxx.h:335, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const void*’ to type ‘void*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:633, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const int (*)[3]’ to type ‘int (*)[3]’ casts away > qualifiers [-Wcast-qual] > > > > mpicxx.h:639, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const int (*)[3]’ to type ‘int (*)[3]’ casts away > qualifiers [-Wcast-qual] > > > > mpicxx.h:1087, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const MPI_Request* {aka const int*}’ to type ‘MPI_Request* > {aka int*}’ casts away qualifiers [-Wcast-qual] > > > > mpicxx.h:1412, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const void*’ to type ‘void*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1690, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char*’ to type ‘char*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1691, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char**’ to type ‘char**’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1698, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char*’ to type ‘char*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1699, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char**’ to type ‘char**’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1712, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char**’ to type ‘char**’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1713, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char***’ to type ‘char***’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1713, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const int*’ to type ‘int*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1727, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char**’ to type ‘char**’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1728, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const char***’ to type ‘char***’ casts away qualifiers > [-Wcast-qual] > > mpicxx.h:1728, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const int*’ to type ‘int*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1851, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const void*’ to type ‘void*’ casts away qualifiers > [-Wcast-qual] > > > > mpicxx.h:1872, GNU C Compiler 4 (gcc), Priority: Normal > > cast from type ‘const void*’ to type ‘void*’ casts away qualifiers > [-Wcast-qual] > > > > A minimum working example of just translating the MPICH mpi.h header using > > > > mpicxx.mpich -o /tmp/test.o -c -fPIC -O3 -g -cxx=g++ -Wall -Wextra > -Wcast-qual -Wconversion -Wformat=2 -Wformat-nonliteral -Wformat-security > -Winit-self -Wmissing-format-attribute -Wmissing-include-dirs -Wpacked > -Wunreachable-code -Wunused -Wno-unused-parameter -Wswitch-enum > -Wredundant-decls -pedantic -Wswitch-default -fno-strict-aliasing > -funroll-loops -mfpmath=sse -fopenmp -fno-omit-frame-pointer -msse3 -fPIC > -std=c++11 /usr/include/mpich/mpi.h > > > > produces the same compiler warnings on the following environment (mpich > packages from the debian repository): > > > > debian@debian:~$ mpicxx.mpich -v > > mpicxx for MPICH version 3.2 > > Using built-in specs. > > COLLECT_GCC=g++ > > COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/6/lto-wrapper > > Target: x86_64-linux-gnu > > Configured with: ../src/configure -v --with-pkgversion='Debian 6.3.0-18' > --with-bugurl=file:///usr/share/doc/gcc-6/README.Bugs > --enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr > --program-suffix=-6 --program-prefix=x86_64-linux-gnu- --enable-shared > --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext > --enable-threads=posix --libdir=/usr/lib --enable-nls --with-sysroot=/ > --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes > --with-default-libstdcxx-abi=new --enable-gnu-unique-object > --disable-vtable-verify --enable-libmpx --enable-plugin > --enable-default-pie --with-system-zlib --disable-browser-plugin > --enable-java-awt=gtk --enable-gtk-cairo > --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-6-amd64/jre --enable-java-home > --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-6-amd64 > --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-6-amd64 > --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar > --with-target-system-zlib --enable-objc-gc=auto --enable-multiarch > --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 > --enable-multilib --with-tune=generic --enable-checking=release > --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu > > Thread model: posix > > gcc version 6.3.0 20170516 (Debian 6.3.0-18) > > > > Sincerely, > > Vincent > > > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/ _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

GCC Compiler Warnings in mpicxx.h when compiling with -Wcast-qual
by Vincent Bode 22 Oct '17

22 Oct '17

Dear MPICH Discuss, When compiling a program that includes the header mpi.h with the GCC Compiler option -Wcast-qual several compiler warnings occur in mpicxx.h: In file included from /usr/include/mpich/mpi.h:2231:0, /usr/include/mpich/mpicxx.h: In member function 'virtual void MPI::Datatype::Set_attr(int, const void*)': /usr/include/mpich/mpicxx.h:335:90: warning: cast from type 'const void*' to type 'void*' casts away qualifiers [-Wcast-qual] MPIX_CALLWORLD( MPI_Type_set_attr( (MPI_Datatype) the_real_datatype, v2, (void *)v3 )); ^ /usr/include/mpich/mpicxx.h:87:17: note: in definition of macro 'MPIX_CALLWORLD' { int err = fnc ; if (err) MPIR_Call_world_errhand( err ); } ^ There appear to be 17 compiler warnings of this type: mpicxx.h:335, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const void*' to type 'void*' casts away qualifiers [-Wcast-qual] mpicxx.h:633, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const int (*)[3]' to type 'int (*)[3]' casts away qualifiers [-Wcast-qual] mpicxx.h:639, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const int (*)[3]' to type 'int (*)[3]' casts away qualifiers [-Wcast-qual] mpicxx.h:1087, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const MPI_Request* {aka const int*}' to type 'MPI_Request* {aka int*}' casts away qualifiers [-Wcast-qual] mpicxx.h:1412, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const void*' to type 'void*' casts away qualifiers [-Wcast-qual] mpicxx.h:1690, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char*' to type 'char*' casts away qualifiers [-Wcast-qual] mpicxx.h:1691, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char**' to type 'char**' casts away qualifiers [-Wcast-qual] mpicxx.h:1698, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char*' to type 'char*' casts away qualifiers [-Wcast-qual] mpicxx.h:1699, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char**' to type 'char**' casts away qualifiers [-Wcast-qual] mpicxx.h:1712, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char**' to type 'char**' casts away qualifiers [-Wcast-qual] mpicxx.h:1713, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char***' to type 'char***' casts away qualifiers [-Wcast-qual] mpicxx.h:1713, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const int*' to type 'int*' casts away qualifiers [-Wcast-qual] mpicxx.h:1727, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char**' to type 'char**' casts away qualifiers [-Wcast-qual] mpicxx.h:1728, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const char***' to type 'char***' casts away qualifiers [-Wcast-qual] mpicxx.h:1728, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const int*' to type 'int*' casts away qualifiers [-Wcast-qual] mpicxx.h:1851, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const void*' to type 'void*' casts away qualifiers [-Wcast-qual] mpicxx.h:1872, GNU C Compiler 4 (gcc), Priority: Normal cast from type 'const void*' to type 'void*' casts away qualifiers [-Wcast-qual] A minimum working example of just translating the MPICH mpi.h header using mpicxx.mpich -o /tmp/test.o -c -fPIC -O3 -g -cxx=g++ -Wall -Wextra -Wcast-qual -Wconversion -Wformat=2 -Wformat-nonliteral -Wformat-security -Winit-self -Wmissing-format-attribute -Wmissing-include-dirs -Wpacked -Wunreachable-code -Wunused -Wno-unused-parameter -Wswitch-enum -Wredundant-decls -pedantic -Wswitch-default -fno-strict-aliasing -funroll-loops -mfpmath=sse -fopenmp -fno-omit-frame-pointer -msse3 -fPIC -std=c++11 /usr/include/mpich/mpi.h produces the same compiler warnings on the following environment (mpich packages from the debian repository): debian@debian:~$ mpicxx.mpich -v mpicxx for MPICH version 3.2 Using built-in specs. COLLECT_GCC=g++ COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/6/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 6.3.0-18' --with-bugurl=file:///usr/share/doc/gcc-6/README.Bugs --enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-6 --program-prefix=x86_64-linux-gnu- --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-gnu-unique-object --disable-vtable-verify --enable-libmpx --enable-plugin --enable-default-pie --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-6-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-6-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-6-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --with-target-system-zlib --enable-objc-gc=auto --enable-multiarch --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 6.3.0 20170516 (Debian 6.3.0-18) Sincerely, Vincent _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] progress engine options
by Jeff Hammond 20 Oct '17

20 Oct '17

I think TPE stands for thread progress engine and the question is about the asynchronous progress agent enabled by MPICH_ASYNC_PROGRESS=1. This agent is very high level and I doubt there’s a good way to make it support blocking mode. You’d be better off writing a new async progress engine in the device code. Another idea is to not touch MPICH and instead look at OFI/libfabric, which supports different progress modes and may be more amenable to what you want. You should be able to get pointers from Sean Hefty and friends on the associated email list. Jeff On Fri, Oct 20, 2017 at 4:26 PM Halim Amer <aamer(a)anl.gov> wrote: > Perhaps you could tell us a bit more about > > 1) what you mean by "progress engine to be enabled". Any device > implementation in MPICH has a progress engine enabled all the time. > There is no such thing as enable/disable the progress engine. Removing > the progress engine would simply cause processes to hang > > 2) what TPE stands for > > Halim > www.mcs.anl.gov/~aamer > > On 10/20/17 12:12 PM, Pau Farré wrote: > > Hi > > > > Here at BSC (Antonio Peña's group) we were doing some modifications to > > mpich that require the progress engine to be enabled, but we found that > > the TPE polling was introducing some overhead. > > > > We were wondering if mpich supports a 'blocking' mode for the TPE. And > > if so, It is compatible with Intel Omni-Path? > > > > Thanks. > > Pau Farré > > > > http://bsc.es/disclaimer > > _______________________________________________ > > discuss mailing list discuss(a)mpich.org > > To manage subscription options or unsubscribe: > > https://lists.mpich.org/mailman/listinfo/discuss > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss -- Jeff Hammond jeff.science(a)gmail.com http://jeffhammond.github.io/ _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] progress engine options
by Halim Amer 20 Oct '17

20 Oct '17

Perhaps you could tell us a bit more about 1) what you mean by "progress engine to be enabled". Any device implementation in MPICH has a progress engine enabled all the time. There is no such thing as enable/disable the progress engine. Removing the progress engine would simply cause processes to hang 2) what TPE stands for Halim www.mcs.anl.gov/~aamer On 10/20/17 12:12 PM, Pau Farré wrote: > Hi > > Here at BSC (Antonio Peña's group) we were doing some modifications to > mpich that require the progress engine to be enabled, but we found that > the TPE polling was introducing some overhead. > > We were wondering if mpich supports a 'blocking' mode for the TPE. And > if so, It is compatible with Intel Omni-Path? > > Thanks. > Pau Farré > > http://bsc.es/disclaimer > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

progress engine options
by Pau Farré 20 Oct '17

20 Oct '17

Hi Here at BSC (Antonio Peña's group) we were doing some modifications to mpich that require the progress engine to be enabled, but we found that the TPE polling was introducing some overhead. We were wondering if mpich supports a 'blocking' mode for the TPE. And if so, It is compatible with Intel Omni-Path? Thanks. Pau Farré http://bsc.es/disclaimer _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] Error in documentation : MPI_Igather
by Guo, Yanfei 17 Oct '17

17 Oct '17

Hi Hugo, Thank you for pointing that out. We will fix it in the next release. In the meantime, it is tracked by this issue https://github.com/pmodels/mpich/issues/2793. Best, Yanfei Guo Assistant Computer Scientist MCS Division, ANL On 10/17/17, 4:30 AM, "Hugo TABOADA" <hugo.taboada23(a)gmail.com> wrote: Hello, I want to signal an error in the website documentation. In this URL : https://www.mpich.org/static/docs/v3.2/www3/MPI_Igather.html recvcount number of elements in receive buffer (significant only at root) (non-negative integer)should be : recvcount number of elements for any single receive (integer, significant only at rootWe can see this page 200 of the MPI standard, June 4, 2015. Hugo T. _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Error in documentation : MPI_Igather
by Hugo TABOADA 17 Oct '17

17 Oct '17

Hello, I want to signal an error in the website documentation. In this URL : https://www.mpich.org/static/docs/v3.2/www3/MPI_Igather.html *recvcount *number of elements in receive buffer (significant only at root) (non-negative integer) should be : *recvcount *number of elements for any single receive (integer, significant only at root We can see this page 200 of the MPI standard, June 4, 2015. Hugo T. _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] upgrade from mpich2
by Min Si 13 Oct '17

13 Oct '17

Hi Shuwei, It is hard to answer your question. It is unclear to us what specific OS or network environment you are using. As far as I can tell, hydra works great on most cluster environments. Please also read https://wiki.mpich.org/mpich/index.php/Hydra. You can try the newest version on a few nodes and ping us if you meet any issue. Min On 10/3/17 4:02 PM, Shuwei Zhao wrote: > Dear all MPICH family, > Before my group is using mpich2-1.0.8 which is using mpd as the > message process managenent, now we wanna to migrate to the newest > version which is using hydra as the process manager. Is there > something need to attention when do migration? > Thanks, > Shuwei > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Re: [mpich-discuss] Could not resolve hostname
by Guo, Yanfei 13 Oct '17

13 Oct '17

Hi Fa-Gung, This looks like your network is not correctly configured. The problem is likely to be the /etc/hosts file. Also, I would suggest you to try the latest stable version of MPICH (v3.2). The MPICH-2 v1.4.1 is a very old version. Yanfei Guo Assistant Computer Scientist MCS Division, ANL On 10/12/17, 10:52 PM, "Fa-Gung Fan" <fagung(a)gmail.com> wrote: Hi, I installed mpich2-1.4.1p1 on CentOS 6.9 following the README instruction. Configure, make, and make install steps went fine. But, when running CPI example on local machine, I got: $ pwd /home/ffan/Downloads/mpich2-1.4.1p1 $ mpiexec -n 2 ./examples/cpi ssh: Could not resolve hostname CentOS-6.localdomain: Name or service not known What could have been wrong? Thank you very much for your help. Fa-Gung Fan _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0