August 2015 - discuss - mailman3.cels.anl.gov

Error : Assertion Failed
by Riddhi Modi 11 Aug '15

11 Aug '15

I have one major problem.I established mpich-3.1.4 cluster on 2 machines and run cpi.c program with 2 process which run on 2 different machines.But there is an error like Assertion Failed in file src/mpid/ch3/channels/nemesis/netmod/tcp/socksm.c at line 597: hdr.datalen!=0 Internal ABORT -process 1 how can i solve this kind of error.??/ any idea?? _______________________________________________ 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] static building of mpich
by Balaji, Pavan 09 Aug '15

09 Aug '15

Did you do make and make install? -- Pavan From: Doha Ehab Reply-To: "discuss(a)mpich.org<mailto:[email protected]>" Date: Sunday, August 9, 2015 at 7:24 PM To: "discuss(a)mpich.org<mailto:[email protected]>" Subject: [mpich-discuss] static building of mpich Hello, i'm trying to cross compile mpich , I have extracted mpich-3.1.3 the I created a folder name build inside it . In order to statically build mpich I used the following commands: export CC="/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-gcc" export AR="/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-ar" export CFLAGS="-march=armv5te -static -I/home/doha/buildroot-2015.02/output/host/usr/include/" export LDFLAGS="-L/home/doha/buildroot-2015.02/output/host/usr/lib/" cc=/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-gcc ./configure --prefix=/home/doha/mpich-3.1.3/build/ --host=arm-linux --disable-shared --with-pm=hydra --with-device=ch3:nemesis --disable-fortran but nothing appears in the build folder while it shows no errors on the terminal _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

static building of mpich
by Doha Ehab 09 Aug '15

09 Aug '15

Hello, i'm trying to cross compile mpich , I have extracted mpich-3.1.3 the I created a folder name build inside it . In order to statically build mpich I used the following commands: export CC="/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-gcc" export AR="/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-ar" export CFLAGS="-march=armv5te -static -I/home/doha/buildroot-2015.02/output/host/usr/include/" export LDFLAGS="-L/home/doha/buildroot-2015.02/output/host/usr/lib/" cc=/home/doha/buildroot-2015.02/output/host/usr/bin/arm-linux-gcc ./configure --prefix=/home/doha/mpich-3.1.3/build/ --host=arm-linux --disable-shared --with-pm=hydra --with-device=ch3:nemesis --disable-fortran but nothing appears in the build folder while it shows no errors on the terminal _______________________________________________ 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] Problems with running make
by Balaji, Pavan 06 Aug '15

06 Aug '15

Sorry, windows is not supported. -- Pavan From: hirak chatterjee Reply-To: "discuss(a)mpich.org<mailto:[email protected]>" Date: Thursday, August 6, 2015 at 10:04 AM To: "discuss(a)mpich.org<mailto:[email protected]>" Subject: [mpich-discuss] Problems with running make Sir, I am working with windows 8 (64 bit). After configuring successfully,following the export step ./configure --prefix=/home/hirak/mpich-install --enable-timer-type=gettimeofday CC="gcc-4" CPP="cpp-4" CXX="g++-4" CXXCPP="cpp-4" 2>&1 | tee c.txt I am getting the error repeatedly function 'MPIX_Grequest_class_allocate' part of alias cycle. The text file is attached herewith. Plz help. Hirak -- Hirak Chatterjee<https://www.facebook.com/hirakc1> Research Scholar Department of Chemistry Assam University, Silchar Assam-788011 INDIA Phone No: +918486882543 (ASSAM) +919475265107 (WEST BENGAL) _______________________________________________ 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] [MPICH compilation with NAG Fortran compiler]
by Thakur, Rajeev 06 Aug '15

06 Aug '15

We run nightly tests with the NAG compiler, and they build fine (see the 5th column from the right at http://www.mpich.org/static/cron/tests/) Can you try configuring with just the —prefix option to see if one of the options is the cause. Shared libraries are built by default. Rajeev > On Aug 6, 2015, at 6:25 AM, Maxim Abalenkov <maxim.abalenkov(a)gmail.com> wrote: > > Dear Madam or Sir, > > I’m facing a problem trying to compile MPICH with NAG Fortran compiler. The compilation breaks with an unrecognised option ‘—dynamic lib’ error. Please see the make log file called ‘m.txt’. > > To configure MPICH version 3.2b4 I’m using the attached configure script. Please see the script for the MPICH configuration options. Configuration and compilation log files called ‘c.txt’ and ‘m.txt’ are also attached. > > I would greatly appreciate your help in compiling MPICH with NAG Fortran. Thank you. Have a good day. > > Product name and version: NAG Fortran Compiler Release 6.0 (Hibiya) Build 1021 > OS: Mac OS X Yosemite 10.10.4 (64-bit) > > — > Best wishes, > Maxim Abalenkov > > > > _______________________________________________ > discuss mailing list discuss(a)mpich.org > To manage subscription options or unsubscribe: > https://lists.mpich.org/mailman/listinfo/discuss > <mpich-configure-nag.sh><c.txt><m.txt> _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

Problems with running make
by hirak chatterjee 06 Aug '15

06 Aug '15

Sir, I am working with windows 8 (64 bit). After configuring successfully,following the export step ./configure --prefix=/home/hirak/mpich-install --enable-timer-type=gettimeofday CC="gcc-4" CPP="cpp-4" CXX="g++-4" CXXCPP="cpp-4" 2>&1 | tee c.txt I am getting the error repeatedly function 'MPIX_Grequest_class_allocate' part of alias cycle. The text file is attached herewith. Plz help. Hirak -- *Hirak Chatterjee <https://www.facebook.com/hirakc1>* *Research ScholarDepartment of ChemistryAssam University, SilcharAssam-788011INDIAPhone No: +918486882543 (ASSAM) +919475265107 (WEST BENGAL)* _______________________________________________ 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] resized2.c issues
by Seo, Sangmin 06 Aug '15

06 Aug '15

Hello Mike, As shown in https://trac.mpich.org/projects/mpich/ticket/2088, the issue has been resolved recently. Can you try the current MPICH master or 3.2b4 (http://www.mpich.org/downloads/)? Thanks, Sangmin On Jul 31, 2015, at 1:40 PM, Michael L. Stokes <Michael.Stokes(a)uah.edu<mailto:[email protected]>> wrote: I'm running mpich-3.1.2/test/mpi/io/resized2 with the output file size is 20, should be 16 newbuf[0] is 55, should be 10, newbuf[1] is 10, should be 55, newbuf[2] is 55, should be 55, newbuf[3] is 55, should be 20 Found 2 errors I have run this test case across a number of machines. It is listed as belonging to ticket2088. Can anyone tell me if this is an unresolved issue or some error on my part? Thanks Mike _______________________________________________ discuss mailing list discuss(a)mpich.org<mailto:[email protected]> 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 : Nemesis engine query by Viswanath
by Roy, Hirak 06 Aug '15

06 Aug '15

Hi Viswanath, I have some opportunity to work with both boost and mpi. Just wanted to mention that boost::mpi::environment calls MPI_Init internally. Your program does not work since you have declared boost::mpi::environment as a global object and then you have also called MPI_Init. Either you remove the global objects (env) or you initialize that inside main. Here is the snippet of the program : int main(int argc, char *argv[]) { mpi::environment env(argc, argv, boost::mpi::threading::multiple); mpi::communicator world; if (world.rank() ==0) { cout << "#Master: Thread support :" << env.thread_level() <<endl; } } Thanks, Hirak Balaji, Pavan balaji at anl.gov <mailto:discuss%40mpich.org?Subject=Re%3A%20%5Bmpich-discuss%5D%20Nemesis%20engine%20query%20by%20Viswanath&In-Reply-To=%3C8A837C95-3617-4F5C-9ED1-DFA6840913A5%40anl.gov%3E> Mon Aug 3 10:48:20 CDT 2015 * Previous message: [mpich-discuss] Nemesis engine query by Viswanath<http://lists.mpich.org/pipermail/discuss/2015-August/004112.html> * Next message: [mpich-discuss] question about valgrind errors I am getting<http://lists.mpich.org/pipermail/discuss/2015-August/004113.html> * Messages sorted by: [ date ]<http://lists.mpich.org/pipermail/discuss/2015-August/date.html#4114> [ thread ]<http://lists.mpich.org/pipermail/discuss/2015-August/thread.html#4114> [ subject ]<http://lists.mpich.org/pipermail/discuss/2015-August/subject.html#4114> [ author ]<http://lists.mpich.org/pipermail/discuss/2015-August/author.html#4114> ________________________________ I first tried your program after removing the boost declarations, and it worked correctly: #include<cstdio> #include<cstdlib> #include<cstring> #include<iostream> #include<vector> #include<pthread.h> #include<mpi.h> using namespace std; #define MASTER 0 int main(int argc, char *argv[]) { int rankID, numprocessors,provided; int returnvalue = MPI_Init_thread(&argc,&argv,MPI_THREAD_MULTIPLE,&provided); printf("\n VALUE OF PROVIDED IS %d",provided); printf("\n VALUE OF MPI_THREAD_MULTIPLE IS %d",MPI_THREAD_MULTIPLE); if(provided < MPI_THREAD_MULTIPLE) { printf("\n THREAD LIBRARY DOESN'T HAVE MULTITHREADING SUPPORT:"); } MPI_Finalize(); return 0; } With respect to Boost, yes, MPICH works correctly with Boost. But it is possible that something is not setup correctly in your installation of Boost. Did you make sure you are linking with the correct MPI library? -- Pavan From: Viswanath Krishnamurthy Reply-To: "discuss at mpich.org<https://lists.mpich.org/mailman/listinfo/discuss><mailto:discuss at mpich.org<https://lists.mpich.org/mailman/listinfo/discuss>>" Date: Monday, August 3, 2015 at 8:28 AM To: "discuss at mpich.org<https://lists.mpich.org/mailman/listinfo/discuss><mailto:discuss at mpich.org<https://lists.mpich.org/mailman/listinfo/discuss>>" Subject: Re: [mpich-discuss] Nemesis engine query by Viswanath Hi, Thanks for the previous post on MPI_Init_thread. Here is the sample program which failed #include<cstdio> #include<cstdlib> #include<cstring> #include<iostream> #include<vector> #include<pthread.h> #include<mpi.h> #include <boost/mpi/environment.hpp> #include <boost/mpi/communicator.hpp> #include <boost/serialization/vector.hpp> #include <boost/mpi.hpp> using namespace std; using namespace boost::serialization; using namespace boost::archive; namespace mpi = boost::mpi; // using mpi library under boost namespace mpi::environment env; mpi::communicator world; #define MASTER 0 // Aim of this program is to print the Rank of each process and size of number of MPI Run int main(int argc, char *argv[]) { // Declarations and prototypes int rankID, numprocessors,provided; // Any serial code goes in here // Initialize the MPI Environment over here //MPI_Init(&argc,&argv); int returnvalue = MPI_Init_thread(&argc,&argv,MPI_THREAD_MULTIPLE,&provided); printf("\n VALUE OF PROVIDED IS %d",provided); printf("\n VALUE OF MPI_THREAD_MULTIPLE IS %d",MPI_THREAD_MULTIPLE); if(provided < MPI_THREAD_MULTIPLE) { printf("\n THREAD LIBRARY DOESN'T HAVE MULTITHREADING SUPPORT:"); //exit(1); } // Terminate the MPI Environment MPI_Finalize(); } Error got: Assertion failed in file /home/viswa/libraries/mpich-3.1.4/src/include/mpiimplthreadpost.h at line 163: depth > 0 && depth < 10 internal ABORT - process 1 internal ABORT - process 0 I tried to use boost library in my MPI Code. Is MPI_Init_thread designed to work with Boost libraries?. If not, please suggest an alternative to send vectors(STL types) using MPI. MPI version - 3.1.4 CPU Info: Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-7 Thread(s) per core: 1 Core(s) per socket: 4 Socket(s): 2 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 44 Stepping: 2 CPU MHz: 3458.000 BogoMIPS: 6916.00 Hypervisor vendor: VMware Virtualization type: full L1d cache: 32K L1i cache: 32K L2 cache: 256K L3 cache: 12288K NUMA node0 CPU(s): 0-7 _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0

[MPICH compilation with NAG Fortran compiler]
by Maxim Abalenkov 06 Aug '15

06 Aug '15

1 0

Re: [mpich-discuss] Nemesis engine query by Viswanath
by Balaji, Pavan 03 Aug '15

03 Aug '15

I first tried your program after removing the boost declarations, and it worked correctly: #include<cstdio> #include<cstdlib> #include<cstring> #include<iostream> #include<vector> #include<pthread.h> #include<mpi.h> using namespace std; #define MASTER 0 int main(int argc, char *argv[]) { int rankID, numprocessors,provided; int returnvalue = MPI_Init_thread(&argc,&argv,MPI_THREAD_MULTIPLE,&provided); printf("\n VALUE OF PROVIDED IS %d",provided); printf("\n VALUE OF MPI_THREAD_MULTIPLE IS %d",MPI_THREAD_MULTIPLE); if(provided < MPI_THREAD_MULTIPLE) { printf("\n THREAD LIBRARY DOESN'T HAVE MULTITHREADING SUPPORT:"); } MPI_Finalize(); return 0; } With respect to Boost, yes, MPICH works correctly with Boost. But it is possible that something is not setup correctly in your installation of Boost. Did you make sure you are linking with the correct MPI library? -- Pavan From: Viswanath Krishnamurthy Reply-To: "discuss(a)mpich.org<mailto:[email protected]>" Date: Monday, August 3, 2015 at 8:28 AM To: "discuss(a)mpich.org<mailto:[email protected]>" Subject: Re: [mpich-discuss] Nemesis engine query by Viswanath Hi, Thanks for the previous post on MPI_Init_thread. Here is the sample program which failed #include<cstdio> #include<cstdlib> #include<cstring> #include<iostream> #include<vector> #include<pthread.h> #include<mpi.h> #include <boost/mpi/environment.hpp> #include <boost/mpi/communicator.hpp> #include <boost/serialization/vector.hpp> #include <boost/mpi.hpp> using namespace std; using namespace boost::serialization; using namespace boost::archive; namespace mpi = boost::mpi; // using mpi library under boost namespace mpi::environment env; mpi::communicator world; #define MASTER 0 // Aim of this program is to print the Rank of each process and size of number of MPI Run int main(int argc, char *argv[]) { // Declarations and prototypes int rankID, numprocessors,provided; // Any serial code goes in here // Initialize the MPI Environment over here //MPI_Init(&argc,&argv); int returnvalue = MPI_Init_thread(&argc,&argv,MPI_THREAD_MULTIPLE,&provided); printf("\n VALUE OF PROVIDED IS %d",provided); printf("\n VALUE OF MPI_THREAD_MULTIPLE IS %d",MPI_THREAD_MULTIPLE); if(provided < MPI_THREAD_MULTIPLE) { printf("\n THREAD LIBRARY DOESN'T HAVE MULTITHREADING SUPPORT:"); //exit(1); } // Terminate the MPI Environment MPI_Finalize(); } Error got: Assertion failed in file /home/viswa/libraries/mpich-3.1.4/src/include/mpiimplthreadpost.h at line 163: depth > 0 && depth < 10 internal ABORT - process 1 internal ABORT - process 0 I tried to use boost library in my MPI Code. Is MPI_Init_thread designed to work with Boost libraries?. If not, please suggest an alternative to send vectors(STL types) using MPI. MPI version - 3.1.4 CPU Info: Architecture: x86_64 CPU op-mode(s): 32-bit, 64-bit Byte Order: Little Endian CPU(s): 8 On-line CPU(s) list: 0-7 Thread(s) per core: 1 Core(s) per socket: 4 Socket(s): 2 NUMA node(s): 1 Vendor ID: GenuineIntel CPU family: 6 Model: 44 Stepping: 2 CPU MHz: 3458.000 BogoMIPS: 6916.00 Hypervisor vendor: VMware Virtualization type: full L1d cache: 32K L1i cache: 32K L2 cache: 256K L3 cache: 12288K NUMA node0 CPU(s): 0-7 _______________________________________________ discuss mailing list discuss(a)mpich.org To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

1 0