- Boost-users - lists.boost.org

Re: [Boost-users] mpi isend to group
by Andreas Schäfer 31 Dec '12

31 Dec '12

On 23:12 Sun 30 Dec , Philipp Kraus wrote: > Sorry I have forgot a "main information": This calls are a preexecution of > the algorithm. The main algorithm is cycled and uses a MPI blocking > communication, so only the preexecution must be a little bit weak. Sorry, I don't quite get it. Could you elaborate on this? What is in this context the "preexecution", what is the main algorithm, and what do you mean by "must be a little bit weak"? > Thanks for the code, but this code is based on OpenMPI. My program must > be also work wirth MPI CH2 (MPI implemention on Windows based systems), > so I would like to create a boost-only solution. The code uses the MPI standard interface, it contains nothing specific to Open MPI (BTW: it's not called "OpenMPI", but that's just nitpicking). It thus works with MPICH2, too -- no matter if the OS is Linux or Windows. > I do it at the moment with: > > while (thread_is_running) { > > if (!l_mpicom.rank()) > for(std::size_t i=1; i < l_mpicom.size(); ++i) > l_mpicom.isend(i, 666, l_task.getID()); > else > if (boost::optional<mpi::status> l_status = l_mpicom.iprobe(0, 666)) > { > std::size_t l_taskid = 0; > l_mpicom.recv( l_status->source(), l_status->tag(), l_taskid ); > } > } You could rebuild the binary tree communication scheme I've illustrated using Boost MPI (I'm too lazy). I'm pretty sure calling isend() without ever waiting for completion is illegal according to the MPI standard and will result in a memory leak (plus the code won't scale thanks to the for loop). Cheers -Andreas -- ========================================================== Andreas Schäfer HPC and Grid Computing Chair of Computer Science 3 Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany +49 9131 85-27910 PGP/GPG key via keyserver http://www.libgeodecomp.org ========================================================== (\___/) (+'.'+) (")_(") This is Bunny. Copy and paste Bunny into your signature to help him gain world domination!

2 3

Some boost headers use Objective-C keywords
by Reiche, Albert F (Albert) 30 Dec '12

30 Dec '12

I am using boost 1.49.0. An example of this is boost\local.hpp" which includes "boost\locale\message.hpp" which defines "id" (an Objective-C keyword) as a local variable in one of its inline functions. This causes an issue when boost headers are included in any code that is parsed by the Objective-C compiler and makes it impossible to include some boost headers directly in Objective-C source files. Would be nice if you could avoid using Objective-C keywords in all public headers. Albert

5 5

mpi isend to group
by Philipp Kraus 30 Dec '12

30 Dec '12

Hello, I have got a problem to create a mpi::isend call. I have got a thread loop like: while (thread_is_running) { std::size_t id = 0; if (!mpicom.rank()) { try { id = getID(); mpicom.isend(id) } catch (...) { } } else { mpicom.ireceive(id); } if (id > 0) { mpi::barrier(); do something with id } boost::thread::yield(); } I would like to create a non-blocking communication over all nodes, so my node with rank 0 checks if a dataset exists, if it exists, the id should be send to all other hosts. After all hosts reveived this id, they should start the calculation. The nodes must be synchronized before the "working part" is startet, but if there is no data send to the host, they should do nothing. I don't know how to use isend to send a message to all hosts. I know only the mpi::any_source flag but I can not find an any_destination. IMHO the ireceive call must be written with iprobe like if (boost::optional<mpi::status> l_status = mpicom.iprobe( 0, myTag )) { std::size_t id = 0; mpicom.recv( l_status->source(), l_status->tag(), id ) } Is there a way to create a non-blocking group communication? How can I send if data exists, a message to all nodes and receive it on each node also with non-blocking communication? Thanks for your help Phil }

3 5

boost thread with class
by Philipp Kraus 30 Dec '12

30 Dec '12

Hello, I use a thread group for creating a blocking thread calculation. Each thread is created with a boost::bind to a member function, but I would like to create a own class for the thread eg class myThread { public some methods private some properties } The group should be boost::thread_group g; for(int i=0; i < 10; i++) g.add( myThread( i ) ) g.join_all(); Which structure must be implementated for the class myThread and from which class must be derived? Thanks Phil

3 5

How to use boost serialization and standard arrays?
by D. R. Evans 30 Dec '12

30 Dec '12

I've spent quite a bit of time searching and reading, but I can't understand why the following is happening. Given a simple function: (non-boost headers) . . . #include <boost/archive/binary_oarchive.hpp> #include <boost/serialization/array.hpp> . . . void archive_data(void) { ofstream ofs(context.archive_name()); boost::archive::binary_oarchive ar(ofs); array<int, 10> test; ar << test; } I get: /usr/include/boost/serialization/access.hpp:118:9: error: ‘struct std::array<int, 10u>’ has no member named ‘serialize’ ---- The documentation says: > This group of headers includes templates which implement serialization for > Standard Library or Boost Library templates. Any program which uses these > templates can invoke serialization of objects of these types just by > including the corresponding header. But it seems that including <boost/serialization/array.hpp> is insufficient. What am I missing? (Of course, in practice I'm not doing this with simple ints, but I've simplified greatly to get to the above example, which still fails.) This is my first attempt at using boost serialization, so I expect I've just misunderstood something quite simple. Doc Evans -- Web: http://www.sff.net/people/N7DR

2 3

socket.is_open returning false when it is in use
by Gonzalo Garramuno 30 Dec '12

30 Dec '12

I have a problem with my program and I traced it to socket.is_open which returns as closed even while the connection is active in my server program. The client, however, can still communicate with that supposedly closed socket. The program is similar to the timeouts example. I can post all the source code if needed.

4 12

compile some code with boost.python by mingw in win7-64bit
by simon zhang 30 Dec '12

30 Dec '12

I decided to make my program compatible with windows environment.But I have very little programming experience on windows.There are some errors need help. Environment: - os: win7-64bit, - ide: codeblocks12.11, - python: Python 2.7.3 Windows X86-64 Installer (Windows AMD64 / Intel 64 / X86-64 binary [1] -- does not include source) - compiler: mingw that come from codeblocks installation package. - boost: boost1.52 I only copy and test this "hello" code that come from ".\boost_1_52_0\libs\python\example\tutorial" Code: #include <iostream>#include <boost/python/module.hpp>#include <boost/python/def.hpp> char const* greet(){ return "hello, world"; } BOOST_PYTHON_MODULE(hello_ext){ using namespace boost::python; def("greet", greet);} using namespace std;int main(){ cout << "Hello world!" << endl; return 0;} Error info: mingw32-g++.exe -Wall -fexceptions -g -ID:\boost\include\boost-1_52 -IC:\Python27\include -c E:\project\snail-MW\test1\main.cpp -o obj\Debug\main.o mingw32-g++.exe -LD:\boost\lib\ -LC:\Python27\libs -o bin\Debug\test1.exe obj\Debug\main.o -static obj\Debug\main.o: In function `inithello_ext': E:/project/snail-MW/test1/main.cpp:11: undefined reference to `boost::python::detail::init_module(char const*, void (*)())' obj\Debug\main.o: In function `ZNK5boost6python9type_info4nameEv': D:/boost/include/boost-1_52/boost/python/type_id.hpp:165: undefined reference to `boost::python::detail::gcc_demangle(char const*)' obj\Debug\main.o: In function `ZNK5boost6python15to_python_valueIRKPKcEclES5_': D:/boost/include/boost-1_52/boost/python/converter/builtin_converters.hpp:161: undefined reference to `boost::python::converter::do_return_to_python(char const*)' ................ It have also do some errors that only compile by the command of "bjam toolset=gcc variant=release " in the station ".\boost_1_52_0\libs\python\example\tutorial".

1 0

thread-safe boost::random_device / boost:mt19937
by Philipp Kraus 28 Dec '12

28 Dec '12

Hello, I'm using boost::random_device & boost::mt19937 for creating random numbers. I would like to use both within different threads, are both object thread-safe, so I can use it without any thread locking? Thanks Phil

3 4

filesystem : home directory
by berenger.morel＠neutralite.org 27 Dec '12

27 Dec '12

Hello. I am would like to use path like "~/.config" in the filesystem. I've searched in documentation and find nothing about it, so I ended by doing something I do not really like, but which sounds to work (I did not tried it on windows, but I my software is not a linux-only one and so I hope people will use it on other OSes. My problem is that I only have an XP partition and there are no development softwares on it. I could add some, but it would mean I need to reboot to test a software, and on an old windows... :/ ), at least on linux. Here are copy/past (with todos, just ignore them) of related portions of my code: =========================== = utils.cpp = =========================== #ifdef WIN32 #define HOME "HOMEPATH" ///\todo check that it should not be %APPDATA% instead of %HOMEPATH% #else #define HOME "HOME" #endif ... std::string getUserConfigDir(void) { ///\todo find a better solution to follow freeDesktop.org's recommmandations std::string homepath(getenv(HOME));///\note flawfinder say I should use getenv with care const fs::path usualUserConfig(homepath + "/.autorealm/config"); if(fs::exists(usualUserConfig)) return usualUserConfig.string(); const fs::path xdgUserConfig(homepath + "/.config/autorealm/config"); if(fs::exists(xdgUserConfig)) return xdgUserConfig.string(); return std::string(); } =========================== Here, the thing I do not like is to need to use the getenv() function, which is portable, but introduces an #ifdef condition. More important, if I had used things like "~/config/" the source code would have been easier to read and would have avoided one string allocation and some concatenations. But maybe they would have been into filesystem... Another problem which will come from the usage of my system is that it only supports WIN32 and linux, maybe Mac OS, maybe other POSIX systems. As you can see, I can only say "maybe" for other OSes, and for non POSIX OSes (is there is any which is not MSW?) I have just no clue. If that feature was supported by a library, at least it would avoid hard-coded OS's dependent stuff. So, is there is a better way to portably use the $HOME shortcut linux users are used ('~')? If it is any other symbol, I do not care at all, too, I just want an unified way to do that. PS: Of course, if you have other remarks about that code snippet or other things, I would be grateful to hear them.

1 0

Fusion variadic vs. fusion numbered vector
by Андрей - 23 Dec '12

23 Dec '12

Hi Is there any converter ( such as boost::fusion::as_vector ) in fusion for convert numbered form type of boost::fusion::vector to variadic ? Another question is why boost::fusion::make_vector ad boost::fusion::as_vector returns numbered forms of vector but boost::fusion::vector_tie - variadic ?

1 0