- Boost-users - lists.boost.org

How to create distributed_property_map
by Николай Кинаш 05 Jan '14

05 Jan '14

Hello. I try to calculate brandes_betweenness_centrality. #include <boost/graph/use_mpi.hpp> #include <boost/mpi/environment.hpp> #include <boost/mpi/communicator.hpp> #include <boost/graph/distributed/mpi_process_group.hpp> #include <boost/graph/parallel/process_group.hpp> #include <boost/graph/adjacency_list.hpp> #include <boost/graph/distributed/adjacency_list.hpp> #include <boost/graph/small_world_generator.hpp> #include <boost/random/linear_congruential.hpp> #include <boost/graph/distributed/betweenness_centrality.hpp> #include <boost/graph/distributed/graphviz.hpp> #include <boost/graph/rmat_graph_generator.hpp> #include <boost/random/linear_congruential.hpp> #include <iostream> #include <algorithm> #include <math.h> using namespace boost; using namespace graph; using namespace distributed; namespace mpi = boost::mpi; typedef boost::adjacency_list<listS, distributedS<mpi_process_group,vecS>, undirectedS> Graph; typedef boost::unique_rmat_iterator<boost::minstd_rand, Graph> RMATGen; struct no_self_loops { no_self_loops() {} bool operator()(const std::pair<int,int>& e) { return e.first != e.second; } }; int main(int argc, char* argv[]) { boost::mpi::environment env(argc, argv); boost::minstd_rand gen; unsigned int SCALE = atoi(argv[1]); unsigned int EDGE_FACTOR = atoi(argv[2]); unsigned int N = pow(2, SCALE); unsigned int E = EDGE_FACTOR * N; Graph g( boost::make_filter_iterator<no_self_loops> ( RMATGen(gen, N, E, 0.45, 0.15, 0.15, 0.25), RMATGen()), boost::make_filter_iterator<no_self_loops>(RMATGen(), RMATGen()), N); typedef boost::property_map<Graph, vertex_index_t>::type LocalMap; boost::parallel::distributed_property_map<mpi_process_group, LocalMap, ||||||> centrality_map; //brandes_betweenness_centrality(g, centrality_map); } What should be here "|||||" ? I see http://www.boost.org/doc/libs/1_55_0/libs/graph_parallel/doc/html/distribut… template<typename ProcessGroup, typename LocalPropertyMap, typename Key, typename GhostCellS = gc_mapS> Key: The key_type of the distributed property map, which must model the Global Descriptor concept But don`t understand, is not work - typedef property_traits<LocalMap>::key_type key_type; Can you provide example?

1 0

Is there a 1 line wonder for counting iterator and back_inserter - say a counting iterator where the start and step can be specified?
by Brian Davis 05 Jan '14

05 Jan '14

Is there a 1 line wonder for counting iterator and back_inserter - say a counting iterator where the start and step can be specified? The following code creates an vector of angles from zero to N-1 * angle step; std::vector angles; float angle_step = 0.275; int number_angles = 293; std::copy( boost::counting_iterator<float>(0), boost::counting_iterator<float>(number_angles), std::back_inserter(angles) ); std::copy( boost::make_transform_iterator(angles.begin(), boost::bind1st(std::multiplies<float>(), angle_step )), boost::make_transform_iterator(angles.end(), boost::bind1st(std::multiplies<float>(), angle_step )), angles.begin() ); Is there or could there be a counting iterator or other mechanism to do this in 1 line? Say where the start and step could be specified? Thanks

2 1

Re: [Boost-users] Boost-users Digest, Vol 3673, Issue 1
by Markus Lenger 04 Jan '14

04 Jan '14

Hi Larry! The error must be somewhere else in your code. The following program behaves exactly as expected: #include <boost/regex.hpp> #include <string> #include <iostream> int main() { boost::match_results<std::string::const_iterator> match; boost::match_flag_type flags = boost::match_default; boost::regex re(std::string("^XXX_(\\d{3})_\\d{3}$"), boost::regex::perl | boost::regex::icase); std::string the_text = "XXX_123_456"; bool rc = boost::regex_search(the_text,match,re,flags); if (rc) { std::cout << "Number is " << match[1] << std::endl; return 0; } std::cout << "No match found" << std::endl; return 1; } --------- max@switters:~/tmp$ g++ test_regex.cpp -l boost_regex max@switters:~/tmp$ ./a.out Number is 123 Am 2014-01-03 18:00, schrieb boost-users-request(a)lists.boost.org: > Message: 1 > Date: Thu, 2 Jan 2014 16:40:55 -0500 > From: "Larry" <lknain(a)nc.rr.com> > To: <boost-users(a)lists.boost.org> > Subject: [Boost-users] [Regex] Subexpression matching during search > Message-ID: <D06A753E760144CDA0A00B03E0D0DFEE@XV2W> > Content-Type: text/plain; charset="utf-8" > > I have probably missed something in the doc that answers my query but I just can?t see it. > > This is using Perl syntax. > > I had a regex: > > ^(XXX_\d{3,3})_\d{3,3}$ > > It compiles OK but does not match anything in a regex_search. If I remove the parenthesis > > ^XXX_\d{3,3}_\d{3,3}$ > > I get matches. Any clues as to what I am missing? > > > Extract of regex related code lines: > > reg.regex = new boost::regex(reg.reg,boost::regex::perl | boost::regex::icase); > > boost::match_results<string::const_iterator> match; > boost::match_flag_type flags = boost::match_default; > std::string::const_iterator xstart, xend; > > xrc = boost::regex_search(xstart,xend,match,*((*rgi).regex),flags); > > Larry

2 1

[Regex] Subexpression matching during search
by Larry 03 Jan '14

03 Jan '14

I have probably missed something in the doc that answers my query but I just can’t see it. This is using Perl syntax. I had a regex: ^(XXX_\d{3,3})_\d{3,3}$ It compiles OK but does not match anything in a regex_search. If I remove the parenthesis ^XXX_\d{3,3}_\d{3,3}$ I get matches. Any clues as to what I am missing? Extract of regex related code lines: reg.regex = new boost::regex(reg.reg,boost::regex::perl | boost::regex::icase); boost::match_results<string::const_iterator> match; boost::match_flag_type flags = boost::match_default; std::string::const_iterator xstart, xend; xrc = boost::regex_search(xstart,xend,match,*((*rgi).regex),flags); Larry

2 2

[ASIO] Async TCP/UDP server design help
by Terry Lazoki 02 Jan '14

02 Jan '14

I'm having hard time understanding the boost asio design and how i can use it to tackle my problem. My problem: Designing a tcp/udp server that is time efficient. 1. User sends packet through tcp 2. Server receives packet and processes it to build a large tree (n > 20000) 3. Server then parses the new tree and sends a summary packet through udp to another client Conditions/Limitations: - Has to be time efficient, receiving, processing and sending has to be done as fast as possible. - Processing the packet is heavy and can cause delays if its done in the same thread - Parsing the tree is also a little heavy when sending it through udp, can cause delays too Design 1 (2 threads): Thread 1: ASIO TCP and UDP. TCP receives packets and adds it to a circular buffer. While at the same time, processes the shared tree and sends packet through udp. Thread 2: Server grabs packet from circular buffer and processes it Some problems here: - I dont understand the async part here. What happens when tcp receives a packet, if the current thread is continuously parsing the tree and sending packets through udp. Do i need to separate the udp part to another thread? - I have 3 parts of the program that i need to run concurrently, well its really 2 parts and receiving packets through tcp. Any help is appreciated. Thanks,

3 2

Problems installing old BOOST version
by squid squid 01 Jan '14

01 Jan '14

Greetings Everyone I'm a phd student in chemometrics and drug design. I am fairly new to ubuntu C programing although I'm rather experienced in C programing in windows. I have recently come accross a number of codes written in 2005 which I'm interested in learning how they work. Those codes needs BOOST library to complie, however they won't compile on the newest BOOST version installed on my ubuntu 12.04. I set the gcc compiler on lenient so that it ignores all those error messages. The code did compile and ran afterwards. However, when I used GDB debugger to watch how the program flows I noticed that there are likely errors in the way the program runs probably due to using a different BOOST version rather than it's original. Hence I like to install the BOOST version corresponding to the code I downloaded. To do that, I installed Ubuntu 6.02 and downloaded BOOST 1.33.1 both of which were created in late 2005. I downloaded BOOST from: http://www.boost.org/users/history/ but I didnt found any detailed instruction on how to install it. Only vague description on using BOOST jam, I played around with BOOST jam for quite awhile and got roughly the installation outline as follows: -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* Unpack the trazball in /usr/share/ goto the directory /usr/share/boost_1_33_1/tools/build/jam_src run the script ./build.sh the script creats an executable called bjam, copy that file into /usr/share/boost_1_33_1/ run bjam with: ./bjam "-sTOOLS=gcc" install -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* My problems starts when running the ./build.sh script, resulting in tons of error messages and no bjam file created. The errors look like: -*-*-*-*-*-*-*-*-*-*-*-*-*-* ... ... search.c:187: error too many arguments to function 'timestamp' In file included from subst.c:2: jam.h:398:24: error: sys/types.h: No such file or directory jam.h:399:23: error: sys/stat.h: No such file or directory ... ...lots more errors.... -*-*-*-*-*-*-*-*-*-*-*-*-* I would greatly appreciate any info on how to deal with this issue. Sincerely

1 0

Mysterious early data termination using filtering_iostreams zlib_compressor
by Kenneth Adam Miller 31 Dec '13

31 Dec '13

I'm using intel's pin tool (pintool.org) and the boost library's filtering_stream in order to compress data before it goes to a file. I already have functioning code that does that, but at the very end of the execution of the target, right before the PIN_AddFiniFunction callback gets executed the stream of bytes in the compressed version ends (I assume this because the decompressed version is generated by subsequently opening another filtering_stream and using a zlib_decompressor and reading the entire compressed file out). I don't know why the stream ends early... This makes getting help kind of hard, because the files that are being produced are hundreds of megabytes, and differ by just a little kb. So even the diff is large, but I think the salient point is that the data included in the compressed/decompressed version is the same aside from the end. To make this easy for anybody helping, I've included a link to my repo where I've created a reproducible mini example, which should be able to be downloaded, built, and run without any hassel on the latest ubuntu: https://github.com/KennethAdamMiller/Boost-PinTool (It's a drop in replacement for debug trace; you just drop the appropriate files into the DebugTrace folder in your PIN_ROOT/source/tools/, then type make. There are very clear instructions! :) ) link to boost's filtering_stream: http://www.boost.org/doc/libs/1_54_0/libs/iostreams/doc/classes/filtering_s… After you build and run the tool, you should see debugtrace.out and decompressed.xml (the xml part I'm too lazy to change even as I type; it's just text). Go to the bottom of each, you can see that debugtrace.out ends with the line containing "eof" properly, but that the decompressed.xml does not... Can anybody tell me what the issue is? I'm quite certain that everything *should* have went into the FilteringStream that went into the ofstream... please help, I have no idea why it's terminating this way...

1 0

Matrix Vector Multiplication in MPI (Coordinate Storage Format)
by ozhan fenerci 30 Dec '13

30 Dec '13

I wrote a serial code for a matrix-vector multiplication in a Coordinate Storage Format(COO). COO method has already made a big speed-up. My matrix is in a tridiagonal form. I want to parallelize the code and see how the parallelizm improves the code. My matrix could grow up to 2.000.000*2.000.000. Any efficient parallelization approach doesn't come to my mind (I have written some basic mpi programs but not much more. So my experience is limited.) Could you give me some headstart? Thanks. Özhan Fenerci #include <boost/numeric/ublas/banded.hpp> #include <boost/numeric/ublas/io.hpp> #include <vector> #include <iostream> #include "mpi.h" #include <stdio.h> int main(int argc, char *argv[] ) { using namespace boost::numeric::ublas; using std::cout; using std::endl; int numprocs, rank, chunk_size, i,j; int max, mymax,rem; int totalNonZeros; //totalNonZeros of TriDiagonal Matrix MPI_Status status; //b=A*x banded_matrix<double> A (10, 10, 1, 1); //vector<double> x(10); //vector<double> b(10); int* x = new int[10]; int* b = new int[10]; /* Initialize MPI */ MPI_Init( &argc,&argv); MPI_Comm_rank( MPI_COMM_WORLD, &rank); MPI_Comm_size( MPI_COMM_WORLD, &numprocs); //MPI_Status status; if (rank == 0) { for (signed i = 0; i < signed (A.size1 ()); ++ i) { for (signed j = std::max (i - 1, 0); j < std::min (i + 2, signed (A.size2 ())); ++j) A (i, j) = i + j; x[i]= 1; } } if(rank == 0) { cout << A <<endl<<endl; totalNonZeros = 0; for (signed i = 0; i < signed (A.size1 ()); ++ i) { for (signed j = std::max (i - 1, 0); j < std::min (i + 2, signed (A.size2 ())); ++ j) if( A(i,j) !=0) totalNonZeros++; } } cout<<"Number ot NonZeros "<<totalNonZeros<<endl; if(rank == 0) { int * row_idx = new int[totalNonZeros]; // row index array int * col_idx = new int[totalNonZeros]; // col index array int * val = new int[totalNonZeros]; //val array int index=0; for (signed i = 0; i < signed (A.size1 ()); ++ i) { for (signed j = std::max (i - 1, 0); j < std::min (i + 2, signed (A.size2 ())); ++ j) if( A(i,j) !=0) { row_idx[index] = i; col_idx[index]=j; val[index] = A(i,j); index++; } } //Clear out result for ( int k=0; k<A.size1(); ++k) { b[k]= 0; } //Matrix-vector Multiply for (int index=0 ; index <totalNonZeros;++index){ b[row_idx[index]] = b[row_idx[index]]+ val[index]*x[col_idx[index]]; cout<<b[3]<<endl; } //delete row_idx; // delete col_idx; //delete val; } MPI_Finalize(); return 0;

2 1

unit and functional testing of boost statechart
by rahul baruah 29 Dec '13

29 Dec '13

Hello All, I like to do unit and functional testing of our implemented statemachine written using boost - statechart. Is there any framework for that?? anyone did any unit testing and suggestion?? Regards Rahul Protim Baruah

2 1

[Boost] Command Line to run a specific Boost Test Case using BOOST_AUTO_TEST_CASE_TEMPLATE
by Hemanth Choudary M V (RBEI/ETA1) 28 Dec '13

28 Dec '13

Hi, I would like to run this boost test case using the command line. Please note that the list contains data types that have 2 or more words. I need to know what to pass in the Command Line for -run_test=__________ #define BOOST_TEST_MODULE example #include <boost/test/included/unit_test.hpp> #include <boost/test/test_case_template.hpp> #include <boost/mpl/list.hpp> //____________________________________________________________________________// typedef boost::mpl::list<long long,unsigned char> test_types; BOOST_AUTO_TEST_CASE_TEMPLATE( my_test, T, test_types ) { BOOST_CHECK_EQUAL( sizeof(T), (unsigned)4 ); } //_ Mit freundlichen Grüßen / Best Regards, Hemanth Choudary MV RBEI/ETA1

3 2