April 2010 - Boost-users - lists.boost.org

[Serialization] std::set<std::string> restore performance
by Soumen 07 Apr '10

07 Apr '10

Recently we're profiling our tool for restoration for a large electronic design. And we're restoring two archive of same number of elements - one uses vector as container while the other uses set. And it appears that restoration of set is much much slower than the vector. Both the restoration called from same parent which is called just once. And vector restoration took 2% of parent while set restoration took 83% of parent. And most of set restoration time is spent in : boost::serialization::stl::archive_input_unique<boost::archive::binary_iarchive, std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> > >::operator ()(boost::archive::binary_iarchive&, std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> >&, unsigned) Distribution to callers: 128886 times boost::serialization::stl::load_collection<boost::archive::binary_iarchive, std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> >, boost::serialization::stl::archive_input_unique<boost::archive::binary_iarchive, std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> > >, boost::serialization::stl::no_reserve_imp<std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> >> >(boost::archive::binary_iarchive&, std::set<std::string<char, std::char_traits<char>, std::alloca Contributions from descendants: 128886 times (99.98%) boost::archive::detail::basic_iarchive::reset_object_address( const*, const*) 128886 times ( 0.01%) std::set<std::string<char, std::char_traits<char>, std::allocator<char>>, std::less<std::string<char, std::char_traits<char>, std::allocator<char>>>, std::allocator<std::string<char, std::char_traits<char>, std::allocator<char>>> >::insert(std::string<char, std::char_traits<char>, std::allocator<char>> const&) Is there any known performance problem for std::set<std::string> restoration? Or do we need to do some special to overcome this performance problem of std::set<std::string>? Since we need to use frequent lookup while inserting, we can't switch to vector. So, can I get better performance with hash_set or tr1::unordered_set if those are supported for boost::serialization? We're using 1.35 version of boost. Regards, ~ Soumen -- View this message in context: http://old.nabble.com/-Serialization--std%3A%3Aset%3Cstd%3A%3Astring%3E-res… Sent from the Boost - Users mailing list archive at Nabble.com.

2 1

Hitting a mysterious memory limit for shared mutex
by Stefan Stefansson 07 Apr '10

07 Apr '10

Hi. I'm trying to expose a "framebuffer API" via a C interface in a shared library. The idea is to have something like the following scenario: 1. Call "init" to initialize a sharable mutex and a shared memory segment for the framebuffer. 2. A process that wants to read the framebuffer would obtain a shared lock by calling ruauv_camReadLock() and get a handle on the mutex back (as a void*). 3. After doing what it needs to the framebuffer (which usually means copying it) it would call ruauv_camReadUnlock(void*) and send in the handle it got from step 2. I've implemented this and a test application that simply loops and does this repeatedly but the problem I have is that after exactly 65501 iterations I get an error stating: terminate called after throwing an instance of 'boost::interprocess::interprocess_exception' what(): Cannot allocate memory Here is a link to the relevant code that I'm using: http://pastebin.com/4ERTWEzj If it's of any relevance, I'm creating a shared library from the code above the main() in this snippet and then I'm compiling the main() into an executable and linking with the library. As I say, the for loop is able to iterate 65501 times and then fails the next time (the failure occurs at line 33 in the snippet above). I would really appreciate any help. -- Kær kveðja | With best regards Stefán Freyr Stefánsson Tölvunarfræðideild | School of Computer Science Sérfræðingur | Specialist Beint/Direct +354 599 6534 | GSM/Mobile +354 822 1667 Háskólinn í Reykjavík | Reykjavik University stefan(a)ru.is Menntavegur 1, Nauthólsvík | 101 Reykjavík | Iceland Sími/Tel +354 599 6200 | Fax + 354 599 6201 www.ru.is

2 2

bjam buil "hangs"
by Jean-Pierre Bergamin 07 Apr '10

07 Apr '10

Hello everyone I'm trying to build boost 1.42 on XP with Visual Studio 2005 with the following commands in the Visual Studio Command Prompt: > set boost_dir=C:\Programme\boost\boost_1_42_0 > bjam --build-type=complete --with-thread --with-test --with-system --with-serialization --with-regex --with-program_options --with-filesystem --with-date_time --prefix=%boost_dir% --includedir=%boost_dir% --libdir=%boost_dir%\lib --exec-prefix=%boost_dir% install Now bjam "hangs", using 50% CPU (one core fully utilized), after this output: D:\Downloads\boost_1_42_0>bjam --build-type=complete --with-thread --with-test --with-system --with-serialization --with-regex --with-program_options --with-filesystem --with-date_time --prefix=%boost_dir% --includedir=%boost_dir% --libdir=%boost_dir%\lib --exec-prefix=%boost_dir% install WARNING: No python installation configured and autoconfiguration failed. See http://www.boost.org/libs/python/doc/building.html for configuration instructions or pass --without-python to suppress this message and silently skip all Boost.Python targets warning: No toolsets are configured. warning: Configuring default toolset "msvc". warning: If the default is wrong, your build may not work correctly. warning: Use the "toolset=xxxxx" option to override our guess. warning: For more configuration options, please consult warning: http://boost.org/boost-build2/doc/html/bbv2/advanced/configuration.html warning: Building Boost.Regex with the optional Unicode/ICU support disabled. note: Please refer to the Boost.Regex documentation for more information note: this is a strictly optional feature. Interestingly enough, just running "bjam --build-type=complete" without any other options, starts the build as expected. Any hints how to debug this? Regards James

2 4

[spirit][karma+phoenix] Show result of member function
by Ravi 07 Apr '10

07 Apr '10

Hi, How can karma be used to show the result of applying a member function to objects in a container? Consider the following code (which does not compile) to show the real and the imaginary parts of complex numbers stored in a vector: --8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<-- #include <boost/spirit/include/karma.hpp> #include <boost/spirit/include/phoenix_bind.hpp> #include <boost/spirit/include/phoenix_core.hpp> #include <boost/spirit/include/phoenix_operator.hpp> #include <complex> #include <iostream> #include <vector> namespace test { template <typename Container> void show_container( const Container &ctr ) { typedef typename Container::value_type complex_t; typedef typename complex_t::value_type real_t; const real_t& ( complex_t::*real_func )() const = &complex_t::real; const real_t& ( complex_t::*imag_func )() const = &complex_t::imag; namespace karma = boost::spirit::karma; namespace phoenix = boost::phoenix; std::cout << karma::format( ( karma::double_[ // What goes here to extract the real part? karma::_1 = phoenix::bind( real_func, karma::_val ) ] << ',' << karma::double_[ karma::_1 = phoenix::bind( imag_func, karma::_val ) ] ) % karma::eol, ctr ); } } // namespace test int main( int, char *[] ) { std::vector< std::complex<double> > vec( 5 ); test::show_container( vec ); return 0; } --8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<----8<-- The issue is that "karma::_1 = phoenix::bind( real_func, karma::_val )" seems unacceptable to the compiler (gcc 4.4.3, boost 1.42, Linux x86_64). The example above is a simplified version of much more complex code. Regards, Ravi

3 5

Boost.Interprocess: Can I share the multi_index array between processes?
by arm2arm 07 Apr '10

07 Apr '10

Hello, I would like to load multiindex container by one process lets say serverside_loader. And access that array by another clientside_analyser process? How to do that? Lets assume that I have a container: <code> struct particleID { int ID;// real ID for particle from Gadget2 file "ID" block unsigned int IDf;// postition in the file particleID(int id,const unsigned int idf):ID(id),IDf(idf){} bool operator<(const particleID& p)const { return ID<p.ID;} unsigned int getByGID()const {return (ID&0x0FFF);}; }; struct ID{}; struct IDf{}; struct IDg{}; typedef multi_index_container< particleID, indexed_by< ordered_unique< tag<IDf>, BOOST_MULTI_INDEX_MEMBER(particleID,unsigned int,IDf)>, ordered_non_unique< tag<ID>,BOOST_MULTI_INDEX_MEMBER(particleID,int,ID)>, ordered_non_unique< tag<IDg>,BOOST_MULTI_INDEX_CONST_MEM_FUN(particleID,unsigned int,getByGID)> > > particlesID_set; </code> kind regards Arman. -- View this message in context: http://old.nabble.com/Boost.Interprocess%3A-Can-I-share-the-multi_index-arr… Sent from the Boost - Users mailing list archive at Nabble.com.

4 4

Boost & OpenMP
by Robert Jones 07 Apr '10

07 Apr '10

Hi All Some vague navel-gazing follows, please bear with me.... I've recently encountered OpenMP, and wondered if there's any activity within Boost to exploit this. Although OpenMP is new to me I appreciate it's been around a while, so apologies in advance if this is all old hat! In a nutshell as I understand it OpenMP is an open standard for explicity directing the compiler to exploit potential parallelism in the code by embedding #pragmas when statements or loops can be parallelised, and specifying in the pragma which variables are common and which are per execution path. The compiler then sorts out threads and locking etc. We have seen some dramatic performance gains using OpenMP. Ok, some (possibly very naive) thoughts - In some ways this is an alternative to Boost thread pools. - Using explicit pragmas is an intermediate step, and ultimately it might be hoped that a compiler would be able to do all this without external direction. - The whole OpenMP notion is too specific for a general library like Boost It occurs to me that there might be benefit in parallelised standard algorithms (for_each, find, etc), although these might have slightly different behaviour to the existing single thread versions. For example, find() would find any occurance of a value, not necessarily the first. Has any work been done in this area, and would it be of interest to anyone? Are there any known show-stoppers that blow the whole notion out of the water? Thanks for listening, - Rob.

5 5

Important: Server Upgrade Wednesday, April 7
by Doug Gregor 07 Apr '10

07 Apr '10

The main Boost server will be down for a Trac upgrade on Wednesday, April 7 from 8am-9am Eastern time. The following services will be affected: Web server (boost.org) Subversion Trac Mailing list archives Mailing lists should still function as normal. My apologies for the late notice. - Doug

1 0

compile problem with wstring in cross-compilation for ARM
by Jürgen Blümer 07 Apr '10

07 Apr '10

Hi, with boost 1.42.0 I have tried to cross-compile an application on a Kubuntu 9.04 for an ARM processor. The application is processing ini-files, so the file 'boost/property_tree/ptree.hpp' is included. During compilation I got the errors: ../../../boost/boost/property_tree/ptree_fwd.hpp:91: error: 'wstring' is not a member of 'std' ../../../boost/boost/property_tree/string_path.hpp:60: error: expected unqualified-id before '&' token ../../../boost/boost/property_tree/string_path.hpp:60: error: expected ',' or '...' before '&' token ../../../boost/boost/property_tree/string_path.hpp:60: error: ISO C++ forbids declaration of 'parameter' with no type ../../../boost/boost/property_tree/string_path.hpp: In function 'std::string boost::property_tree::detail::dump_sequence(int)': ../../../boost/boost/property_tree/string_path.hpp:62: error: 's' was not declared in this scope By looking in the 'hpp'-files referenced in the error message I found, that in 'boost/property_tree/ptree_fwd.hpp' a '#ifndef BOOST_NO_CWCHAR' is only used on line on line 109 for encapsulating wstring usage. Using this ifndef for encapsulating also the other erroneous lines, the error vanishes and everything is compiled. Is there a reason, why the ifndefs are not inserted at the other position? Thanx for help Jürgen Blümer

1 0

[Boost.UTF] --run_test doesn't match long names
by Adam Nielsen 07 Apr '10

07 Apr '10

Hi all, Is there any documentation on permissible test names? Some of my names don't work correctly with the --run_test parameter. I can put up to 15 characters and end with '*' and the tests will be run, but once I specify the 16th character (with a wildcard or without) it prints the error 'no test cases matching filter'. Is there a length limit on test names? I didn't see any mention of one in the docs and didn't get any compile warnings. If there is a limit, is there an easy way to extend it? Thanks, Adam.

2 2

boost lib search path
by mh00219＠surrey.ac.uk 06 Apr '10

06 Apr '10

hello, I am using boost header files in my program. I need to compile it using gcc 4.3.4. but I dont know how to give the search paths of boost libraries to my linker. please reply me with answers. thank you

2 1