- Boost - lists.boost.org

[thread_pool|range_ex|bind] Extrange behaviour with fibonacci example and parallel_sort
by vicente.botet 07 Feb '09

07 Feb '09

Hi, I have tried to do a parallel_sort using the Boost.ThreadPool library and I get some realy extrange. I have copied the fibonacci example and adapted it to do a parallel sort. See attached file. When I set the defines FIBO this works some times, don't finish others and crash the others. I have added some traces (define COUT) and I get better results but not always correct. I suspected that the threads in the thread pool needs to be stoped. I have added a pool shutdown but nothing changes. Oliver, could you take a look? When I define X_SORT the compilation even do not finish. What is wrong on my definition of parallel_sort? Am I using boost:bind incorrectly? Is there some issue with binding function templates? Can some one help me? Thanks in advance, Vicente P.S. I'm using Boost 1.37 + The file compiles only with the lates version of the Boost.Range library http://www.boostpro.com/vault/index.php?action=downloadfile&filename=range_… and the Boost.ThreadPool version 17 http://www.boostpro.com/vault/index.php?action=downloadfile&filename=boost-… $ gcc --version gcc (GCC) 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) Copyright (C) 2004 Free Software Foundation, Inc. This is free software; see the source for copying conditions. There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. Debug session $ gdb x_sort2 GNU gdb 6.8.0.20080328-cvs (cygwin-special) Copyright (C) 2008 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "i686-pc-cygwin"... (gdb) help List of classes of commands: aliases -- Aliases of other commands breakpoints -- Making program stop at certain points data -- Examining data files -- Specifying and examining files internals -- Maintenance commands obscure -- Obscure features running -- Running the program stack -- Examining the stack status -- Status inquiries support -- Support facilities tracepoints -- Tracing of program execution without stopping the program user-defined -- User-defined commands Type "help" followed by a class name for a list of commands in that class. Type "help all" for the list of all commands. Type "help" followed by command name for full documentation. Type "apropos word" to search for commands related to "word". Command name abbreviations are allowed if unambiguous. (gdb) run Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 3056.0x128] [New thread 3056.0x68c] std::sort: reverse 0..100000 0.05 s [New thread 3056.0xfe4] [New thread 3056.0x1c8] 55 Program exited normally. (gdb) run Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 4028.0xd48] [New thread 4028.0xef4] std::sort: reverse 0..100000 0.03 s [New thread 4028.0x9ec] [New thread 4028.0x5bc] [New thread 4028.0x3bc] [Switching to thread 4028.0x3bc] Quit (gdb) run The program being debugged has been started already. Start it from the beginning? (y or n) y Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 2220.0x4dc] [New thread 2220.0xe58] std::sort: reverse 0..100000 0.05 s [New thread 2220.0x9b0] [New thread 2220.0x778] [New thread 2220.0x264] Program received signal SIGINT, Interrupt. [Switching to thread 2220.0x264] 0x7c87635d in KERNEL32!GetConsoleCharType (Quit (gdb) run The program being debugged has been started already. Start it from the beginning? (y or n) y Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 3964.0xf40] [New thread 3964.0xedc] std::sort: reverse 0..100000 0.05 s [New thread 3964.0x320] [New thread 3964.0xe44] [New thread 3964.0x544] [Switching to thread 3964.0x544] Quit (gdb) q The program is running. Exit anyway? (y or n) y Vicente@viboes1 ~/SVN/bin.v2/vbe_interthreads/libs/interthreads/test/x_sort2.tes t/gcc-3.4.4/debug/link-static/threading-multi $ gdb x_sort2 GNU gdb 6.8.0.20080328-cvs (cygwin-special) Copyright (C) 2008 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html> This is free software: you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. Type "show copying" and "show warranty" for details. This GDB was configured as "i686-pc-cygwin"... (gdb) run Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 1076.0x31c] [New thread 1076.0x9b8] std::sort: reverse 0..100000 0.05 s [New thread 1076.0xaec] [New thread 1076.0xdc0] 19239 [unknown (0xDC0)] x_sort2 1076 _cygtls::handle_exceptions: Exception: ST ATUS_ACCESS_VIOLATION 59227 [unknown (0xDC0)] x_sort2 1076 open_stackdumpfile: Dumping stack trace t o x_sort2.exe.stackdump Program received signal SIGSEGV, Segmentation fault. [Switching to thread 1076.0xdc0] 0x610b6132 in pthread_cond::~pthread_cond () from /usr/bin/cygwin1.dll (gdb) run The program being debugged has been started already. Start it from the beginning? (y or n) y Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 1876.0x8dc] [New thread 1876.0x938] std::sort: reverse 0..100000 0.05 s [New thread 1876.0x624] [New thread 1876.0xe28] 55 Program exited normally. (gdb) run Starting program: /home/Vicente/SVN/bin.v2/vbe_interthreads/libs/interthreads/te st/x_sort2.test/gcc-3.4.4/debug/link-static/threading-multi/x_sort2.exe [New thread 3240.0xef8] [New thread 3240.0xfb8] std::sort: reverse 0..100000 0.05 s [New thread 3240.0x5d4] [New thread 3240.0xe04] [New thread 3240.0x884] [Switching to thread 3240.0x884] Quit (gdb) q The program is running. Exit anyway? (y or n) y

1 0

Re: [boost] [lexical_cast] A suggestion
by Anders Dalvander 07 Feb '09

07 Feb '09

> some bashing snipped... Why not compromise? 1. Make the lexical_cast changes for optional that Andrew Troschinetz suggested. 2. Add the following convert_to.hpp file: #include <boost/lexical_cast.hpp> #include <boost/optional.hpp> namespace boost { template <typename Target, typename Source> inline Target convert_to(const Source& arg) { return lexical_cast< optional<Target> >(arg).get_value_or(Target()); } template <typename Target, typename Source> inline Target convert_to(const Source& arg, const Target& fallback) { return lexical_cast< optional<Target> >(arg).get_value_or(fallback); } } 3. Use convert_to instead of lexical_cast. 4. Everyone happy? Regards, Anders Dalvander

1 0

[config] [random] usage of BOOST_NO_MEMBER_TEMPLATE_FRIENDS
by Steven Watanabe 07 Feb '09

07 Feb '09

AMDG Boost.Random uses BOOST_NO_MEMBER_TEMPLATE_FRIENDS to decide whether it is safe to define the stream operators. #if !defined(BOOST_NO_OPERATORS_IN_NAMESPACE) && !defined(BOOST_NO_MEMBER_TEMPLATE_FRIENDS) template<class CharT, class Traits> friend std::basic_ostream<CharT,Traits>& operator<<(std::basic_ostream<CharT,Traits>& os, const uniform_int& ud); template<class CharT, class Traits> friend std::basic_istream<CharT,Traits>& operator>>(std::basic_istream<CharT,Traits>& is, uniform_int& ud); #endif However, the documentation of BOOST_NO_MEMBER_TEMPLATE_FRIENDS says that BOOST_NO_MEMBER_TEMPLATE_FRIENDS is defined if the compiler doesn't support template<class P> friend class frd; It seems from https://svn.boost.org/trac/boost/ticket/1068 that these do not correspond exactly. Are there any compilers that we still care about, that don't support this use of friend? Also, it sounds to me like the use of BOOST_NO_OPERATORS_IN_NAMESPACE is wrong because it is documented as "Compiler requires inherited operator friend functions to be defined at namespace scope" and there is no inheritance here. Is this right? In Christ, Steven Watanabe

2 1

Re: [boost] [date_time] Time zone improvements
by Rutger ter Borg 07 Feb '09

07 Feb '09

Jeff Garland wrote: > Very cool :-) One question -- is there a portable solution where a user > could deploy a files onto a Windows system or is it a Unix only solution? Portable, big endian binary files. The variant I'm proposing is also based on the POSIX standard http://www.opengroup.org/onlinepubs/000095399/basedefs/xbd_chap08.html: "TZ This variable shall represent timezone information. The contents of the environment variable named TZ shall be used by the [snip] functions, and by various utilities, to override the default timezone. The value of TZ has one of the two forms (spaces inserted for clarity): :characters or: std offset dst offset, rule " The second variant is what's currently in Boost.Date_Time as Posix Time Zone, having a direct specification of the time zone. I'll be talking about the first variant, containing :characters. Although this is implementaiton specific, most POSIX-compatible systems have implemented the Zoneinfo database for this variant, e.g., Linux-based systems, many BSDs, Mac OS X, etc. The :characters point to any other zoneinfo-compatible binary file, either with an absolute or relative path. A natural step for changing Boost.Date_Time would be to make changes such that the Posix Time Zone supports both formats. I.e., variant 1) posix_time_zone( ":/etc/localtime" ); or posix_time_zone( ":/usr/share/zoneinfo/Europe/Amsterdam" ); variant 2) posix_time_zone( "EST-5EDT,M4.1.0,M10.5.0" ); I think both should work as it is from the same standard, and seems to be a natural and consistent starting point to look at changing the interface. Do you agree? > Actually, I believe the abstract timezone is able to support this, but > there > isn't a concrete implementation. It would be great to have the concrete > implementation to prove out the abstraction (obviously from the text below > you think there needs to be changes). As far as I can tell, the abstract time zone class currently is not able to represent variant 1): the abstract zone class is geared towards a single-lined definition of a time zone, whereas variant 1) may be regarded as a series of definitions, i.e., a map of < transition_time_point_in_utc, { gmt_offset, abbreviation, is_dst, ... } > looking up utc_to_local is a matter of doing a lower_bound search on the transition time points. The right-hand side of the map above is widely implemented with a struct called ttinfo. A proposal from my side would be to change the abstract time zone class in such way that it is able to represent both variants. The easiest approach here is probably to introduce the ttinfo structure, and make the current library working with that. variant 1): has a map of transition points, returns a (pointer to a) ttinfo structure. May hold more than 2 different ttinfo structures. variant 2): always has 2 ttinfo structures (std/dst). Uses dst-calculations/functions to lookup the transitions point, returns the right (pointer to a) ttinfo structure. What do you think? Cheers, Rutger

2 3

[filesystem] relative path from arbitrary file system location?
by Kenny Riddile 07 Feb '09

07 Feb '09

Is there any existing functionality for getting a relative path from a full path, where the desired relative path is relative to an arbitrary directory path? If not, I'm assuming I would just have to walk up both full paths until I find their intersection...etc

2 4

[who's using boost] matlap!
by Thorsten Ottosen 07 Feb '09

07 Feb '09

Hi follow boosters, I'm just installing matlap 2008, and looked in the bin/win32 directory. It contains nothing less than boost_date_time-vc80-mt-1_34_1.dll boost_filesystem-vc80-mt-1_34_1.dll boost_graph-vc80-mt-1_34_1.dll boost_iostreams-vc80-mt-1_34_1.dll boost_prg_exec_monitor-vc80-mt-1_34_1.dll boost_program_options-vc80-mt-1_34_1.dll boost_regex-vc80-mt-1_34_1.dll boost_serialization-vc80-mt-1_34_1.dll boost_signals-vc80-mt-1_34_1.dll boost_unit_test_framework-vc80-mt-1_34_1.dll boost_wave-vc80-mt-1_34_1.dll boost_wserialization-vc80-mt-1_34_1.dll :-) best regards Thorsten

5 6

[New Feature Suggestion] Fast RTTI Implementation
by 류인환 07 Feb '09

07 Feb '09

Hello there. Let me introduce the fast rtti system. Many other library and system(ex. game engines) has there own rtti system. But normally their rtti system consume Big O(n) for dynamic casting another type. So, I implemented fast rtti Big O(1) time consuming rtti system. I used bit signature for class inhertance hierarchy expression. Thanks for reviewing my mail and source code

5 4

[filesystem] : basic_path : Feature request
by Rob 06 Feb '09

06 Feb '09

Whilst using filesystem::path (v1.37) , I found I wanted the following functionality: Typedefs added for reference and const_reference and a push_back function, as follows typedef string_type& reference; typedef const string_type& const_reference; inline void push_back(const_reference path) { *this /= path; } This would allow use of std::back_inserter with stl algorithms such as Boost::filesystem::path src("dir1/./dir2"); Boost::filesystem::path target; Std::remove_copy(src.begin(), src.end(), std::back_inserter(target), "."); I would also like to see pop_back() but this may not be as simple to implement, I'll have to check the code Rob PS - My first post so I hope I'm on the right list, right message format, etc.

4 4

Using Boost::Thread in DLL fails on LoadLibrary
by Michael Mzik 06 Feb '09

06 Feb '09

Hi there, I am rather new to Boost and this is my first trouble (win32 debug): I am using Boost::Thread inside a DLL which is loaded by my main program via LoadLibrary(..), but doing so leads to a "Debug Assertion" in "dllinit.cpp" (only in debug configuration!). my DLL links with: #pragma comment( lib, "libboost_thread-vc80-mt-gdp-1_37" ) #pragma comment( lib, "libboost_date_time-vc80-mt-gdp-1_37" ) #pragma comment( lib, "stlportstld.5.0" ) with preprocessor defines: _STLP_DEBUG BOOST_ALL_DYN_LINK BOOST_THREAD_USE_DLL BOOST_LIB_DIAGNOSTIC I am using STLport 5.0.1 and I have boost compiled successfully against STLport (at least I hope so..), I am not using pthreads. No release/debug DLL mixups, all DLLs are available (I checked with the dependency walker). The only thing that bugs me is that there is no "libboost_thread-vc80-mt-gdp-1_37.DLL" generated during boost-build.. but honestly I havent the smallest clue whats going wrong.. Hopefully anyone of you can help me out.. thank a lot in advance and greetings, Michael

1 0

[pool] an O(1) implementation of object_pool::destroy()
by Ben Muzal 06 Feb '09

06 Feb '09

I was looking at object_pool and noticed that there is a way to make object_pool::destroy() an O(1) operation instead of an O(n) operation. If I understand the problem correctly, object_pool::destroy() is O(n) because it keep the free list sorted. It keeps the free list sorted so that object_pool::~object_pool() runs in O(n) time instead of in O(n^2) time. However, I believe there is an implementation of object_pool::~object_pool() that will run in O(n) time without requiring a sorted free list. My suggestion is that ~object_pool() could use a mark-sweep strategy: It would first step through the free list and mark each free chunk as such. It would then go back and examine each chunk in the pool and call the destructor on the ones not marked as free. The each chunk's 'next' pointer could be used to store the marked-as-free flag so that no extra memory is needed. This may be related to Ticket #290. A similar mark-sweep method could be used for pool::release_memory(). The free list would have to be reconstructed afterworlds but that is also easy. Just step back through all of the chunks and connect the ones that are marked as free. While I am on the subject of the pool library, I would also like to see a pool constructor that allows the user to set the initial size of the pool and whether or not the pool was allowed to grow. This would be very useful for the memory limited, real time environments in which memory pools are popular. Any thoughts on these suggestions? If asked, I could help work on an implementation. --Ben

3 3