- Boost-users - lists.preview.boost.org

[parameter] Parameter-Enabled Class Templates: combining keyword and deduction for the same parameter
by Joaquín Mª López Muñoz 28 Jul '07

28 Jul '07

The docs explain that a parameter can be named: typedef parameter::parameters< required<tag::threading_policy> ... > class_signature; ... class_<threading_policy<single_threaded>,...>; or else be deduced: typedef parameter::parameters< required< deduced<tag::threading_policy>, is_base_and_derived<threading_policy_marker,_> > ... > class_signature; ... class multithreaded: threading_policy_marker{...}; class_<multithreaded,...>; Now, is it possible to combine these two options into one, so that the the two syntaxes class_<threading_policy<single_threaded>,...>; class_<multithreaded,...>; are accepted? If so, how? Thank you, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

5 7

Program Options: parse_config_file: How to ignore unknown options?
by Olaf van der Spek 28 Jul '07

28 Jul '07

Hi, How do I tell this function to ignore unknown options?

2 1

Boost Threads and Cancel Method
by Jeffery Cavallaro 27 Jul '07

27 Jul '07

Per the C++0x discussions earlier in the week, and the calls for the Bjarne, et al to make some definitive decisions regarding a threading API, I would ask the Boost thread developers not to wait and please add a cancel method now. A lot of the scheduling priority stuff can wait, but it is going to be real hard to turn people from pthreads until some sort of a cancel method, however basic, is available. I believe that this is the state of things; however, it not, someone please feel free to correct me. Jeffery Cavallaro Postini Engineering 650-486-8342 ----------------------------------------------------------- This message may contain confidential and/or privileged information. This information is intended to be read only by the individual or entity to whom it is addressed. If you are not the intended recipient, you are on notice that any review, disclosure, copying, distribution or use of the contents of this message is strictly prohibited. If you have received this message in error, please notify the sender immediately and delete or destroy any copy of this message.

1 0

program_options: documentation dead end
by Steven Mackenzie 27 Jul '07

27 Jul '07

I am trying to understand composing, multitoken and zero_tokens methods of the typed_value class. This link http://www.boost.org/doc/html/boost/program_options/typed_value.html#id6986… is to the composing method from the reference page for the typed_value class: it suggests looking at the is_composing documentation, which is unfortunately empty. Can this be fixed for Boost 1.34.1? Also, can someone explain to me what the multitokens method does. The doc states "Specifies that the value can span multiple tokens." What is the definition of a token here? Finally, what does zero_tokens() do? Thanks in advance Steven

8 25

[Important] Boost moving to Subversion on Tuesday, July 31st
by Doug Gregor 27 Jul '07

27 Jul '07

============================================= Boost will be moving to Subversion on Tuesday, July 31st. ============================================= === What's happening === We are converting the existing Boost CVS repository, hosted by SourceForge, over to Subversion. The contents of the CVS repository will join the Boost sandbox and (beta) web site in the Boost Subversion repository, hosted by the Open Systems Laboratory at Indiana University. === Timeline === At 8:00 AM EDT on Tuesday, July 31, we will be cutting off access to the current Boost CVS repository. The Boost Subversion repository, mailing list archives, Boost Trac, beta.boost.org, and anything else hosted at OSL will also be unavailable during this time. We expect to have everything back online by midnight Tuesday. I will send another note to the mailing lists when the conversion is complete. The mailing lists and the main Boost web site should still work during this time. === What you should do === If you have any outstanding changes to Boost in your local CVS tree, commit them before 8:00am EDT on Tuesday. Any changes to your local CVS tree that have not been committed to CVS will be lost and will have to manually be moved over Subversion. The Boost CVS repository will not be reopened. If you have not done so yet, familiarize yourself with Subversion, read the Boost Trac page about the Boost Subversion repository, and, if you have write access to CVS, request a Subversion account: http://svn.boost.org/trac/boost/wiki/BoostSubversion === Who's in charge? === I will be the primary contact person for questions about and problems with the Boost Subversion repository. Most questions should be directed to the Boost mailing list. Account requests should go to all of the moderators at boost-owner(a)lists.boost.org. === How you can help === In several Boost tools and many places in Boost's documentation, we refer to the Boost CVS repository. We need help upgrading these tools and fixing the documentation to refer to the Boost Subversion repository. Specifically: * regression.py, used by regression testers, must be updated to use Subversion * Various developer and release-manager documentation talks about Boost's CVS, but should now refer to Subversion. Better yet: migrate this documentation over to the Boost Trac, where it is easier to maintain and update, and won't be a part of the Boost distribution * Write Subversion-specific documentation on the Trac. Quick-start guides, tips and tricks, etc. would be helpful. One particularly important bit: using svnmerge.py to simplify the handling of release branches. Doug Gregor Open Systems Lab @ Indiana University

1 0

portable lock-free reference counting
by Bryan Green 27 Jul '07

27 Jul '07

Hello, I have a class that contains a reference count. I would like to maintain the count without using a mutex. Can anyone recommend a way to do this portably, using Boost? I see the needed mechanics in 'detail/sp_counted_base.hpp', but since it's in 'detail', I'm not sure what the most portable (and maintainable) way to use it is. Thanks, -Bryan

1 0

boost 1.34.1 cygwin build error.
by chun ping wang 27 Jul '07

27 Jul '07

I change user-config.jam to add "using python : : /usr/bin/python2.5 ;" $ ./bjam --toolset=gcc --prefix=/usr/boost-install install Building Boost.Regex with the optional Unicode/ICU support disabled. Please refer to the Boost.Regex documentation for more information (don't panic: this is a strictly optional feature). /usr/src/boost_1_34_1/tools/build/v2/tools/stage.jam:424: in object(installed-sh ared-lib-generator)(a)4.run *** argument error * rule virtual-target.register ( target ) * called with: ( object(file-target)@1396 object(file-target)@1397 ) * extra argument object(file-target)@1397 /usr/src/boost_1_34_1/tools/build/v2/build/virtual-target.jam:880:see definition of rule 'register' being called /usr/src/boost_1_34_1/tools/build/v2/build/generators.jam:840: in try-one-genera tor-really /usr/src/boost_1_34_1/tools/build/v2/build/generators.jam:902: in try-one-genera tor /usr/src/boost_1_34_1/tools/build/v2/build/generators.jam:1110: in construct-rea lly /usr/src/boost_1_34_1/tools/build/v2/build/generators.jam:1186: in generators.co nstruct /usr/src/boost_1_34_1/tools/build/v2/tools/stage.jam:160: in construct /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:1242: in object(install-t arget-class)(a)85.generate /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:765: in generate-really /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:738: in object(main-targe t)(a)246.generate /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:895: in targets.generate- from-reference /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:1169: in generate-depende ncies /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:1214: in object(alias-tar get-class)(a)89.generate /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:765: in generate-really /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:738: in object(main-targe t)(a)248.generate /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:895: in targets.generate- from-reference /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:1169: in generate-depende ncies /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:1214: in object(alias-tar get-class)(a)97.generate /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:765: in generate-really /usr/src/boost_1_34_1/tools/build/v2/build/targets.jam:738: in object(main-targe t)(a)252.generate /usr/src/boost_1_34_1/tools/build/v2/build-system.jam:402: in load /usr/src/boost_1_34_1/tools/build/v2/kernel/modules.jam:261: in import /usr/src/boost_1_34_1/tools/build/v2/kernel/bootstrap.jam:132: in boost-build /usr/src/boost_1_34_1/boost-build.jam:9: in module scope

2 6

Creating an equivalent to java's System.currentTimeMillis
by Philippe Mondon 27 Jul '07

27 Jul '07

Hi, I'm trying to build an equivalent to java's System.currentTimeMillis() using boost framework, and I've done something like that : #include <boost/date_time/posix_time/posix_time.hpp> using namespace boost::gregorian; using namespace boost::posix_time; //.... Other code long Utils::getCurrentTime() { ptime now(microsec_clock::universal_time()); date d(1970,boost::date_time::Jan,01); ptime torigin(d); //midnight //time_duration tduration = now - torigin; time_period tperiod(torigin,now); //long returnVal = tduration.total_milliseconds(); long returnVal = tperiod.length().total_milliseconds(); cout << "Current time : " << returnVal << endl; return returnVal; } Unfortunately, it gives only a 8 digit number (in java you've got a 13 digit number when using currentTimeMillis) so I suspect that it does not give a duration since Jan the 1st 1970 (which is java reference). Do you have an idea of how to achieve the writing of this piece of code? Thanks, regards. Philippe Mondon Equipe PAC/Architecture DIMC Calyon Ce message et ses pièces jointes (le "message") est destiné à l'usage exclusif de son destinataire. Si vous recevez ce message par erreur, merci d'en aviser immédiatement l'expéditeur et de le détruire ensuite. Le présent message pouvant être altéré à notre insu, CALYON Crédit Agricole CIB ne peut pas être engagé par son contenu. Tous droits réservés. This message and/or any attachments (the "message") is intended for the sole use of its addressee. If you are not the addressee, please immediately notify the sender and then destroy the message. As this message and/or any attachments may have been altered without our knowledge, its content is not legally binding on CALYON Crédit Agricole CIB. All rights reserved.

2 2

[interprocess] Problems on Mac OS X 10.4
by Mark A. Thalman 27 Jul '07

27 Jul '07

We are using the boost interprocess library for some cross-platform interprocess communication. It seems to be working fine on Windows but we are having some issues on the Mac OS (10.4.10, using gcc 4). One thing is that the remove call always deletes the shared memory file, regardless of whether the other process is using it. Does anyone know if this has been tested? Also, the only version I can find is in CVS. Is this where I should be getting the source from, or is there an "official" release somewhere? -- Mark

1 0

[statechart] completion event
by Francisco Santos 27 Jul '07

27 Jul '07

Dear Mr. Andreas Huber, Are completion transitions and events supported by the Statechart Boost library? If so, how do you recommend they should be implemented in code. I have not found examples of statemachines containing transitions *without* explicit triggers in your library's doc page (http://www.boost.org/libs/statechart/doc/index.html) Thank you, Francisco Santos PS: I have included the definition of a completion transition and completion event for reference: A completion transition is a transition without an explicit trigger, although it may have a guard defined. When all transition and entry actions and activities in the currently active state are completed, a completion event instance is generated. This event is the implicit trigger for a completion transition. The completion event is dispatched before any other queued events and has no associated parameters. For instance, a completion transition emanating from a concurrent composite state will be taken automatically as soon as all the concurrent regions have reached their final state. If multiple completion transitions are defined for a state, then they should have mutually exclusive guard conditions. [1] "Completion transitions and completion events", p. 144, Unified Modeling Language Specification, v1.4.2, 2004

2 4

[tests] Signals not well covered in unit tests (in contrast to exceptions)?
by Jens Seidel 27 Jul '07

27 Jul '07

Hi, I have the following problem with a C++ project on Linux using the Boost.test framework: I use the "resource acquisition is initialization" technique to ensure that all resources I allocate are properly released. That's why I have helper classes with a resource allocation in the constructor and deallocation in the destructor. One example of such a helper class is an indentation for my logging code: LOG("blabla"); { LogIndendation indent; // start indentation of output LOG("a message"); divide_by_zero(); LOG("a further message"); } // reset indentation LOG("blabla"); The output is normally: blabla a message a further message blabla Please note that divide_by_zero (which perform an integer division by zero) causes a SIGFPE fatal arithmetic error and the signal handler of Boost.test detects this. Nevertheless the destructor of LogIndendation is *not* called and all further tests I have create wrong indented output (which I compare with the expected output as another test too). So I get: test1 blabla a message unknown location(0): fatal error in "Test1": signal: SIGFPE (arithmetic exception) test2 blabla If I do not generate a signal but throw an exception instead, the destructor is called so this is fine. (Please note that g++ does normally not call a destructor with such a technique if the exception is not catched at all (probably because all resources are expected to be cleaned up, once the process terminates???), but Boost.test catchs all exceptions - good.) I also tried to write my own SIGFPE signal handler which just throws an exception. But this handler is not called, probably it's overwritten by Boost's one. Probably Boost doesn't call the previous handler because it normally terminates the application. But maybe it's possible for Boost to save the initial handler in main() and later to compare the current handler with the initial one. If they are equal handle signals as usual (output an error) otherwise do nothing and call the previous handler? Has anyone a good solution for me, or is there a problem with Boost.test and signals? How to solve this? Jens

1 0

boost 1.34.1 build.bat
by chun ping wang 27 Jul '07

27 Jul '07

For some reason i can no longer specify the toolset to use ...when i run build.bat mingw on my windows machine.. it defaults to msvc. Inside user-config.jam I have using mingw : : C://Dev-Cpp//bin//gcc ; but it keeps using msvc build.

3 3

Any update on a form of boost::threads getting into the C++ 0x standard?
by Graham Reitz 26 Jul '07

26 Jul '07

Our group has had a lot of success with boost::threads and grown comfortable with its use. Thanks, Graham

8 11

Boost with XCode 2.4.1
by Scott Froberg 26 Jul '07

26 Jul '07

Hi All, I'm very new around here. Is anyone using Boost with Apple's XCode 2.4.1? I'm having some trouble with adding the libraries correctly. Thanks, Scott ____________________________________________________________________________________ Get the free Yahoo! toolbar and rest assured with the added security of spyware protection. http://new.toolbar.yahoo.com/toolbar/features/norton/index.php

5 4

[Lambda] Trouble using istreambuf_iterators
by Scott Meyers 26 Jul '07

26 Jul '07

The following code using Lambda with vector iterators compiles without complaint: vector<char> v; find_if(v.begin(), v.end(), _1 == 'x'); Ditto for this code, which uses istream_iterators: istream_iterator<char> b1(cin); istream_iterator<char> e1; find_if(b1, e1, _1 == 'x'); But the same code with istreambuf_iterators gets rejected: istreambuf_iterator<char> b2(cin); istreambuf_iterator<char> e2; find_if(b2, e2, _1 == 'x'); Error messages from VC8 and gcc 4.1.1 are below. Am I doing something wrong, is there a bug in Lambda regarding istreambuf_iterators, or is this a restriction that I didn't see documented somewhere? Thanks, Scott VC8 says: lambda.cpp D:\Apps\Microsoft Visual Studio 8\VC\INCLUDE\algorithm(87) : error C2664: 'bool boost::lambda::lambda_functor<T>::operator ()<char>( A &) const' : cannot convert parameter 1 from 'char' to 'char &' [Instantiation traceback omitted] gcc 4.1.1. says: stl_algo.h:188: error: no match for call to '( boost::lambda::lambda_functor< boost::lambda::lambda_functor_base< boost::lambda::relational_action<boost::lambda::equal_action> , boost::tuples::tuple< boost::lambda::lambda_functor< boost::lambda::placeholder<1> >, const char > > > ) (char)' stl_algo.h: In function 'istreambuf_iterator<char, char_traits<char> > __find_if( istreambuf_iterator<char, char_traits<char> > , istreambuf_iterator<char, char_traits<char> > , boost::lambda::lambda_functor< boost::lambda::lambda_functor_base< boost::lambda::relational_action<boost::lambda::equal_action> , boost::tuples::tuple< boost::lambda::lambda_functor< boost::lambda::placeholder<1> >, const char > > >, input_iterator_tag )': [Instantiation traceback omitted]

4 11

Tuple use/overuse?
by Graham Reitz 26 Jul '07

26 Jul '07

I have been finding that the boost::tuple library is so convenient that I am concerned I may be overusing it. For example, in situations where I might use a struct I am tend to prefer a tuple. Where is a good place to draw the line in situations like this? Or is this the intent to use it as often as possible? Thanks much, Graham

6 14

[mpl] BOOST_MPL_AUX_LAMBDA_SUPPORT for classes without nested ::type
by Joaquín Mª López Muñoz 26 Jul '07

26 Jul '07

Hello, The following program does compile fine on a conformant compiler: #include <boost/mpl/apply.hpp> #include <boost/mpl/aux_/lambda_support.hpp> #include <boost/mpl/assert.hpp> #include <boost/type_traits/is_same.hpp> using namespace boost; using namespace boost::mpl; template<typename T> struct foo { BOOST_MPL_AUX_LAMBDA_SUPPORT(1,foo,(T)) }; template<typename T> struct bar { typedef T type; BOOST_MPL_AUX_LAMBDA_SUPPORT(1,bar,(T)) }; typedef apply1<foo<_1>,int>::type t1; typedef apply1<bar<_1>,int>::type t2; BOOST_MPL_ASSERT((is_same<t1,foo<int> >)); BOOST_MPL_ASSERT((is_same<t2,int>)); int main(){return 0;} Note that foo<_1> applied to int resolves to foo<int> whereas bar<_1> resolves to int, which is correct because the former class does not have a nested ::type while the latter does. This is documented in the MPL reference at: http://www.boost.org/libs/mpl/doc/refmanual/quote.html Now, if I compile the same program with a defective compiler (MSVC 6.0) not supporting PTS, I've got the following: ...\boost\boost\mpl\aux_\preprocessed\msvc60\apply_wrap.hpp(77) : error C2039: 'type' : is not a member of 'result_<int>' ...\boost\mpl\aux_\preprocessed\msvc60\bind.hpp(184) : see reference to class template instantiation 'boost::mpl::apply_wrap1<class foo_rebind,int>' being compiled whereas the code associated with bar<> works OK. So, it seems like BOOST_MPL_AUX_LAMBDA_SUPPORT does not provide lambda support for classes without nested ::type: 1. Is this so by design and/or documented somewhere? 2. If not, could it be fixed? Thank you, Joaquín M López Muñoz Telefónica, Investigación y Desarrollo

2 1

Compile probelms
by alex＠gorex.com.hk 26 Jul '07

26 Jul '07

Hi, I need help on compile Boost.Python. I keep having an error "cannot open file boost_python-vc80-mt-gd-1_34.lib" when I issue a commond bjam -sTOOLS=vc-8_0. I searched some notes from web and I added "import python" or "using python" already. However the problem still exists. anyone can advise?

2 1

Re: [Boost-users] [thread] comparison with Intel Threading BuildingBlocks
by Ovanes Markarian 26 Jul '07

26 Jul '07

On Wed, July 25, 2007 15:54, Andrew Holden wrote: > Toon Knapen wrote: >> I just received (again) some doc on the Intel Threading >> Building Blocks >> (TBB) (http://www.threadingbuildingblocks.org/) >> >> I was wondering how the boost.thread developers/supporters >> would compare TBB with boost.thread (btw: the book of James >> Reinders on TBB mentions boost.thread) ? > > The two libraries are meant to solve different problems. Boost.thread > handles situations where the programmer wishes to explicitly manage > threads, often giving each thread a distinct task to perform. It would > be useful for situations like worker threads in a GUI program. It is > also useful I/O-bound threads, which would cause TBB to waste a > processor core. > > TBB, on the other hand, is useful for situations in which you have a > parallelizable task to perform, and you would rather not worry about > threading details, like a parallel for loop. Do you mean that TBB is more like a "future" which might be added to the upcoming tr2? >[...] With Kind Regards, Ovanes Markarian

3 2

program_options error class
by ALiX 26 Jul '07

26 Jul '07

In the latest implementation of boost (1.34), the error class in boost::program_options is defined as a subclass of std::logic_error. This is, however, not stated in the documentation. In the documentation the class error is described as if it has no superclass from which it inherits. In my programs, can I safely rely on the fact that error inherits from std::logic_error, and so comfortably use the what() function to extract the error messages? Might this change in later implementations? Kind regards, /ALiX

2 3

Fixed sized containers
by Dowie, Martin (SELEX) (UK) 26 Jul '07

26 Jul '07

All, I've noticed the addition of template class "array" to TR1. As I work in the embedded field this is going to be extrememly useful to me (thank you!). I was wondering if there were any plans to expand on this and standardize other statically fixed containers? Cheers -- Martin SELEX Sensors and Airborne Systems Limited Registered Office: Sigma House, Christopher Martin Road, Basildon, Essex SS14 3EL A company registered in England & Wales. Company no. 02426132 ******************************************************************** This email and any attachments are confidential to the intended recipient and may also be privileged. If you are not the intended recipient please delete it from your system and notify the sender. You should not copy it or use it for any purpose nor disclose or distribute its contents to any other person. ********************************************************************

1 0

[statechart] asynchronous machine's post event question.
by Jiang Miao 26 Jul '07

26 Jul '07

Hi, According to the code struct Working:sc::state< Working, Active > { Working(my_context ctx):my_base(ctx) { cout<<__FUNCTION__<<endl; outermost_context_type& machine=outermost_context(); // machine is a asynchronous machine. A machine.post_event( EvtTerminate() ); B machine.my_scheduler().queue_event( machine.my_handle(),boost::intrusive_ptr<EvtTerminate>(new EvtTerminate() ) ); }; ~Working(void) { cout<<__FUNCTION__<<endl; } typedef sc::transition< EvtIdleWork,Idle > reactions; }; Is there any difference between A and B? Which is the better and recommended or Is there another better way to post event? Thank you.

2 1

Re: [Boost-users] Support for sp_counted_base for HP Itanium aCC compiler
by Baruch Zilber 26 Jul '07

26 Jul '07

>Hi Baruch, > >thank you for doing this. > >Can you, please, post your code here (I guess, it is a header file like >sp_counted_base_gcc_ia64.hpp)?. It will be reviewed by the Boost community >and we'll review it at HP. And, then, it can be added to Boost, hopefully, >in Boost 1.35 time frame. > >Thanks again, > Boris I copy the file sp_counted_base_gcc_ia64.hpp to sp_counted_base_acc_ia64.hpp and change the following methods: atomic_increment(),atomic_decrement() and atomic_conditional_increment() The change was: #include <machine/sys/inline.h> inline void atomic_increment( int * pw ) { _Asm_mf(); static_cast<int>(_Asm_fetchadd(_FASZ_W, _SEM_REL, (void*)pw, +1, _LDHINT_NONE) + 1); } inline int atomic_decrement( int * pw ) { _Asm_mf(); return (static_cast<int>(_Asm_fetchadd(_FASZ_W, _SEM_REL, (void*)pw, -1, _LDHINT_NONE) - 1) - 1); } inline int atomic_conditional_increment( int * pw ) { return _Asm_mov_to_ar((_Asm_app_reg)_AREG_CCV, *pw, (_Asm_fence)(_UP_CALL_FENCE | _UP_SYS_FENCE | _DOWN_CALL_FENCE | _DOWN_SYS_FENCE)), _Asm_mf(), (_Asm_cmpxchg((_Asm_sz)4, (_Asm_sem)_SEM_REL, pw, *pw + 1, (_Asm_ldhint)_LDHINT_NONE)); } Thanks Baruch This message and the information contained herein is proprietary and confidential and subject to the Amdocs policy statement, you may review at http://www.amdocs.com/email_disclaimer.asp

2 4

Re: [Boost-users] Tuple use/overuse?
by Roman Perepelitsa 26 Jul '07

26 Jul '07

> Joel de Guzman <joel(a)boost-consulting.com> on Thu, 26 Jul 2007 06:59:46 +0800 > > - bind expression with tuple becomes a nightmare > > Why? Ah access to the members? It's a nightmare indeed. I'd say boost::get is not bindable at all. typedef tuple<int, double> two; vector<two> src; vector<int> dst; // ouch const int & (*get_first)(const cons<two::head_type, two::tail_type> &) = get<0>; transform(src.begin(), src.end(), dst.begin(), get_first); Would not it be cool to write something like this instead? transform(src.begin(), src.end(), dst.begin(), get<0>()); And it's not hard to implement it. namespace boost { namespace tuples { template <int N> struct getter { template <class HT, class TT> typename access_traits< typename element<N, cons<HT, TT> >::type >::non_const_type operator()(cons<HT, TT> & c) const { return get<N>(c); } template <class HT, class TT> typename access_traits< typename element<N, cons<HT, TT> >::type >::const_type operator()(const cons<HT, TT> & c) const { return get<N>(c); } }; } template <int N> tuples::getter<N> get() { return tuples::getter<N>(); } } Does fusion have something like this? Regards, Roman Perepelitsa Deutsche Bank Moscow +7 (495) 660-74-08 --- This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and delete this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden. Please refer to http://www.db.com/en/content/eu_disclosures.htm for additional EU corporate and regulatory disclosures.

1 0

Re: [Boost-users] [thread] comparison with Intel Threading BuildingBlocks
by Ovanes Markarian 26 Jul '07

26 Jul '07

On Wed, July 25, 2007 17:06, Andrew Holden wrote: > Ovanes Markarian wrote: >> Thanks for you link, a very interesting one ;) I ment this >> one in my previous post: >> http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2007/n2276.html > > Another interesting technology. I'll need to examine it more closely. > It looks similar to TBB. Intel's implementation of TBB uses tasks and > thread pools. Instead of futures, it uses canned parallel algorithms > that WILL complete before returning, thus freeing the user from worrying > that values might not be ready. > > It looks like the futures run at a somewhat lower level than TBB, which > is certainly not a bad thing. I can already see the beginnings of some > ideas where it might do things that TBB can't, like any situation where > you wish to launch a separate thread (or task) and forget about it. I'm > also intrigued by the option of waiting an a specific return value from > the thread instead of joining the entire thread, like in boost.thread. > > I find myself wordering if this would be a worthy addition to > boost.thread. It looks like you could use a boost.thread condition > object in the implementation of a future. AFAIK Peter Dimov did some technology evaluation and implementation for futures. I also saw that Howard Hinnant was working on futures. May be there are already some implemenations which prove the concept. Boost Vault contains them at: http://boost-consulting.com/vault/index.php?&direction=0&order=&directory=C… With Kind Regards, Ovanes Markarian

6 5