October 2014 - Boost-users - lists.preview.boost.org

NT² eand Boost.SIMD release 3.1
by Joel FALCOU 06 Oct '14

06 Oct '14

NT2 is a Matlab inspired, high-level API for scientific computation in C++. Boost.SIMD is a candidate library for inclusion in Boost dealing with the portable encpsulation of SIMD instructions sets. This following releases updates NT2 and Boost.SIMD libraries to version 3.1 This new release of NT2 and Boost.SIMD focused on fixing performance issues and add some new support for architecture and parallel algorithms. The main new features include: - Linear Algebra functions can now use GPGPU through the MAGMA library whenever possible. - Reduction and Scan operations now works with more shared memory back-end - Altivec support is now complete for the VMX ISA We wish to thanks Ian Masliah for his work on MAGMA, Antoine Tran Tan for the shared memory support and Pierre Esterie participation to the Altivec ISA completion. Special thanks to Tim Blenchmann and khegeman for their help and patch submisisons throughout this development. The release Changelog is here : https://github.com/MetaScale/nt2/releases/tag/3.1.0 The files can be download from : http://nt2.metascale.org/downloads/ The Documentation (in progress) is here : http://nt2.metascale.org/doc/html/index.html **Next release announcment** This release is the last C++03 compliant version of both NT2 and Boost.SIMD. The next release - 4.0 - will be C++11 only. The new minimal version of supported compilers will be : - g++ 4.6 - Microsoft Visual Studio 12 (2013) - clang 3.4 - icc 12.1 The main reason for this change is to be able to properly fix the compile-time issue that one may experience. It'll also allow us to use some C++11 feature to simplify or streamline some of the user interface. * The NT2 Team

1 0

[bimap] Look ups on the vector_of collection type
by Arindam Mukherjee 05 Oct '14

05 Oct '14

Would it make sense to define a bimap like this: boost::bimap< boost::bimaps::set_of<std::string>, boost::bimaps::vector_of<std::string>> mymap; And then make the right view (vector_of) more amenable to lookups by performing a sort followed by a binary search on it? Thanks.

1 0

Flat multi set operation costs
by Paolo Bolzoni 03 Oct '14

03 Oct '14

Dear list, In a project of mine I am using a sorted vector as a multimap. The point was that insert, delete and retrieval operations arrived in batches (lots of inserts, lots of retrieval, lots of deletes...) so I essentially saved time sorting the vector when a new kind of operation arrived. I recently noticed boost has the flat multiset implementation, so I was wondering do it follows the same idea? If it does I would be nice to kick-out some code and use the library for ease of maintenance. Your faithfully, Paolo

2 1

boost python wrapping class with getter/setter with const string& arg and return type
by MM 03 Oct '14

03 Oct '14

Hi class C { public: const std::string& get_name() const; void set_name(const std::string&); private: std::string name_; }; given this class C, how would I expose it to python with the class property name? class_<C>("C"). .add_property("name", &C::get_name, &C::set_name); with the proper includes as in the tutorial, it fails to compile. MM

1 0

[Test] Global Fixtures
by Mauricio Carneiro 03 Oct '14

03 Oct '14

I am trying to use global fixtures in my tests but I may be misinterpreting the documentations. The following toy example doesn't work : #include <boost/test/unit_test.hpp> struct TruthFixture { int x = 3; int f (int a) { return a + 3; } }; BOOST_GLOBAL_FIXTURE( TruthFixture ) BOOST_AUTO_TEST_CASE(test_variable) { printf("x = %d\n", x); } BOOST_AUTO_TEST_CASE(test_function) { printf("f(2) = %d\n", f(2)); } the error message I get is the following: test/test.cpp:14:22: error: use of undeclared identifier 'x'; did you mean 'TruthFixture::x'? printf("x = %d\n", x); ^ TruthFixture::x test/test.cpp:4:7: note: 'TruthFixture::x' declared here int x = 3; ^ test/test.cpp:14:22: error: invalid use of non-static data member 'x' printf("x = %d\n", x); ^ test/test.cpp:18:25: error: use of undeclared identifier 'f'; did you mean 'TruthFixture::f'? printf("f(2) = %d\n", f(2)); ^ TruthFixture::f test/test.cpp:6:7: note: 'TruthFixture::f' declared here int f (int a) { ^ test/test.cpp:18:25: error: call to non-static member function without an object argument printf("f(2) = %d\n", f(2)); ^ 4 errors generated. clearly the BOOST_GLOBAL_FIXTURE declaration is not doing anything. But if I drop the global fixture declaration and revert to the per function declaration, it works: #include <boost/test/unit_test.hpp> struct TruthFixture { int x = 3; int f (int a) { return a + 3; } }; BOOST_FIXTURE_TEST_CASE(test_variable, TruthFixture) { printf("x = %d\n", x); } BOOST_FIXTURE_TEST_CASE(test_function, TruthFixture) { printf("f(2) = %d\n", f(2)); } then everything works perfectly as intended. Any input on what I am doing wrong with the *global* fixture declaration? Thank you! Mauricio Carneiro, Ph.D. http://www.broadinstitute.org/~carneiro/

2 1

[Boost] [DateTime] dividing a time_period in N time_periods
by Pablo Madoery 02 Oct '14

02 Oct '14

Hi, I want to make a method which divides a time_period in N time periods, something like this: list<time_period> splitTimePeriod(time_period timePeriod, int n); I actually have made it and although it seems to work it results very inefficient regarding the time it took to divide a large time_period. The method I did is this: list<time_period> splitTimePeriod(time_period timePeriod, int n) { double secsTime = dateTime::diffSeconds(timePeriod.end(), timePeriod.begin()); // tiempo en segundos de cada período de tiempo double ti = secsTime / n; list<time_period> timePeriods; // tiempo de comienzo del primer período de tiempo ptime tpBegin = timePeriod.begin(); // tiempo de fin del último período de tiempo ptime tpLastEnd = timePeriod.end(); for (int i = 0; i < n; i++) { int tUnits = ti * time_duration::ticks_per_second(); // tiempo de comienzo del siguiente período de tiempo ptime tpNextBegin = dateTime::increment(tpBegin, tUnits); // tiempo de fin del período de tiempo ptime tpEnd; if (i == n - 1) { tpEnd = tpLastEnd; } else { tpEnd = dateTime::decrement(tpNextBegin, 1); // guarda para evitar que por cuestiones // de redondeo, el tiempo final de un período // sea mayor que el pasado inicialmente // como argumento de entrada if (tpEnd > tpLastEnd) { tpEnd = tpLastEnd; } } time_period p(tpBegin, tpEnd); // se inserta período de tiempo en la lista timePeriods.push_back(p); // se actualiza el comienzo del siguiente período de tiempo tpBegin = tpNextBegin; } return timePeriods; } Other methods I use are these: double diffSeconds(const ptime &dateA, const ptime &dateB) { time_duration diff = dateA - dateB; return diff.total_seconds(); } ptime increment(const ptime &date, int n) { time_iterator it(date, dateTime::unit); for (int i = 0; i < n; i++) { ++it; } return *it; } ptime decrement(const ptime &date, int n) { time_iterator it(date, dateTime::unit); for (int i = 0; i < n; i++) { --it; } return *it; } const time_duration unit = microseconds(1); ################################################# The question is: is there a better way to accomplish this?

3 6

[mpl] struct definition and mpl sequence in one shot?
by Adam Romanek 01 Oct '14

01 Oct '14

Hi! I have a problem and I'm not sure how to approach it. Currently my code looks like this: struct Definition { typedef boost::mpl::int_<5> A; typedef boost::mpl::int_<3> B; typedef boost::mpl::int_<6> C; typedef boost::mpl::int_<1> D; typedef boost::mpl::vector<A, B, C, D> Seq; }; Here, the number N in mpl::int_< N > denotes some arbitrary decimal number. Then some other code calculates the sum of these numbers up to the type defined by "key", e.g. for Definition::D, the sum is 5 + 3 + 6 (A + B + C). This needs to be done at compile-time. That's why I use mpl::vector and some appropriate meta-programming. I'm wondering whether it would be possible to provide such struct definition without the need to repeat type names in the mpl::vector for Seq type. In other words, I'd probably need a bunch of macros which would allow me to write code like this: struct Definition { FIELD(A, 5); FIELD(B, 3); FIELD(C, 6); FIELD(D, 1); GEN_SEQ() // if needed }; And then Definition::A would still refer to boost::mpl::int_<5>, or would at least allow me to access the boost::mpl::int_<5> somehow, and Definition::Seq would give me the appropriate MPL sequence. Of couse this is just my imagination. The code might look different, I'm just looking for options. Thanks in advance. Adam Romanek

1 0

[Build] -finline-functions no longer supported in clang 3.5
by Mauricio Carneiro 01 Oct '14

01 Oct '14

Hi all, turns out clang 3.5 is no longer supporting <http://llvm.org/klaus/clang/commit/6590426aeb5275ec33dac2877f9349bbbb2d4b2e…> -finline- <http://llvm.org/klaus/clang/commit/6590426aeb5275ec33dac2877f9349bbbb2d4b2e…> functions <http://llvm.org/klaus/clang/commit/6590426aeb5275ec33dac2877f9349bbbb2d4b2e…>, which it claims never did anything anyway, so this is their way to take it out of usage. My release builds (which are done solely through boost build) automatically add this parameter to the compilation command line, and thus is now failing (since I use -Werror). Is there a way to modify the "release" variant not to use that line anymore? Also, is there a way to extend the command line parameters that the release variant will use? Thank you, Mauricio Carneiro, Ph.D. http://www.broadinstitute.org/~carneiro/

1 0

boost::log native syslog sink.
by svante karlsson 01 Oct '14

01 Oct '14

I'm using ubuntu 14.04 with the included boost 1.54 and I'm trying to set up a syslog sink. I've cut this from the samples and added "ident". void init_syslog() { boost::shared_ptr< boost::log::core > core = boost::log::core::get(); // Create a backend boost::shared_ptr< sinks::syslog_backend > backend(new sinks::syslog_backend( keywords::facility = sinks::syslog::local0, keywords::use_impl = sinks::syslog::native, keywords::ident = "mqtt-proxy" )); // Set the straightforward level translator for the "Severity" attribute of type int backend->set_severity_mapper(sinks::syslog::direct_severity_mapping< int >("Severity")); // Wrap it into the frontend and register in the core. // The backend requires synchronization in the frontend. core->add_sink(boost::make_shared< sink_t >(backend)); } The logs end up in syslog but I cant prefix those by program name (and processnumber ) - I've expected the logs to be marked as 'mqtt-proxy' I get logs like Sep 29 03:47:40 ubuntu : ad569c57-780a-431b-bde3-7ede49e7c77f, bridge create but what I would like is similar to Sep 29 03:47:40 ubuntu mqtt-proxy: ad569c57-780a-431b-bde3-7ede49e7c77f, bridge create Am I missing something or is ident not supported? /svante

1 0

[Threads] and OpenMP
by Stephan Menzel 01 Oct '14

01 Oct '14

Hi all, I would like to ask around about a strategy on how to tacke the following threading problem: I am maintaining a boost intensive portable C++ client-server application and recently have included a highly complex 3rd party algorithm into my application that does CPU intense stuff. Long running stuff. Often in the range of days. While at the same time being heavy in OpenMP usage. Now of course requirements are that the user has to be able to cancel this. The algorithm runs in a boost threads controlled background thread. Now I began to place interruption points throughout the 3rd party code but soon had to realize that a lot of CPU time is being handled in OpenMP threaded code regions. So the question is: How can I more or less gracefully shut this down? How do interruption points work in an omp parallel for loop? Do I have a chance at all? My constraints here are a bit forgiving. Given the complexity of the 3rd party code I have already accepted the notion that this is gonna be a memleak monster. Luckily, there's no disk or other hardware access so I'm ready to live with a certain amount of mem not being cleaned up. Yet the backend thread communicates with the spawning class by calling a mutexed callback when it's ready so the user knows it's cancelled. This is protected by a boost mutex as well. And juding from the horrible threading "clogs" that I observed in my first trials I suppose this is gonna be a problem as well. And input is appreciated. And yes, I am aware that this is gonna be an ugly mess. Cheers, Stephan

3 3