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

mpfr_float stack allocation
by Daniel Brake 10 May '16

10 May '16

Hi Boost Users, I use Boost.Multiprecision's wrappers for MPFR, particularly the boost::multiprecision::mpfr_float. I've had good success with this library and its types so far, enjoying performance gains because of expression templates, and easy access to mpfr and gmp for floats, ints, and rationals. However, in my complex number type, I am struggling with named real temporaries in arithmetic operators. I am looking for help with multiplication, and I can generalize to all other operations needing a real temporary later. Here's my current code for multiplication, implemented in terms of *=. /** Complex multiplication. uses a single temporary variable 1 temporary, 4 multiplications */ complex& operator*=(const complex & rhs) { mpfr_float re_cache(real_*rhs.real_ - imag_*rhs.imag_); // cache the real part of the result imag_ = real_*rhs.imag_ + imag_*rhs.real_; real_ = re_cache; return *this; } I am using the naive evaluation formula, for better or worse. But the particular method of evaluation doesn't matter to me right now, it's the temporary variable re_cache used to cache the real part is killing me. Profiling using Allinea, gprof, and Callgrind reveals that a significant percentage of time in the program is consumed creating and destroying re_cache, as well as other temporaries scattered throughout my code. In an effort to eliminate heap allocations, I first tried replacing my using mpfr_float = boost::multiprecision::number<boost::multiprecision::mpfr_float_backend<0,boost::multiprecision::allocate_dynamic>, boost::multiprecision::et_off>; with allocate_stack, just for these named temporary variables. However, the documentation <http://www.boost.org/doc/libs/1_60_0/libs/multiprecision/doc/html/boost_mul…> states that allocate_stack only works in fixed precision, so the 0 digits indicating variable precision should cause problems? I expected 0 digits to fail to compile, but compile it does. The conversion from the type with allocate_stack to allocate_dynamic is not provided by the = operator, so I can write re_cache.convert_to<mpfr_float_dynamic>(). But using variable precision with stack allocation appears to cause all sorts of problems. Everything depending on complex multiplication in my program breaks, indicated by massive failures in my test suite. Thus, I currently regard allocate_stack as a non-solution. My question fundamentally regards how to deal with re_cache and the *= operator for complex multiplication. Anyone have ideas on how to get rid of constant heap de/allocation of re_cache without inducing a ton more arithmetic? Using a static for it is not a solution, both because I anticipate multithreading in this application in the future, and the fact that precision will vary though the run, so I'd have to check the precision of re_cache on every evaluation. The C version of the program I am re-implementing used OpenMP for threads, and used a thread-id-indexed global for re_cache. That solution then forces OpenMP onto anyone wanting to use the complex class in multiple threads. Hence, I view this as a non-solution, too. Again, my initial thought was 'heap allocation is the problem here, so I'll use the stack allocated backend'. But variable precision doesn't appear to work with allocate_stack. And statics are no good. Any thoughts? Thanks for your time, Daniel Brake Postdoc University of Notre Dame Applied and Computational Math and Stats

2 1

Re: [Boost-users] [boost] Boost 1.61.0 Release Candidate 2
by Marshall Clow 09 May '16

09 May '16

On Sat, May 7, 2016 at 4:31 AM, Vladimir Prus <vladimir.prus(a)gmail.com> wrote: > > The second release candidates for the 1.61.0 release are now available at: > > http://boost.cowic.de/rc/boost_1_61_0_rc2.zip > http://boost.cowic.de/rc/boost_1_61_0_rc2.7z > http://boost.cowic.de/rc/boost_1_61_0_rc2.tar.bz2 > http://boost.cowic.de/rc/boost_1_61_0_rc2.tar.gz > > The SHA256 checksums are as follows: > > fa1c34862b7ba8674ed6e064a14667a11830c6252f702d9458451834b74f7815 > boost_1_61_0_rc2.7z > a547bd06c2fd9a71ba1d169d9cf0339da7ebf4753849a8f7d6fdb8feee99b640 > boost_1_61_0_rc2.tar.bz2 > a77c7cc660ec02704c6884fbb20c552d52d60a18f26573c9cee0788bf00ed7e6 > boost_1_61_0_rc2.tar.gz > 02d420e6908016d4ac74dfc712eec7d9616a7fc0da78b0a1b5b937536b2e01e8 > boost_1_61_0_rc2.zip > > The changes relative to RC1 are MSVC fixes for signals2 and graph > libraries. > > As always, the release managers would appreciate it if you download > the candidate of your choice and give building it a try. Please report > both success and failure, and anything else that is noteworthy. > > I successfully built the .bz2 release with Apple's clang with C++03/11/14/1z. -- Marshall

1 0

Boost ASIO steady_timer and date change
by Sabino Colonna 09 May '16

09 May '16

Hello, I'm using a steady_timer to perform an asynchronous wait in the following manner: boost::asio::io_service io_service; boost::asio::steady_timer my_timer(io_service); void onTimeout(const boost::system::error_code& e) { // Print something } ... my_timer.expires_from_now(std::chrono::seconds(600)); my_timer.async_wait(onTimeout); io_service.run(); If system 'date' does not change, everything is ok. But if I change 'date' by hand while the timer is running, it expires before 600 seconds (more or less after 5 minutes) and 'onTimeout' is called. With synchronous wait, all is ok. According to documentation, steady_timer should not be affected by time change. Which may be the problem? Thanks, Sabino

2 2

Re: [Boost-users] Compiler limitation (MSVS 2015 update 2) or real error using boost::async?
by James Swift 06 May '16

06 May '16

Hi, in case anyone wants to play with the example I provided but doesn't have access to MSVC you can do so online here. http://rextester.com/BAX37541 kind regards, James

2 2

converting to float128 from double
by Cooper, Bridgette R D 04 May '16

04 May '16

Hi, I have a function that takes as an argument a vector of doubles and tries to convert it to a vector of float128. When I accumulate the values in the float128 vector it looks like it does double additions and only casts up at the last step. The initial vector is defined as const std::vector<double> The float128 vector is defined as std::vector<float128> epoint(num) And I try to do the casting (inside a loop) via: epoint.at(i)=(float128(e8.at(i))); If I do the accumulation on the original vector with no casting into a float128, I get the same as accumulating the vector that should be of float128 type. How should I be doing the casting? Thanks,

4 13

Compiler limitation (MSVS 2015 update 2) or real error using boost::async?
by James Swift 04 May '16

04 May '16

Hi, I'm getting an error with the following code only with Visual C++. (VS 2015 update 2. I also tried preview of VS 15 but it's the same) Clang and GCC are OK with this but perhaps VS is just more strict but correct? The error is deep in boost internals and I can't work out which object is move only and/or has it's copy constructor deleted or if it's true at all! Any help greatly appreciated. kind regards, James Swift --------------------------------- #define BOOST_THREAD_PROVIDES_FUTURE #define BOOST_THREAD_PROVIDES_FUTURE_CONTINUATION #define BOOST_THREAD_PROVIDES_EXECUTORS #include <boost/thread/future.hpp> class TestEx { public: void submit(std::function<void()>) { } bool try_executing_one() { return false; } bool closed() { return false; } void close() { } }; int main() { TestEx tex; boost::async( tex, [](){}); return 0; } ----------------------------------- 1>------ Build started: Project: BoostFuture, Configuration: Debug Win32 ------ 1> BoostFuture.cpp 1>c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3801): error C2280: 'boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>::invoker(const boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type> &)': attempting to reference a deleted function 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\detail\invoker.hpp(731): note: compiler has generated 'boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>::invoker' here 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3800): note: while compiling class template member function 'boost::detail::shared_state_nullary_task<Rp,Cont>::shared_state_nullary_task(const boost::detail::shared_state_nullary_task<Rp,Cont> &)' 1> with 1> [ 1> Rp=Rp 1> ] 1> c:\program files (x86)\microsoft visual studio 14.0\vc\include\xutility(287): note: see reference to function template instantiation 'boost::detail::shared_state_nullary_task<Rp,Cont>::shared_state_nullary_task(const boost::detail::shared_state_nullary_task<Rp,Cont> &)' being compiled 1> with 1> [ 1> Rp=Rp 1> ] 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3858): note: see reference to class template instantiation 'boost::detail::shared_state_nullary_task<Rp,Cont>' being compiled 1> with 1> [ 1> Rp=Rp 1> ] 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3873): note: see reference to function template instantiation 'void boost::detail::future_executor_shared_state<Rp>::init<T,Executor>(Executor &,Fp &&)' being compiled 1> with 1> [ 1> Rp=Rp, 1> T=boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>, 1> Executor=TestEx, 1> Fp=boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type> 1> ] 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3873): note: see reference to function template instantiation 'void boost::detail::future_executor_shared_state<Rp>::init<T,Executor>(Executor &,Fp &&)' being compiled 1> with 1> [ 1> Rp=Rp, 1> T=boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>, 1> Executor=TestEx, 1> Fp=boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type> 1> ] 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boost_1_60\boost\thread\future.hpp(3964): note: see reference to function template instantiation 'boost::future<void> boost::detail::make_future_executor_shared_state<Rp,boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type>,Executor>(Executor &,Fp &&)' being compiled 1> with 1> [ 1> Executor=TestEx, 1> Fp=boost::detail::invoker<main::<lambda_4383c995bef90b35d4f40a119c795fc2>,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type,boost::tuples::null_type> 1> ] 1> c:\users\james\documents\visual studio 2015\projects\boostfuture\boostfuture\boostfuture.cpp(34): note: see reference to function template instantiation 'boost::future<void> boost::async<TestEx,main::<lambda_4383c995bef90b35d4f40a119c795fc2>>(Executor &,F &&)' being compiled 1> with 1> [ 1> Executor=TestEx, 1> F=main::<lambda_4383c995bef90b35d4f40a119c795fc2> 1> ] ========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

1 0

udp::endpoint - port change
by Alexander Carôt 04 May '16

04 May '16

Hello all, probably trivial but I am currently stuck with this: I create a UDP endpoint via udp::resolver::query query(udp::v4(),IP,port); destinationEndpoint = new udp::endpoint(*resolver.resolve(query)); but after the creation I want to change the port without deleting and recreating it with different parameters. How can I do this ? Thanks in advance, best Alex

3 2

Re: [Boost-users] Extending boost graph edge class
by Rafael Gasperetti 03 May '16

03 May '16

Hi, The third option fit in my case. I haver a class that I need to be the edge of the graph, and i have another one that I need to be the vertex. Using this way, My classes Can be used inside the graph to calculate distance, scc, etc. Thanks,Rafael -------- Mensagem original -------- De : Daniel Hofmann <daniel(a)trvx.org> Data: 03/05/2016 19:23 (GMT-03:00) Para: boost-users(a)lists.boost.org Assunto: Re: [Boost-users] Extending boost graph edge class Can you elaborate on "extend the edge class"? Here are some possible interpretations: 1/ You want to optimize the memory consumption for large graphs. The edges and vertices still fit into uint32_t: using graph_t = compressed_sparse_row_graph<directedS, no_property, no_property, no_property, uint32_t /* Vertex */, uint32_t /* EdgeIndex */>; 2/ You want to store arbitrary metadata on your edges and retrieve this edge data based on your graph and a specific edge descriptor: struct edge_data_t { int distance = 0; }; using graph_t = compressed_sparse_row_graph<directedS, no_property, edge_data_t>; // .. auto duration = [&graph](auto edge) { return graph[edge].duration; }; auto total = accumulate(edges(graph) | transformed(duration), 0); This example is using "bundled properties". There is also the possibility to store edge data external to the graph and not couple the edge data's lifetime with the graph (e.g. if you need edge weights for a particular algorithm in your use-case). 3/ You want to provide your own graph representation and make it a model for the BGL graph concepts: http://www.boost.org/doc/libs/1_60_0/libs/graph/doc/leda_conversion.html http://www.boost.org/doc/libs/1_60_0/libs/graph/doc/graph_concepts.html Here are some more BGL examples that should get you started: https://github.com/daniel-j-h/cppnow2016 Cheers, Daniel J H On 05/03/2016 08:37 AM, Rafael Gasperetti wrote: > > > Hi, > My name is Rafael and I need to extend the edge class of de graph part > of the boost lib, but I always find problems trying to do this. Can you > help me with examples and explanations about how to do this? > I read a lotação of places trying to do this but no help was found, > please help me. > > Thanks a lot, > Rafael > > > _______________________________________________ > Boost-users mailing list > Boost-users(a)lists.boost.org > http://lists.boost.org/mailman/listinfo.cgi/boost-users > _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

Extending boost graph edge class
by Rafael Gasperetti 03 May '16

03 May '16

Hi,My name is Rafael and I need to extend the edge class of de graph part of the boost lib, but I always find problems trying to do this. Can you help me with examples and explanations about how to do this?I read a lotação of places trying to do this but no help was found, please help me. Thanks a lot,Rafael

2 1

[boost.multiprecision] print cpp_int with a thousands separator
by Tassilo Glander 03 May '16

03 May '16

Dear List, how can I control the thousands separator behavior when streaming a cpp_int to std::cout? I think it is not possible without changing the global locale, as cpp_int.str uses lexical_cast, which uses the global locale object to read the thousands separator character. This prevents me from having a custom formatting like below. Is there another way? Best, Tassilo #include <iostream> #include <locale> #include <iomanip> int main() { struct my_numpunct : std::numpunct<char> { protected : virtual char do_thousands_sep() const override { return '.' ; } virtual std::string do_grouping() const override { return "\03" ; } }; cpp_int testNum(1000); std::cout.imbue( std::locale( std::cout.getloc(), new my_numpunct ) ) ; std::cout << testNum << '\n' ; // still get 1000, desired is 1.000 }

1 0