- Boost - lists.boost.org

[type_erasure] Performance
by Julien Nitard 31 Jul '12

31 Jul '12

Hi all, I took time to write a simple benchmark for boost::any. Please find the code attached. It requires C++11. I could only test with g++ because clang 3.0/Ubuntu segfaults on the code. I'd be more than happy to get feedback on whether the tests are relevant. I think the compiler can't impact the result by optimizing away the payloads because there's the virtual call barrier preventing it from doing so. The results are quite good for a library that is not optimized yet. Here are some typical results on my machine (though results may vary slightly on each run): Times are in seconds for 100 million iterations, using -O3 -std=c++0x and no other special flags. * Classic inheritance replacement: Inherited: 0.261134 Erased: 0.260092 Erased + ctor: 0.343596 Erased + ctor 3 concepts: 0.295584 * boost::function replacement: boost::bind: 0.260356 boost::function: 0.469181 std::function: 0.515477 te::any + boost::bind: 0.494241 te::any + lambda: 0.379945 Basically that means that TypeErasure has a cost that's zero or at worse neglectible when compared to normal inheritance. It probably means that those simple use cases can't be optimized since the theoretical performance is achieved. If you create the any object each time, then there's something like a 30% loss. That's against an empty virtual function which means that for non trivial payloads the cost is still low. Good job ! The only thing I can't explain is why creating an any with 3 concepts in an mpl::vector is faster than a single primitive concepts. Any ideas ? Nice as well is that callable<> matches or bests boost::function in the simple cases I have tested (likely to be common thanks to C++ lambdas). Actually, I guess that the results would be even better in "normal" circumstances because the way I wrote the test any has 2 virtual call to perform. I guess that in that case again, optimization is not required before acceptance and release. Regards, Julien

1 0

[multiprecision] Huge testcase datafile in commit r79802
by Lars Viklund 31 Jul '12

31 Jul '12

Hi list. My boost-commit@ mbox made my mail agent go into a nice crawl today when displaying the commit details for r79802. It turns out that it's about 6.3 MiB large, most of which is a massive 100k line test case in sandbox/big_number/libs/multiprecision/performance/delaunay_data.txt Will this behemoth file end up in the real repository in case of library acceptance, and even worse, will it end up in the deployed tarballs? Are there any guidelines on keeping tests neat, or possibly generating performance test data if it's excessively huge? While in the discussion, has there been any thoughts of separating documentation from the tree, and/or having codegen of the source files that are clearly generated (vector200.hpp, etc.) in the build tree? It's very straining on disk space to have multiple Boost trees lying around, particularly if you're working in some kind of homedir environment. Combating-bloat'ly-yours, -- Lars Viklund | zao(a)acc.umu.se

2 2

[type_erasure]suggest tuple doc contain requires clause specifying T... are unique
by Larry Evans 31 Jul '12

31 Jul '12

http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… Specifies that the T... are placeholders; however, it doesn't specify that they be unique; hence, tuple<concept,_a,_a> x(int(0),std::vector<int>(0)); would satisfy the current description but wouldn't be right. Of course no one would intentionally do that; however, if the Tags are calculated in some meta-program, it might happened accidentally. Hence the requires class should also say, "If T... are not unique, no diagnostic occurs", or something similar. The reason I think "no diagnostic occurs" is that I've looked briefly in tuple.hpp, at the: ::boost::type_erasure::cons<Concept, T...> impl; in the class tuple definition and the cons definition, and I don't see any check for duplicates. -regards, Larry

2 3

[preprocessor metaprogramming] tools
by Nathan Ridge 31 Jul '12

31 Jul '12

Hello, I've noticed preprocessor metaprogramming has become more popular of late, both as a means to provide a cleaner alternative interface to a library (e.g. the DEFINE and ADAPT macros in Boost.Fusion or the proposed macro wrappers for Boost.TypeErasure) and as a means to emulate new language features (e.g. Boost.LocalFunction, Boost.Generic, and Boost.Contract). Sadly, IDE support for writing macro metaprograms is fairly poor in comparison to IDE support for writing regular C++ code. I wanted to share with you something I've found very useful for tracing/debugging macro metaprograms: the Eclipse CDT IDE has a "macro stepper" that allows you to step through the expansion of a macro step-by-step from the initial invocation to the final expansion. (This feature can be accessed by using the "Explore Macro Expansion" context menu item on a macro invocation.) I would be curious to hear what tools you have found useful for writing or debugging macro metaprograms. Regards, Nate

7 11

Re: [boost] BlueGo Beta 0.0.1
by Julian Amann 30 Jul '12

30 Jul '12

> How can I get the sources? the sources can be found at https://bitbucket.org/Vertexwahn/bluego Julian

1 0

[date_time] Who is in charge?
by Philippe Vaucher 30 Jul '12

30 Jul '12

According to the tracker, it looks like the last ticket with activity is https://svn.boost.org/trac/boost/ticket/5257 which is 16 months ago. Also, according to the mailing list I see a lot of discussion/patches dieing out of inactivity ("[boost] [date_time] warning removal", "[boost] [Date Time] bug in documentation example (Localization Demonstration)"). I'm not here to point fingers, we all have busy times where we just cannot handle open-source responsabilities anymore, but maybe this lib should be switched to "no active maintainers" to avoid any confusion? Also, and probably I should have begin with this, is there a page that lists who is the maintainer of what lib beside the Trac? I couldn't find any, but I didn't look very hard either. Philippe

11 18

[type_erasure]any(U,static_binding) stores wrong type
by Larry Evans 30 Jul '12

30 Jul '12

When the attached is run, it produces output: ./static_binding.exe std::is_copy_constructible<ph_erasee_type>=0 53:ph_ctor_arg<boost::type_erasure::_a>::CTOR.default:my_instance=0 creating a_fst from a_ctor_arg 64:ph_ctor_arg<boost::type_erasure::_a>::CTOR.copy:my_instance=1:(void*)this=0x66ab50 :*fst_cast=ph_erasee<boost::type_erasure::_a>.my_instance=0:(void*)fst_cast=0x66ab50 dynamic_cast<ph_ctor_arg_type*>(value_base_ptr)=0x66ab50 dynamic_cast<ph_erasee_type*>(value_base_ptr)=0 Indicating that a ph_ctor_arg_type is being stored in a_fst instead of the expected ph_erasee_type. -regards, Larry

2 6

[1.51][thread] Compile errors when chrono is disabled
by Andrey Semashev 29 Jul '12

29 Jul '12

Hi, There is a problem with Boost.Thread on windows, when BOOST_THREAD_USES_CHRONO macro is defined. In win32/basic_timed_mutex.hpp there are references to Boost.Chrono not guarded with this macro, which results in compile errors. Could this be fixed before the 1.51 release? I can apply the change, if permission is granted.

2 5

[offtopic] C++11 useful trick
by Eric Niebler 29 Jul '12

29 Jul '12

I recently ran into a limitation in recursive functions that make use of decltype in trailing return types. I've come up with a work-around that I think is a little non-obvious, so I thought I'd share. The problem came up while trying to implementing something like a variadic back() function. The "obvious" implementation doesn't work: template<typename T> T back(T && t) { return t; } template<typename T, typename ...U> auto back(T && t, U &&... u) -> decltype(::back(static_cast<U &&>(u)...)) // ERROR { return ::back(static_cast<U &&>(u)...); } int main() { int i = ::back(1,2,3); } The problem occurs on the line marked "HERE". Trouble is, the variadic "back" function is not yet in scope until after the return type. However, we'd like to use it when specifying the return type. The solution uses a default function template parameter to make the ::back symbol dependent, thereby deferring it's lookup until phase 2, when the symbol is visible. Check it: struct S { template<typename T> static T back(T && t) { return t; } template<typename T, typename ...U, typename Impl = S> static auto back(T && t, U &&... u) -> decltype(Impl::back(static_cast<U &&>(u)...)) { return Impl::back(static_cast<U &&>(u)...); } }; int main() { int i = S::back(1,2,3); } The trick is the "typename Impl = S" followed by a (now dependent) invocation of Impl::back. Voila. Apologies if this is a known technique, and I'm spamming the list with old news. It's news to me. -- Eric Niebler BoostPro Computing http://www.boostpro.com

9 26

[type_erasure]broken link in any.html#any(const U&, const static_binding&)
by Larry Evans 29 Jul '12

29 Jul '12

The html page fragment: http://steven_watanabe.users.sourceforge.net/type_erasure/libs/type_erasure… contains a link, CopyConstructible, which sends my browser to: http://steven_watanabe.users.sourceforge.net/trunk/doc/html/CopyConstructib… which is not there. Maybe it should go to: http://www.boost.org/doc/libs/1_50_0/libs/utility/CopyConstructible.html or something similar? Also, why couldn't there be any number of arguments, something like: template<typename Map, class... U> explicit any(const static_binding< Map > & binding, U &&... arg); I think this would then allow any constructor to be used for the contained type. Of course then the requirement would have to be modified to be something like: The underlying type has a constructor accepting arguments, U&... arg. -regards, Larry

1 0