- Boost - lists.boost.org

[container] Experimental support for memory expansion in vector and burst allocation in node containers
by Ion Gaztañaga 16 Feb '14

16 Feb '14

Hi to all, Some years later than I expected, I've committed to the master branch (after regression tests in develop were ok) a new feature for Boost.Container: new experimental allocators that support memory expansion and burst allocation in order to improve insertion and erasure times in most containers. The interface of these allocators is a variation of the interface I proposed in N2045. The interface is still experimental and not ready to be used by other containers as it's quite complex to handle and I think it can be further improved. Nevertheless, Boost.Container users can benefit from improved performance just using these allocators. The first experiment using this approach can be found in my old article "Applying classic memory allocation strategies to C++ containers" (http://www.drivehq.com/web/igaztanaga/allocplus/) In that article, Boost.Interprocess containers were modified to support N2045 and DLMalloc was version 2.8.3. In the new version the extended system allocator is based on DLMalloc version 2.8.6 (I've noticed a performance improvement in DLMalloc in the new version) and requires linking a new Boost.Container library. Boost.Container is still header-only if extended allocators are not used. DLMalloc is known to be size and speed efficient for non heavily multithreaded environments, but Boost.Container DLMalloc extension is *not* a replacement for multithreaded allocators (say ptmalloc2, ptmalloc3, nedmalloc, tcmalloc, jemalloc...). Your default system allocator will be more efficient than Boost.Container for highly multithreaded environments. On the other hand, at least in Windows and Linux, the allocator seems to be very efficient in non-multithreaded programs. The following extended allocators are provided: * boost::container::allocator: This extended allocator offers expansion, shrink-in place and burst allocation capabilities implemented as a thin wrapper around the modified DLMalloc. It can be used with all containers and it should be the default choice when the programmer wants to use extended allocator capabilities. * boost::container::node_allocator: It's a Simple Segregated Storage allocator, similar to Boost.Pool that takes advantage of the modified DLMalloc burst interface. It does not return memory to the DLMalloc allocator (and thus, to the system), unless explicitly requested. It does offer a very small memory overhead so it's suitable for node containers that allocate very small value_types and it offers improved node allocation times for single node allocations with respect to boost::container::allocator. * boost::container::adaptive_pool: It's a low-overhead node allocator that can return memory to the system. The overhead can be very low (< 5% for small nodes) and it's nearly as fast as boost::container::node_allocator. It's also suitable for node containers. More details on the implementation of this allocator can be found in http://www.drivehq.com/web/igaztanaga/allocplus/#Chapter3. The goal of these extended allocators in Boost.Container is to start building widespread existing practice in regularly requested features from container users (realloc-like behaviour in vector, amortizing several node allocation in range insertions, etc...). I'm not an allocator expert, nor a HPC guru. These new allocators are just the first step of an experiment that I hope real experts will improve. I will appreciate any comment, bug report or suggestion. Thanks, Ion

1 0

Lost the Gmane interface
by Edward Diener 16 Feb '14

16 Feb '14

Normally I post to this list using the Gmane newsgroup interface. The last message I tried to send via this interface received the response from Gmane: "You have tried posting to gmane.comp.lib.boost.devel, which is a unidirectional mailing list. Gmane can therefore not send this message to that mailing list. " Does anybody know what is happening ? I sent a message to a Gmane administrator but so far no response. Edward Diener

2 1

[intrusive] intrusive/detail/has_member_function_callable_with.hpp(74) : fatal error C1189: #error : "BOOST_PP_ITERATION_START() must be zero (0)"
by Vicente J. Botet Escriba 15 Feb '14

15 Feb '14

Hi, there is an issue in has_member_function_callable_with.hpp that make a lot of test fail. See for example the following compiler errors. Could this be fixed? Thanks, Vicente Test output: teeks99-02j-win2008-64on64 - thread - shared_future__copy_assign_p / msvc-12.0 Rev 085b21ee6f062cfc538438496cc5f6ac2da9217f / Fri, 14 Feb 2014 21:34:08 +0000 Compile [2014-02-15 04:21:38 UTC]: fail call "C:\Program Files (x86)\microsoft visual studio 12.0\vc\vcvarsall.bat" x86 >nul cl /Zm800 -nologo @"C:\local\teeks99-02\j\results\boost\bin.v2\libs\thread\test\shared_future__copy_assign_p.test\msvc-12.0\debug\threading-multi\sync\futures\shared_future\copy_assign_pass.obj.rsp" copy_assign_pass.cpp C:\local\teeks99-02\j\boost_root\boost/intrusive/detail/has_member_function_callable_with.hpp(74) : fatal error C1189: #error : "BOOST_PP_ITERATION_START() must be zero (0)" and Test output: teeks99-05f-Ubuntu12.04-64 - thread - async__async_p / clang-linux-3.0 Rev 91a3fc2ca12c3dfdf088383f8b60fb476e222039 / Fri, 14 Feb 2014 23:21:25 +0000 Compile [2014-02-15 02:19:03 UTC]: fail "clang++-3.0" -c -x c++ -Wextra -Wno-delete-non-virtual-dtor -Wno-long-long -Wunused-function -pedantic -O0 -g -fno-inline -Wall -pedantic -g -pthread -fPIC -Wextra -Wno-delete-non-virtual-dtor -Wno-long-long -Wunused-function -pedantic -DBOOST_ALL_NO_LIB=1 -DBOOST_CHRONO_DYN_LINK=1 -DBOOST_SYSTEM_DYN_LINK=1 -DBOOST_SYSTEM_NO_DEPRECATED -DBOOST_THREAD_BUILD_DLL=1 -DBOOST_THREAD_POSIX -DBOOST_THREAD_THROW_IF_PRECONDITION_NOT_SATISFIED -DBOOST_THREAD_USE_DLL=1 -I".." -o "/mnt/teeks99-05/f/results/boost/bin.v2/libs/thread/test/async__async_p.test/clang-linux-3.0/debug/threading-multi/sync/futures/async/async_pass.o" "../libs/thread/test/sync/futures/async/async_pass.cpp" In file included from ../libs/thread/test/sync/futures/async/async_pass.cpp:34: In file included from ../boost/thread/future.hpp:54: In file included from ../boost/thread/detail/memory.hpp:18: In file included from ../boost/thread/csbl/memory/allocator_arg.hpp:16: In file included from ../boost/container/scoped_allocator.hpp:28: In file included from ../boost/container/allocator_traits.hpp:29: In file included from ../boost/container/detail/memory_util.hpp:30: In file included from ../boost/preprocessor/iteration/detail/iter/forward1.hpp:57: ../boost/intrusive/detail/has_member_function_callable_with.hpp:74:5: error: #error "BOOST_PP_ITERATION_START() must be zero (0)" #error "BOOST_PP_ITERATION_START() must be zero (0)" ^ ../boost/intrusive/detail/has_member_function_callable_with.hpp:305:44: error: no template named 'has_member_function_callable_with_allocate_impl'; did you mean 'intrusive::detail::has_member_function_callable_with_pointer_to_impl'? struct BOOST_PP_CAT( BOOST_PP_CAT(has_member_function_callable_with_ ^

3 2

[GSOC] '14 web page minor typo
by zlynus 15 Feb '14

15 Feb '14

Hi all. I'm Lingyu Zhu, a student working on a GSOC proposal. First of all,Thanks to the admin that 2014 GSoC web page has so many great idea and they are all well-documented, especially the section describing the idea of improving boost.afio,which draw my special attention. However, there is a conspicuous typo in the header line: GGoogle Summer of Code 2014 =\ I wonder if the minor typo will effect google's judgement and thus the acceptance of boost org application. So here I am --this is my first 'bug report' ;) Best regards, Lingyu

2 1

[utility] Missing current_function.hpp
by Edward Diener 15 Feb '14

15 Feb '14

From modular boost: switch to develop git pull git submodule update b2 headers No current_function.hpp anywhere switch to master git pull git submodule update b2 headers No current_function.hpp anywhere What am I missing ?

3 2

NT² 3.0 and Boost.SIMD - Stable release
by Joel Falcou 15 Feb '14

15 Feb '14

The first stable release of the 3.x series of MetaScale’s open-source software is available: NT² 3.0. It also includes its spin-off project, Boost.SIMD (not yet a Boost library). Many Issues have been closed since last beta. The main focus of this release cycle was to fix performances issues and to stabilize some parts of the API. Next step will be to separate NT2 and Boost.SIMD into separate repositories, thus making one more step toward Boost.SIMD submission into review. [ http://www.metascale.org/nt%C2%B2-3-0-and-boost-simd ] *Changes & Bug Fixes:* - Major performance improvements in both libraries - Added experimental support for the HPX runtime for parallelization on shared memory systems - Optimizations of many elemental functions, in particuler log/ex, frexp/ldexp and cbrt - More advanced shuffle pattern recognition for SIMD - Fixed issues with Visual Studio 2013 - More functions are force-inlined to ensure good performance - New functions for conversion between types and mixing of vectors of different sizes - Demo applications usable as benchmarks - Better handling of errors when an unsupported function is called - New API for custom memory allocators Full changelog is available at https://github.com/MetaScale/nt2/releases/tag/3.0.0. *How to get it ?* - Source and binary packages can be downloaded here: http://nt2.metascale.org/downloads - The release is also available on Github : https://github.com/MetaScale/nt2/releases - The main master branch is still under development : https://github.com/MetaScale/nt2

1 0

Refactoring tools/regression and tools/release into separate sub-modules
by Beman Dawes 14 Feb '14

14 Feb '14

The tools used by regression testers and regression reporting, and the tools used by release managers are currently part of the boost-super project. As we gain experience with Git and submodules, it is becoming clear these need to be decomposed into two separate submodules. Among other advantages, that would speed work on these tools as we could better utilize volunteers without having to hand out excessive super-project write permissions. We would like one or two volunteers to do the refactoring. You don't need to be a library maintainer or release manager, although that's fine if you are. You do need git skills, as we do want to preserve history. Comments? --The release managers

5 22

Introducing Boost.Optional co-mantainer
by Fernando Cacciola 14 Feb '14

14 Feb '14

Hello, I am pleased to announce that Andrzej Krzemienski will be a co-mantainer for Boost.Optional His goal is to (finally!) implement: 1. Move semantics emulation (C++-3 compatible). 2. Function value() from tr2::optional that either returns a value or throws. For those of you who don't know Andrzej, he is the main force behind the standarizaztion of std::optional Once again, thank you very much Andrzej ! -- Fernando Cacciola SciSoft Consulting, Founder http://www.scisoft-consulting.com

1 0

General Advice Thankyou - now I need a Trac login
by Fletcher, John P 14 Feb '14

14 Feb '14

Thank you for the helpful responses and a lot of new information. I now need a trac login and to be assigned as the maintainer for Phoenix. I have not found instructions for getting a Trac login. Please help. John

3 4

path of test-binary
by dib 14 Feb '14

14 Feb '14

Hi, within the tests of a (not-yet) boost-library, how can I find out the path where b2 placed the binaries of the tests? Problem is the following: Some of the tests need the path to a shared library, which is built together with the tests. In the lib/mylib/tests/Jamfile.v2 I have the following: [...] # our test lib for shared library tests lib test_library : test_library.cpp ; test-suite mylib : # # shared library # [ run shared_library_load_test.cpp ] [ run shared_library_get_symbol_test.cpp ] # # [...] Inside shared_library_*_test.cpp I need something like this: boost::filesystem::path tstlib("bin.v2/libs/mylib/test/mycompiler/test_library.so"); As boost.build decides by itself where to put the binary, i cannot hardcode the path in the source-files.. Is there a way to get the path from boost.build? If not, how could I solve this problem? -- regards dib

2 1