- Boost-users - lists.boost.org

RE: [Boost-users] Changing the heap allocator
by Jim.Hyslop 15 Apr '04

15 Apr '04

David Barrett-Lennard wrote: > Yes HeapSetInformation() will do it once you have the heap handle. > However, the CRT creates its own heap with a call to HeapCreate(), and > stores the heap handle in a global variable called _crtheap > that is not > exported by the library. Right, of course - I thought that GetProcessHeap() would return the correct heap, but it does not. -- Jim Hyslop Senior Software Designer Leitch Technology International Inc. (http://www.leitch.com) Columnist, C/C++ Users Journal (http://www.cuj.com/experts)

1 0

RE: [Boost-users] Changing the heap allocator
by Jim.Hyslop 14 Apr '04

14 Apr '04

David Barrett-Lennard wrote: > Does anyone know how to enable the LFH (low fragmentation > heap) for the > heap created by the VC CRT? Have you tried HeapSetInformation()? -- Jim Hyslop Senior Software Designer Leitch Technology International Inc. (http://www.leitch.com) Columnist, C/C++ Users Journal (http://www.cuj.com/experts)

2 1

Changing the heap allocator
by David Barrett-Lennard 13 Apr '04

13 Apr '04

Hi, I'm working on a project where I've replaced Microsoft's (terrible) heap allocator with the Doug Lea allocator. There seems to be some cases where the Windows XP allocator is 500 times slower than it should be. This made my program take minutes rather than seconds to run. To try to understand what kind of allocation patterns cause the behaviour I wrote a small console application (see below). After playing around with different allocation patterns for half an hour I found a case where the pathological behaviour appeared. I'm none the wiser about the problem. For example, moving the malloc(301) statement before the 7 x malloc(1000) makes the problem go away! However, given the ease with which I rediscovered the problem, I wonder why there aren't more complaints about Microsoft's heap allocator. With some allocation patterns I have seen the allocator fall below 1kHz on a Pentium IV 2GHz machine! I've created a library that exports alternative versions of operator new and operator delete, by using a def file as follows EXPORTS ??2@YAPAXI@Z ??3@YAXPAX@Z I've had to do some hacky things to make it work with the STL (because of msvcp60.dll) , but that's another story! Now my question : I want to use boost, but need to work out how to make it use a different allocator. Will I need to rebuild the boost dlls (such as boost_python.dll), against my static library? Regards, David Barrett-Lennard /////////////////////// heap allocation test /////////////////////////// #include <iostream> #include <time.h> int main(int argc, char* argv[]) { for (int count=1 ; count <= 5 ; ++count) { int n1 = 0, n2 = 0; clock_t start = clock(); { const int NUM = 10000; void* L[NUM*7]; for (int i=0 ; i < NUM ; ++i) { for (int j=0 ; j < 7 ; ++j) L[n1++] = malloc(1000); void* leak = malloc(301); ++n2; } for (int k=n1-1 ; k >= 0 ; --k) free(L[k]); } clock_t elapsed = clock()-start; if (elapsed) { std::cout << "Time = " << elapsed << " ms" << " Rate = " << 1000 * (n1 + n2) / elapsed << " Hz" << std::endl; } } return 0; }

3 4

Using boost::regex 1.31.0 on vc6
by Chris Branch 11 Apr '04

11 Apr '04

Hi, I'm trying to use boost::regex 1.31.0 (the v4 source code) on MSVC++6 (sp5), but whenever I compile a project using the new code I get an internal compiler error in perl_matcher_common.hpp and no object is built. I want to use the new v4 code (the older v3 code works, and I'll use that if I have to) but this file is quite important... is there something I can do, or does it not work with VC6? Thanks, Chris

2 2

boost::thread and boost::bind
by GoochRules! 11 Apr '04

11 Apr '04

I am attempting to do this in vs.net 2003: int f(HANDLE, HANDLE, HANDLE); int main(int argc, char* argv[]) { HANDLE x; HANDLE y; HANDLE z; boost::thread* p = new boost::thread(boost::bind(f, x, y, z); p->join(); delete p; } As a contrived example it works. In my project it does not, spewing out 5 errors: 1) c:\src\include\boost\bind.hpp(62) : error C2825: 'F::result_type': cannot form a qualified name c:\src\include\boost\bind\bind_template.hpp(16) : see reference to class template instantiation 'boost::_bi::result_traits<R,F>' being compiled with [ R=boost::_bi::unspecified, F=int (__cdecl *)(HANDLE,HANDLE) ] c:\src\ansa-1.0\src\main\main.cpp(369) : see reference to class template instantiation 'boost::_bi::bind_t<R,F,L>' being compiled with [ R=boost::_bi::unspecified, F=int (__cdecl *)(HANDLE,HANDLE), L=boost::_bi::list3<boost::_bi::list_av_3<fd_xpt,fd_xpt,fd_xpt>:: B1,boost::_bi::list_av_3<fd_xpt,fd_xpt,fd_xpt>::B2,boost::_bi:: list_av_3<fd_xpt,fd_xpt,fd_xpt>::B3> ] 2) c:\src\include\boost\bind.hpp(62) : error C2039: 'result_type' : is not a member of 'operator``global namespace''' 3) c:\src\include\boost\bind.hpp(62) : error C2146: syntax error : missing ';' before identifier 'type' 4) c:\src\include\boost\bind.hpp(62) : error C2955: 'boost::_bi::type' : use of class template requires template argument list c:\src\include\boost\bind.hpp(94) : see declaration of 'boost::_bi::type' 5) c:\src\include\boost\bind.hpp(62) : fatal error C1903: unable to recover from previous error(s); stopping compilation note: fd_xpt is declared as "typedef HANDLE fd_xpt;", and c:\src\ansa-1.0\src\main\main.cpp(369) is similar to "boost::thread* p = new boost::thread(boost::bind(f, x, y, z);" in the example. It appears that the compiler thinks my first argument is a "int (__cdecl *)(HANDLE,HANDLE)" when it is most certainly a "int (__cdecl *)(HANDLE,HANDLE,HANDLE)". It used to take two arguments, but I've since added a third (it was at this very point that these errors crashed the party). I am sure that there is no declaration that specifies only two arguments: I've had two other developers verify this. I have cleaned my project multiple times, even by hand, as well as getting a clean source tree from cvs, and yet I still get these errors. Any suggestions on how to silence these errors are greatly appreciated. --mpeltzer goochrules(a)hotmail.com

3 3

Re: [boost] Re: [python] external references to python details
by Domenico Andreoli 11 Apr '04

11 Apr '04

hi, Lars Gullik Bj�nnes writes: > David Abrahams <dave(a)boost-consulting.com> writes: > > | larsbj(a)gullik.net (Lars Gullik Bj�nnes) writes: > > > >> I just discovered that boost libraries (or only one...) depend on > >> files in python/detail. That should probably be fixed? > > > | Got a suggested patch? > > Not really. > > With the subset I use from boost moving indirect_traits.hpp and > is_xxx.hpp from boost/python/detail/ to boost/detail/ (and fixup > #includes) seems like the simple fix. here at Debian we have the same problem, please see bug #242801 [0]. this suggestion could would fix also our situation. any news? thanks domenico [0] http://bugs.debian.org/242801 -----[ Domenico Andreoli, aka cavok --[ http://filibusta.crema.unimi.it/~cavok/gpgkey.asc ---[ 3A0F 2F80 F79C 678A 8936 4FEE 0677 9033 A20E BC50

1 0

template
by adel.essafi 09 Apr '04

09 Apr '04

hi list he is the source code of my project. it is an algorithms for scheduling DAG taks on machines. there is a template function that does not work well. please , can you guess where is the fault. thanks Accédez au courrier électronique de La Poste : www.laposte.net ; 3615 LAPOSTENET (0,34/mn) ; tél : 08 92 68 13 50 (0,34/mn)

1 0

template
by adel.essafi 09 Apr '04

09 Apr '04

hi list I am developping a scheduling algorithm. here you have the source code of my project. please , can you see what is wrong with it. thanks Accédez au courrier électronique de La Poste : www.laposte.net ; 3615 LAPOSTENET (0,34/mn) ; tél : 08 92 68 13 50 (0,34/mn)

1 0

RE: [Boost-users] Re: Any Success with AIX and GCC?
by James Zappia 09 Apr '04

09 Apr '04

I've compiled gcc on aix 4.3.3 in the past and no matter what c++ program I wrote, threading or no, I needed to specify the -pthread flag to get it to compile. It's an aix runtime thing with c++, pure c programs didn't have this problem. HTH, James -----Original Message----- From: Rene Rivera [mailto:grafik.list@redshift-software.com] Sent: Wednesday, April 07, 2004 12:55 PM To: boost-users(a)lists.boost.org Subject: [Boost-users] Re: Any Success with AIX and GCC? Peter Jones wrote: > Hey, remember back when Rene Rivera said: > >>The gcc toolset is already adding the -pthread option when linking >>multi-threaded variants (lines 143-145). Are you saying you need to use >>-pthread even when linking single-threaded variants? > > > This is what "bjam -sTOOLS=gcc -sBUILD=release" produces without my > change to gcc-tools.jam: Which builds the release single-threaded variants. > gcc-Link-action > bin/boost/libs/filesystem/build/libboost_filesystem.so/gcc/release/shared-li nkable-true/libboost_filesystem-gcc-1_31.so > bin/boost/libs/filesystem/build/libboost_filesystem.so/gcc/release/shared-li nkable-true/libboost_filesystem-gcc-1_31.imp > ld: 0711-317 ERROR: Undefined symbol: .pthread_mutex_lock > ld: 0711-317 ERROR: Undefined symbol: .pthread_mutex_unlock > ld: 0711-317 ERROR: Undefined symbol: .pthread_mutex_init > ld: 0711-317 ERROR: Undefined symbol: .pthread_mutex_destroy Is there a way to tell who is referencing those symbols? As it might not be Boost code but other runtime code. -- Something to try would be to compile an empty *.so. That is compile an empty source file and try doing a "-shared" link. If you get thread errors with an empty shared library, then the problem is that the GCC+AIX runtimes require threading. And I'll fix that on the toolset, and post a patch ;-) -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org - grafik/redshift-software.com - 102708583/icq _______________________________________________ Boost-users mailing list Boost-users(a)lists.boost.org http://lists.boost.org/mailman/listinfo.cgi/boost-users

1 0

ublas and matrix expression
by Russell Hind 09 Apr '04

09 Apr '04

I have an expression a = (X^T * X)^-1 * X^T * y where x is an n * k matrix and y is an n row vector How do I implement this expression in ublas? I can create and populate the matrices and can do the product of matrices and the transpose use prod and trans, but I'm unsure on the ^-1. Can I do this with ublas? (I don't know much about matrices, but have been given the above expression as a way of finding a polynomial fit for a set of data). Thanks in advance Russell

4 14