January 2010 - Boost-users - lists.boost.org

[math][special_functions] Why is cbrt(x) slower than pow(x, 1/3)?
by Tim Odenthal 18 Jan '10

18 Jan '10

Hi! For a (cell) simulation code I have to calculate lots of cubic roots in every timestep. So I tried to improve performance on my first guess, namely pow(x, 1/3) using boost::math::cbrt(). To my astonishment, this is much slower than the original code. I wrote a small test program to check this claim; compiling with g++ (Ubuntu 4.4.1-4ubuntu9) 4.4.1 (-O3) on a Intel Core i7 CPU I get the following timings (averaging the time over 10 trials): average time to compute 5000000 roots with pow(): 0.603 s. average time to compute 5000000 roots with boost::cbrt(): 1.087 s. average time to compute 5000000 roots with improved boost::cbrt(): 1.015 s. average time to compute 5000000 roots with exp(1/3*log()): 0.541 s. My "improved" version allows giving the boost::math::cbrt() method a first guess instead of taking just the value of the number whose root is to be calculated. Actually, the function I called in this instance 5000000 times does a bit more than just calculating the root; I want to calculate the radii of to spheres which overlap so, that their combined volume is equal to the volume of a "mother sphere": /** * @brief calculates the radius of a daughter cell during symmetric division, * assuming the volume of the mother cell is conserved. * * @param MotherRadius radius of mother cell before division * @param Overlap of the two daughter cells, i.e. radius of daughter - * distance from center of daughter to middle-plane of * both cells * * @return for positive overlap returns positive radius, for negative overlap * negative radius **/ double daughterRadiusfOver( double MotherRadius, double Overlap) { // r = 1/2 (2 sqrt(R^6-h^3 R^3)-h^3+2 R^3)^(1/3)+h^2/(2 (2 sqrt(R^6-h^3 R^3)- h^3+2 R^3)^(1/3)) double O3 = Overlap*Overlap*Overlap; double M3 = MotherRadius*MotherRadius*MotherRadius; double step1 = pow(-O3 + 2*sqrt(-O3*M3 + M3*M3) + 2*M3,1.0/3.0);//<-- CUBIC ROOT HERE!!! Change to: // double step1 = boost::math::cbrt(-O3 + 2*sqrt(-O3*M3 + M3*M3) + 2*M3); if ( Overlap < 0 ){ cerr << "Warning in daughterRadiusfOver: Negative overlap given!" <<endl; return -(step1 + Overlap*Overlap/step1)*0.5; } else { return (step1 + Overlap*Overlap/step1)*0.5; } } Why is boost::math::cbrt() no improvement for me - or am I using it the wrong way? For which cases might it be an improvement, or why is it in the library? It's funny, but I found that std::exp(1/3*std::log(x)) is so far the fastest way... Thanks in advance Tim

5 6

boost::format warning
by Henning Meyer 16 Jan '10

16 Jan '10

Hello, when compiling a simple test like this: #include <boost/format.hpp> int main(void) { int c = 0; boost::format("%1% %2%") % (c==42) % c++; } with this command: $ c++ -c -O -Wall -o ftest.o ftest.cc I get this warning: ftest.cc: In function ‘int main()’: ftest.cc:4: warning: operation on ‘c’ may be undefined What did I do wrong? Thanks for the help! Henning $ c++ -v Using built-in specs. Target: i486-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.4.1-4ubuntu8' --with-bugurl=file:///usr/share/doc/gcc-4.4/README.Bugs --enable-languages=c,c++,fortran,objc,obj-c++ --prefix=/usr --enable-shared --enable-multiarch --enable-linker-build-id --with-system-zlib --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.4 --program-suffix=-4.4 --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug --enable-objc-gc --enable-targets=all --disable-werror --with-arch-32=i486 --with-tune=generic --enable-checking=release --build=i486-linux-gnu --host=i486-linux-gnu --target=i486-linux-gnu Thread model: posix gcc version 4.4.1 (Ubuntu 4.4.1-4ubuntu8) $ uname -a Linux treponema 2.6.31-17-generic #54-Ubuntu SMP Thu Dec 10 16:20:31 UTC 2009 i686 GNU/Linux

2 2

[filesystem] boost 1.41, hpux, gcc 440, compilation error solved
by Rui Pedro Fernandes 15 Jan '10

15 Jan '10

Hi, While trying to compile filesystem under HPUX because the WCHAR_MAX define value used was including a ')' character (because the value has a cast). The include file wchar.h defines the correct WCHAR_MAX if macro _INCLUDE_STDC__SOURCE_199901 is defined. I then solved the problem by changing the filesystem's Jamfile.v2 to add the lines: : requirements <host-os>hpux,<toolset>gcc:<define>_INCLUDE_STDC__SOURCE_199901 Don't know however if this is the best way to solve the issue... Thanks, Rui

1 0

Dead link on main page
by Tim St. Clair 15 Jan '10

15 Jan '10

Version 1.40.0.cmake2 (Boost + CMake) goes to no where. -- Cheers, Timothy St. Clair

1 0

[serialization] boost 1.33.1, HP-UX ia64 & gcc-3.4.4
by Frank Bergemann 15 Jan '10

15 Jan '10

Hi, for a porting job i needed to shift from OSF tru64 to HP-UX V3 ia64 (on both using gcc-3.4.4) There i have stream error exception for boost::serialization data import via XML. The exception is thrown here: include/boost/archive/impl/basic_xml_iarchive.ipp: template<class Archive> BOOST_ARCHIVE_OR_WARCHIVE_DECL(void) basic_xml_iarchive<Archive>::load_start(const char *name){ // if there's no name if(NULL == name) return; bool result = this->This()->gimpl->parse_start_tag(this->This()->get_is()); if(true != result){ boost::throw_exception( archive_exception(archive_exception::stream_error) ); } // don't check start tag at highest level ++depth; return; } I.e. it fails to parse the XML data via boost:spirit. I tried to debug with 'gdb', but was impossible for me at the level of boost::spirit parser implementation. Then i was remembering, that there are two sets of boost libs - one "normal" and a "debug" version. So i switched to the debug version - just to see, if i can better dig into the details with 'gdb' with it. And suddenly the problem was gone. I.e. my application works, if i link with libboost_serialization-gcc-d.a But it doesn't work, if i link with libboost_serialization-gcc.a I also tried to extract all the stuff from the boost::serialization, that is used by my application, into a little test program. Just to feed in the same XML string, that fails for my application. I wanted to see, that i can re-produce the problem - but for a minimum program. - for using that next to evaluate deeper. But - unfortunately - for my test programm it works (with both library versions, "debug" and "normal" one). So it is really hard to find the reason for the problem. I also checked about compiler support on the boost.org pages. And actually it was only boost lib 1.38.0, which mentions HP-UX IA64 together with gcc. But that was gcc-4.2.1 (i use 3.4.4.). Before 1.38.0, it was alway the HP native compiler (not gcc). After 1.38.0 HP-UX is not mentioned at all anymore. So i guess i am generally in bad conditions for my "constellation" (HP-UX ia64 V3, gcc-3.4.4, boost 1.33.1). What is the difference about "debug" lib image and "non-debug" lib image? Is it just optimization level for the compiler? Do you have any hint for me? It might be a problem of gcc-3.4.4 itself on hp-ux (i'll check). But it would be interesting to here, if s.o. else had the same problem(?!) - many thanks! Frank Bergemann ____________________________________________________________________ WEB.DE MillionenKlick: Kostenlos tippen, täglich 1 Million gewinnen! http://produkte.web.de/go/07/

2 1

How to use boost-build/bjam with only home directory acces?
by s seelenluft 15 Jan '10

15 Jan '10

Hi, I have successfully installed boost-build and bjam on a Linux workstation where I only had access to my home folder, but I have trouble replicating this on another. I put the boot-build folder at the top level of my home folder and created also a bin folder there into which I copied the content of boost-build folder and the bjam binary. I then added an environmental variable BOOSTROOT with the value '../' and set my jamroot file to <include>$BOOSTROOT. I can now start bjam from folders one level below the home folder. On a different Linux system this fails with the error: Unable to load Boost.Build: could not find "boost-build.jam" --------------------------------------------------------------- Attempted search from /afs/.../trunk-1 up to the root and in these directories from BOOST_BUILD_PATH and BOOST_ROOT: /usr/share/boost-build. Please consult the documentation at 'http://www.boost.org'. The boost-build.jam file sits in the same folder as bjam (~/bin) but on one system bjam finds it and on another it does not. The folder /usr/share/boost-build does not exist and I do not have the permission to create it. I've used boost-jam 3.1.13 on both systems (since it worked fine on one system, I was loth to change anything). This was the current version when I started the whole business. In short: Is there any way to tell bjam to use the local boost-build.jam file? Thanks. Markus

1 0

[karma] using attributes out of order
by Cory Nelson 15 Jan '10

15 Jan '10

Hello, I was fairly familiar with Spirit 1, but I'm now making my first foray into Spirit 2. I'm trying to write something with karma's attribute system. Say I've got a fusion adapted struct: struct s { int foo; double bar; std::string baz; }; The Spirit docs give plenty of examples like: int_ << double_ << string But I can't find any docs showing how to output the members in a different order, like: double_ << string << int_ Is there a recommended way to achieve this? Thanks! -- Cory Nelson

3 2

BoostJam sensitive to spaces in paths, in particular in TMP
by Bilokon, Paul 15 Jan '10

15 Jan '10

Hi all, Just found this, apparently a bug, in BoostJam. I'm using an admittedly old version of Boost, V1.36.0, in this particular instance. I'm using the latest BoostJam, V3.1.17. This occurs on Windows when the TMP environment variable contains spaces. In my case it was "C:\Documents and Settings\xxxxxxxxxxx\Local Settings\Temp" When you try to run something like C:\Programs\Win32\BoostJam\V3.1.17\bjam.exe --build-type=complete --toolset=msvc-90 --with-thread You get MkDir1-quick-fix-for-windows bin.v2\libs\thread 'C:\Documents' is not recognized as an internal or external command, operable program or batch file. if not exist "bin.v2\libs\thread\\" mkdir "bin.v2\libs\thread" ...failed MkDir1-quick-fix-for-windows bin.v2\libs\thread... This is fixed by set TMP=C:\temp I don't know if this merits a patch. Regards, Paul Paul Bilokon, Vice President Citigroup | FX - Options Trading | Quants 33 Canada Square | Canary Wharf | Floor 3 London, E14 5LB Phone: +44 20 798-62191 paul.bilokon(a)citi.com<mailto:paul.bilokon@citi.com>

2 1

read_graphviz in 1.4x series
by Cyrus Hall 15 Jan '10

15 Jan '10

Ciao all- I've been happily using the boost graph library for a number of months now. Thanks to everyone who has contributed to this great piece of software. After upgrading from boost 1.38 to 1.40 in Ubuntu, read_graphviz now prints to std::cerr the graph it has just read. I downloaded to the source code to make sure this wasn't some cruft left over by the Ubuntu package managers, and it seems these print statements are in the main distribution package. This change seems to correspond with a complete rewrite of the scanning and parsing underlying read_graphviz, which seems to have happened between 1.39 and 1.40. 1.41 contains no changes from 1.40 in read_graphviz_new.cpp. The particular offending lines of code are in translate_results_to_graph in read_graphviz_new.cpp, starting at line 760. This is called directly from detail::graph::read_graphviz on line 796. I see no obvious way to toggle these print statements. But I also can't find other messages about this problem, so I'm guessing I'm just missing something. Is there someway I can get read of this debugging output? Cheers, Cyrus

2 3

atomic read/write operations for integral types/pointers, just like tbb::atomic<T>.
by Michal 15 Jan '10

15 Jan '10

Hallo Group Members. Does boost::thread offer something that can act on integral types/pointers in atomic fashion, just like intel's tbb::atomic<T> does? best regards, Michal

1 0