July 2004 - Boost - lists.boost.org

Boost Mathematicians
by Tom Brinkman 07 Jul '04

07 Jul '04

What is the current status of all Boost mathmatical, statistical and econometric libraries currently in development? The collection of boost math libraries that are currently included in the library are world-class and provide a wonderful foundation to build-upon. However, many areas of advanced statistical and mathmatical inference are not currently being addressed by any of the boost libraries. Is there currently a way for those boost developers with a background in mathmatics to coordinate their efforts. If not, why not? Lets get such a discussion started. A continued and regular dialog of mathmatics as it relates to boost library development would be of great use, primaryily to avoid overaping our efforts. If your a mathmetician and agree with me that boost is a great foundation to build-upon, let the group/me know. Points for Discussion: 1) We should keep a seperate list of boost::math libraries that are currently under development and provide a forum for advanced discussions. How can we do this? 2) We need a way to attract the best mathmaticians in the world. How can we make the boost libraries an even more attractive development platform for the mathemiticians? 3) We should create a list of mathmatical algorithims and libraries that are needed. 4) We need some very basic boost specific math-tutorials for our future mathemiticans and potential library writers. 5) We need to advertise to the development community about how strong of a foundation boost is for mathmatical library development. In short, lets tell the world that this is the premiere place for advanced mathmatical library development, as I believe that it could be. Comments? Regards, Tom Brinkman __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

23 70

RE: [boost] Re: [utility] boost::reset
by jhopkin＠reflectionsinteractive.com 07 Jul '04

07 Jul '04

Daniel Frey wrote: > jhopkin(a)reflectionsinteractive.com wrote: > > The STL implementation I have been using seems to allocate a buffer on first > > insertion into any container of a given type, and frees it only when it > > feels like it (often on program shut-down). > > Sounds like a pool allocator. Question: Does .clear() for your STL > return the buffer to the pool or is it still bound to the instance? What > happens in case of the .swap()-trick? I believe .clear() and the swap trick do what's expected per instance. Specifically for vector, clear generally won't release any memory to the pool; the swap trick releases it all. James Hopkin

2 1

RE: [boost] Does anyone have an is_virtual_base_and_derived ?
by SRobb＠reflectionsinteractive.com 07 Jul '04

07 Jul '04

I wrote that very thing a year or so ago for the purpose of testing if a base pointer type can be statically downcast to a derived type. I don't claim to have tested all permutations of its usage, but it certainly worked on the cases I was testing it with. However, it does have the unfortunate side-effect of generating an 'ambiguous base' warning on the later GCCs every time it is used with a class which inherits another non-virtually. (though the traits class still gives the correct answer) If anyone can find a way of preventing this warning, or even an alternate implementation, I'd be glad to hear it. Here is my basic implementation: #include <boost/config.hpp> #include <boost/type_traits/is_convertible.hpp> namespace is_virtual_base_and_derived_detail { #ifdef _MSC_VER #pragma warning(push) #pragma warning(disable : 4250) #endif template <typename BaseImpl, typename DerivedImpl> struct tester : public DerivedImpl, virtual public BaseImpl { tester(); tester(const tester&); }; #ifdef _MSC_VER #pragma warning(pop) #endif template <typename Base, typename Derived, bool IsBaseAndDerived> struct impl; template <typename Base, typename Derived> struct impl<Base, Derived, true> { BOOST_STATIC_CONSTANT(bool, value = sizeof(tester<Base, Derived>) == sizeof(Derived)); }; template <typename Base, typename Derived> struct impl<Base, Derived, false> { BOOST_STATIC_CONSTANT(bool, value = false); }; } template <typename Base, typename Derived> struct is_virtual_base_and_derived : is_virtual_base_and_derived_detail::impl<Base, Derived, boost::is_convertible<Derived*, Base*>::value> { }; template <typename T> struct is_virtual_base_and_derived<T, T> { BOOST_STATIC_CONSTANT(bool, value = false); }; -----Original Message----- From: boost-bounces(a)lists.boost.org [mailto:boost-bounces@lists.boost.org] On Behalf Of Robert Ramey Sent: Tuesday, July 06, 2004 6:01 PM To: boost(a)lists.boost.org Subject: [boost] Does anyone have an is_virtual_base_and_derived ? Type traits includes: ::boost::is_base_and_derived<T,U> Does anyone have a version that might be called is_virtual_base_and_derived<B, D> which would return true iff D is derived virtually from B ? Robert Ramey _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost

1 0

regression tests on irix
by Toon Knapen 07 Jul '04

07 Jul '04

I'm trying to get the regression tests working again on SGI/Irix but I fail to build the html tables. Apparantly the compiler_status tool is looking for a Jamfile in the wrong place. Anybody a clue what could be happening (failing command and its output in attachment) toon generating html tables: /disk2/home/tk/boost_regression/boost/tools/regression/build/run/compiler_status --comment /disk2/home/tk/boost_regression/boost/../regression_comment.html /disk2/home/tk/boost_regression/boost cs-IRIX64.html cs-IRIX64-links.html Could not open Jamfile: /disk2/home/tk/boost_regression/Jamfile **** error return code: 1 ******** errors detected; see standard output for details ******** Failed HTML result table generation.

1 0

Re:[boost] serialization 1.31 compilation error.
by Robert Ramey 07 Jul '04

07 Jul '04

Russell Balest wrote: >Possibly triggered by the using statement: >using namespace boost::serialization; >on line 53 of iserializer.hh , but that's a guess. Using "using" in the header was an error that was pointed out during the formal review. My current copy has fixed this a new version will be available very shortly #20. This will be guaranteed compatible with 1.31. Also very shortly the latest version will be checked into main CVS. So you can wait a little bit or remove the "using from your copy of the iserializer.hpp Robert Ramey

1 0

Interest in strict unit style typing for physics/science calculations
by edA-qa mort-ora-y 07 Jul '04

07 Jul '04

Is there any interest in a library that performs strict type checking on values with unit quantities? The intent is to prevent operation and assigment of unlike types. For example, if A is in meters/second, and B is in seconds: -A * B results in a value with "meters" units -A / B results in a value with "meters/seconds^2" -A = B results in an exception due to a mismatch The code I have so far supports arbitrary unit types via tokens, which have to be defined for a project. Some example code looks like (with several notation options): typedef UnitValue<float> uFloat; UnitType unit_m( 1 ); //define meters type UnitType unit_s( 2 ); //define seconds type uFloat time( unit_s ); time = unit_s( 3.0f ); //3 seconds uFloat speed( unit_m / unit_s ); speed = 15 * speed.Type(); //set speed of 15m/s uFloat accel( unit_m / unit_s / unit_s ); accel = speed / time; //results in 5m/s^2 accel = speed; //throws exception, m/s != m/s^2 time = accel; //throws exception, s != m/s^2 The library has a price for performance compared to native value operations. However, part of the requirements will be a flag (a #define ) that turns off the unit processing for the code, which effectively removes the classes (replaces the UnitValue simply with typedefs of the stored type). -- edA-qa mort-ora-y Idea Architect http://disemia.com/

3 3

Write access to the CVS repository
by Edward Diener 07 Jul '04

07 Jul '04

I wanted to correct a spelling error in one of the Boost CVS documentation files but I didn't have write access to the Boost CVS repository. Is there a way to get this ? If not, should I mention such an error on this NG or is there some other place ?

3 9

RE: [boost] Boost.Test, failed assert output
by Glen Knowles 06 Jul '04

06 Jul '04

> From: Rozental, Gennadiy > I will see what I can do about it. If I remember correctly, > the problem is that I do not know where the error occurred. > So I have no way to print line pointing there. I could print > only line pointing where the error was caught/detected. I am > not sure it is what you want. Ideally you would point to the line, but I can simply look at the output from the assert which has the file/line info. I just want the IDE to notice that there was an error so that I'll know something went wrong. As far as I'm concerned you could make up a file and line if that's what it takes to stop the build, in fact reporting an "unknown location" would probably cause less confusion then reporting the place inside boost.test where the error was detected. Thanks, Glen

1 0

BGL Algorithm (Network Flow)
by Tanton Gibbs 06 Jul '04

06 Jul '04

I'm attempting to do a graph algorithm and am looking for a good BGL algorithm to do this -- any advice would be appreciated. Basically, I have a directed, weighted, bipartite graph. I'll call the two sets of nodes (V,W). Every node in V connects to one or more nodes in W. However, not every node in W is connected to. When the algorithm ends, I want every node in V to be connected to 0 or 1 node in W and no node in W should have more than one connection to it. I also want the total weight of all the edges to be minimized (basically, prune the most costly edges to create the graph), but no edge should be pruned unnecessarily (I want to maximize the number of connected nodes). It looks to me to be a pretty evil problem, but I'm hoping someone has a good suggestion. Thanks! Tanton H. Gibbs, Ph.D. Software Developer Acxiom Corporation

1 0

Re: [ublas-dev] Re: [boost] Re: Boost Mathematicians
by Jeremy Graham Siek 06 Jul '04

06 Jul '04

Hi Michael, On Jul 5, 2004, at 4:52 PM, Michael Stevens wrote: > Hi Jeremy, > > Thanks for chipping in on this. I don't think we have conversed > before, apart > from maybe a few patches I submitted for MTL. Although it would be > great for > everyone to work on uBLAS I think a revived MTL would also be a good > thing. > The design of MTL was very influential. Thanks! > I won't comment on the other issues as I have already posted a few > thoughts on > these matters. > > > > Mmmh, this open pitfalls concerning A*B*C in performance issues > for > > > non experts. > > > > No, the right way to handle this is to have the expression template > > machinery > > identify that a temporary is needed, and create one automatically. > > I wonder if you caught the discussion on this about 6 months ago. I > would > strongly argue that the automatic creation of temporaries in the > subexpression context is a bad thing... > > The roles of the expressions we declare (as in declarative > programming) is > twofold. a) to express the mathematical context b) to express how > this should > be computed. That is, to be able to express how the expression is to > be > evaluated. This gives us control over its numerical properties and > also of > the time and space complexity. > > In principle, uBLAS has a one to one relationship between an > expression and > how it is evaluated. Simply stated this is based on an element by > element > evaluation of assignment. There are shortcuts in the case of sparse > types but > the principle holds. > > An expression A+B+C can be evaluated in this manner trivially. > A*B*C can also be evaluated in this manner, although the element by > element > rule results in an O() that is not what many users expect. > To get the O() that users expect we could automagically insert > temporaries. I > fear this is a tar pit however. What temporary should be chosen? Each > of A B > or C could be a either dense, sparse, compressed, banded, symmetric or > combinations therefore. The type the user expects or is "the best" is > not at > all clear. Although automagically inserting temporaries in non-trivial (especially since it must all be done with template meta-programming) I do not think it is a tar pit. Here is the simply heuristic that I used for MTL 3 to decide where to put temporaries: 1. If an operation needs to read from an argument multiple times, evaluate the argument eagerly into a temporary. For example, a sparse matrix in compressed row format times the sum of two vectors: A * (x + y) In this case, the (x + y) would be evaluated into a temporary before evaluating the multiplication. 2. If an operation writes into its output in multiple passes, then perform the operation in an eager fashion. For example, the sum of a vector and the product of a sparse matrix in compressed column format and a vector. x + (A * y) In this case, the (A * y) would be evaluated into a temporary before evaluating the addition. > > I think the only correct solution is to force the user to specify > this type. > Incidently, the no temporary evaluation should also be an option as > in some > cases (small sizes) it is desirable. uBLAS currently disallows this > to avoid > confusing the user. I should be reinstated in some form however. > > All the best, > Michael > > -- > ___________________________________ > Michael Stevens Systems Engineering > > 34128 Kassel, Germany > Phone/Fax: +49 561 5218038 > > Navigation Systems, Estimation and > Bayesian Filtering > http://bayesclasses.sf.net > ___________________________________ > > _______________________________________________ Jeremy Siek <jsiek(a)osl.iu.edu> http://www.osl.iu.edu/~jsiek Ph.D. Candidate, Indiana University Bloomington C++ Booster (http://www.boost.org) _______________________________________________

3 2