- Boost-users - lists.preview.boost.org

Poisson distribution for large mean
by kgolod0641 18 Mar '03

18 Mar '03

Hi, I am using poisson distribution (boost/random/poisson_distribution.hpp), and have encountered a problem when using a large mean. Let's say I'm setting up my poisson distribution with mean = MEAN, and repicking a sequence of 1000 random numbers. If I average out these 1000 numbers, I would expect to get MEAN (or very close to it). This works OK for mean < 745. Afterwards, no matter how large I set the mean, the same sequence of numbers is generated, with average of 745. Any help would be greatly appreciated!! Thank you, Ksenia

1 0

RE: [Boost-Users] Re: help with mersenne_twister
by scleary＠jerviswebb.com 18 Mar '03

18 Mar '03

> -----Original Message----- > From: Jon Agiato [mailto:JonAgiato@nyc.rr.com] > > Thanks again for all the help! I was looking through > random-generators.html and see that there are in addition to > the Mersenne Twister, Linear Congruential, and Lagged > Fibonacci, five others: const_mod, additive_combined, > shuffle_output, rand48, and inversive_congruential. I am > assuming these are all good RNGs to use for comparison, am I wrong? const_mod: not an RNG; it is used as an implementation helper for other RNGs rand48: uses Linear Congruential algorithm additive_combined, shuffle_output: modifications to the Linear Congruential to try to make it better; so these two require Linear Congruential generator(s) as input > > And lastly (don't want to wear out my welcome.. lol), do > > you think there would be any benefit to using the advanced > > RNG classes in my research? > > Actually, I think it might be a good idea to do so for your > project. The > specializations provided by Boost.Random are well-known > RNGs that have been > published in research papers and used extensively. There > are undoubtedly > many other good specializations out there, but few are used > (mainly because > it is very hard to define what a "good RNG" is). > > Anyway, what I was thinking is that you might make some of your own > specializations that are known to be bad RNGs, and discuss > why they are bad. > Here is where my knowledge about RNGs end; I'm sure that > there are known bad > RNGs, but I don't know what they are. > > -Steve > > That's a great idea Steve, thanks! I am rather new to RNG > research but agree that a comparison of sorts would be good. > Could you or anyone here provide assistance, perhaps a short > example on how one would use one of the advanced RNGs Boost > provides? I use Boost a lot, and think that one of the > things it could really benefit from is more in depth > documentation, especially for those of us struggling to > utilize the library as quickly as possible. Then again, I > know that 1_30_0 should be out soon, so these developers have > enough to do on their plates. Look at the docs again; they explain what is necessary to specialize the generic algorithms. Note that you should be familiar with the generic RNG algorithms before attempting to specialize. For example, looking at the Linear Congruential RNG (which is the only RNG algorithm I'm familiar with), the algorithm is -- as specified in the Boost.Random docs: x(n+1) := (a * x(n) + c) mod m and boost::random::linear_congruential is declared as: template<class IntType, IntType a, IntType c, IntType m, IntType val> class linear_congruential; with a constructor of: explicit linear_congruential_fixed(IntType x0 = 1); Given this, the specialization boost::minstd_rand, declared as: typedef random::linear_congruential<long, 48271L, 0, 2147483647L, 399268537L> minstd_rand; uses the specific Linear Congruential algorithm: x(n+1) := (48271 * x(n)) mod 2147483647 again, x(0) is given by the constructor. (The last template argument is used for implementation validation, and does not affect the algorithm). You can experiment with adding your own specializations, or those defined by other people that are not found in Boost.Random. For example, here's one from "Numerical Recipes in C": typedef boost::random::linear_congruential<long, 69621L, 0, 2147483647L, 0> ParkMiller; As far as the Boost.Random docs go, I think that they are quite good (with the exception that the "val" template parameter is inconsistently documented). Most users of this library just need a random number generator, and the comparision chart with the general recommendation of mt19937 suits the bill for anyone who wants to quickly use the library. What you want to do is not just quickly use a single RNG; you have much more extensive needs, covering a larger range of RNG algorithms. Naturally, more time will be necessary to learn all the algorithms and the Boost implementations of them. -Steve

1 0

RE: [Boost-Users] boost install and library linking
by Michael Hunley 18 Mar '03

18 Mar '03

Yes, but as of the last version I got from CVS (threads/Jamfile dated 1/30/03) it put it in a folder [$(BOOST)/libs/thread/bin-stage] that is not common with other produced libs. This is better than the gristed folders, but still not as useful or convenient as a top level common folder [like $(BOOST)/bin-stage]. When I need to specify in Linux where to look for shared libs, or when I want to add the .lib to link with in MSVC, I prefer the one-stop-shopping for all my boost libs. Perhaps I'm just being lazy, but that is the mother of invention, after all ;). I hear tell that the new v2 boost will have an "install" (don't know if it is a build target or something else) which will probably be far superior still. michael At 12:07 PM 3/17/2003 -0600, you wrote: >Michael Hunley said: > > I posted a solution to this for the v1 jambuild on the jambuild list, > > but not sure if anything will come of it since most of the effort is > > going into v2. My solution (for threads only) was to add a simple > > stage target at the end of the root Jamfile. Here it is: > > > > stage bin > > : ## Source Files ## > > <dll>libs/thread/build/boost_thread > > : ## Tags ## > > <tag><debug>"d" > > : ## default variants ## > > debug > > release > > ; > >Such a target already exists in the thread library's Jamfile (not the root). > >-- >William E. Kempf > > > > >Info: <http://www.boost.org> >Wiki: <http://www.crystalclearsoftware.com/cgi-bin/boost_wiki/wiki.pl> >Unsubscribe: <mailto:boost-users-unsubscribe@yahoogroups.com> > > >Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/

1 0

RE: [Boost-Users] Re: help with mersenne_twister
by scleary＠jerviswebb.com 18 Mar '03

18 Mar '03

> -----Original Message----- > From: Jon Agiato [mailto:JonAgiato@nyc.rr.com] > > Thanks for your suggestions and corrections, you made my > day for sure! I am sort of embarrassed as I have a decent > amount of experience with C++ but this was my first attempt > at writing to a binary file, incorrectly apparently! ;-) If > you do not mind, I would like to ask a few additional questions: I think everyone makes that same mistake; I know I have! :) > I notice there is mt19937 for the mersenne twister, and > mt11213b for the mersenne twister. What is the difference > between the two? Check out the first chart in libs\random\random-generators.html - it compares a few properties of the various specializations. > What other deterministic random number generators does > boost offer, and can I use them in the same fashion as > mt19937 as far as object instantiation? Linear congruential and lagged Fibonacci. All RNGs in Boost.Random follow the same interface, given by the RNG concept documentation in libs\random\random-concepts.html > I have Linux in VMWare as well as a standalone Linux box so > I should be able to utilize the non-deterministic generator. > Do you have any suggestions that I might keep in mind before I begin? No; I've never actually used it myself. > And lastly (don't want to wear out my welcome.. lol), do > you think there would be any benefit to using the advanced > RNG classes in my research? Actually, I think it might be a good idea to do so for your project. The specializations provided by Boost.Random are well-known RNGs that have been published in research papers and used extensively. There are undoubtedly many other good specializations out there, but few are used (mainly because it is very hard to define what a "good RNG" is). Anyway, what I was thinking is that you might make some of your own specializations that are known to be bad RNGs, and discuss why they are bad. Here is where my knowledge about RNGs end; I'm sure that there are known bad RNGs, but I don't know what they are. -Steve

2 1

gcc - cygwin - warning: too many 'l' suffixes in integer constant
by Bruce Adams [TSP Sunbury] 18 Mar '03

18 Mar '03

Hi, I have just begun using boost with gcc-3.2 on a vanilla PC-cygwin platform. Up until deciding to use boost I have been compiling with warnings treated as errors. I am getting warning too many 'l' suffixes in integer constant. See output below. Does anyone know what is going on here? The lines in the boost header file of interest use ULL as suffix. I presume this is intended to mean unsigned long long though I could not track down the #define (ULL being a rather common string). I have discovered that using ull in the preprocessor gives this warning though it is legal when declaring a constant in the language itself. This looks like an inconsistency within gcc. Perhaps because long long isn't officially in ISO C++ yet only C99 (correct me if I'm wrong). I managed to remove all but two occurances of the warning by taking the dubious action of changing ull to ul in boost in cstdint.h & limits.h but I could not track down the remaining instances. Are these suffixes necessary for these constants anyway? If the constant is big enough to need to be a long long the compiler should detect that and do any promotions necessary before comparison, shouldn't it? If its just for buggy compilers (assuming gcc isn't one of them in this case) can't ULL be #defined away? Where can I find this definition? Apologies for wasting bandwidth with this newbie error but can someone recommend a preferred course of action. Thanks & regards, Bruce A. (not actually using long longs in this particular application - yet) gcc -c -x c++ -gstabs -fcheck-new -fstrict-aliasing -fstack-check -frtti -Wall - Wno-long-long -pedantic ReadWord.cpp -o ReadWord.o -I../General -I../Boost/boost_1_29_0 In file included from ../Boost/boost_1_29_0/boost/regex/config.hpp:95, from ../Boost/boost_1_29_0/boost/cregex.hpp:27, from ../Boost/boost_1_29_0/boost/regex.h:26, from ReadWord.cpp:18: ../Boost/boost_1_29_0/boost/cstdint.hpp:202:49: warning: too many 'l' suffixes in integer constant ../Boost/boost_1_29_0/boost/cstdint.hpp:202:105: warning: too many 'l' suffixes in integer constant ../Boost/boost_1_29_0/boost/cstdint.hpp:202:105: warning: too many 'l' suffixes in integer constant ../Boost/boost_1_29_0/boost/cstdint.hpp:202:123: warning: too many 'l' suffixes in integer constant ../Boost/boost_1_29_0/boost/cstdint.hpp:202:195: warning: too many 'l' suffixes in integer constant Note: warning: ISO C++ does not support `long long' warning: ISO C89 forbids long long integer constants are suppressed by -Wno-long-long ============================================================================ Any opinions expressed in this e-mail are those of the individual and not necessarily those of Tyco Safety Products. Any prices for the supply of goods or services are only valid if supported by a formal written quotation. This e-mail and any files transmitted with it, including replies and forwarded copies (which may contain alterations) subsequently transmitted from Tyco Saftey Products are confidential and solely for the use of the intended recipient. If you are not the intended recipient or the person responsible for delivery to the intended recipient, be advised that you have received this e-mail in error and that any use is strictly prohibited. In this event, please notify us via e-mail at 'helpdesk.tepg(a)tycoint.com' or telephone on 0121 255 6499 and then delete the e-mail and any copies of it. ============================================================================

1 0

RE: [Boost-Users] Re: help with mersenne_twister
by scleary＠jerviswebb.com 17 Mar '03

17 Mar '03

> Hello Steve, and thank you for your kind help. I am trying to use > the boost library as part of a project in which I demonstrate the > validity of various random number generation algorithms and > methodologies. I wanted to use boost to demonstrate what can be > achieved through top class C++ generic programming. It sounds like Boost.Random may be a good choice for this. > I used mt19937 as you directed and my example did compile but when I > ran the binary file through Diehard it failed every single test > every time. I am sure this can not be due to the algorithm, but I > can not pinpoint what I could be doing in error. Here is the code I > used to create a binary file for diehard: I am not familiar with the "Diehard" tests. But I see a couple of problems in your example code: you are not creating a binary file, and you are not generating random numbers. > ofstream file("MTRNG.bin", ios::binary); "ios::binary" means open the output text file without \n -> \r\n conversions. It does not make insertions and extractions binary. > file << MT; This means to save the current state of the random number generator in the text output file. To output the generated random numbers (instead of the random number generator state), use: file << MT(); To output the generated random numbers to a binary file, use: const mt19937::result_type x = MT(); file.write((const char *) &x, sizeof(x)); > I was intending on using boost::random to demonstrate various > deterministic and non-deterministic RNGs and their validity as far > as randomness using Diehard. A last note: Boost.Random currently only provides one non-deterministic RNG, and it only works on Linux. -Steve

2 1

RE: [Boost-Users] help with mersenne_twister
by scleary＠jerviswebb.com 17 Mar '03

17 Mar '03

> -----Original Message----- > From: Jon Agiato [mailto:JonAgiato@nyc.rr.com] > > I am having difficulty in understanding the documentation regarding > the use of the pseudo-random number generators in boost::random. > Below, I have created a small sample program of the type I am trying > to create in order to use the various classes in boost::random with > Diehard (requires a binary file of 32 bit integers for testing). > This program does not compile as it says that mersenne_twister > requires template arguments. I understand what that means of > course, but am unsure as to what arguments should go in there. I am > using MSVC++ 7.0 and would really apprechiate any assistance. > Thanks! The Boost.Random library defines fully-generic random number generator algorithms, e.g., boost::random::mersenne_twister, boost::random::linear_congruential, etc. These are only for advanced users. The Boost.Random library also provides specializations of those generic random number generator algorithms for use by normal users. e.g., boost::mt19937 and boost::mt11213b are both specializations of the Mersenne twister; and boost::minstd_rand is a specialization of the linear congruential algorithm. So, unless you have really exceptional random number requirements, you should be using boost::mt19937 instead of boost::random::mersenne_twister: > using boost::random::mersenne_twister; using boost::mt19937; > mersenne_twister MT; mt19937 MT; Then your example should compile. -Steve

2 1

help with mersenne_twister
by agiatojon 17 Mar '03

17 Mar '03

Hello all, I am having difficulty in understanding the documentation regarding the use of the pseudo-random number generators in boost::random. Below, I have created a small sample program of the type I am trying to create in order to use the various classes in boost::random with Diehard (requires a binary file of 32 bit integers for testing). This program does not compile as it says that mersenne_twister requires template arguments. I understand what that means of course, but am unsure as to what arguments should go in there. I am using MSVC++ 7.0 and would really apprechiate any assistance. Thanks! #include <iostream> using std::cout; using std::cin; #include <cstdlib> #include <boost/random/mersenne_twister.hpp> using boost::random::mersenne_twister; #include <iomanip> using std::ios; #include <fstream> using std::ofstream; int main() { ofstream file("MTRNG.bin", ios::binary); const unsigned int N = 10; unsigned int i = 0; mersenne_twister MT; for(; i != N; ++i) { file << MT; } cin.get(); return EXIT_SUCCESS; } Jon Agiato JonAgiato(a)nyc.rr.com

2 1

RE: [Boost-Users] boost install and library linking
by scleary＠jerviswebb.com 17 Mar '03

17 Mar '03

> Clearly, I need to link to one of the boost::test library. How > should I be linking to this library? All libraries that were built > have a .lib extention and are spread over a huge directory > structure. How can I tell which library I need? I'm not that familiar with Boost.Test (yet), so I really can't answer this one. > Is there a way to finnish the build process so that all of the > statically-linked debug libraries are in one place? If somebody > would send me a make file or let me know what is a good approach to > building applications with boost, I'd appreciate it. AFAIK, there is no way to place all the .lib files in a single directory. The phrase "building applications with boost" leads me to state: Boost is not a library. Rather, Boost is a collection of libraries. Most of the Boost libraries (such as Boost.Random) are header libraries, which means that you add ${BOOST} to the include path of your compiler and you're done with the "installation". There are a few Boost libraries (Python, Regex, Signals, Test, and Thread) that produce actual .lib library files. If you are not using those libraries, you do not need to build the .lib files. HTH, -Steve

2 1

Re: [boost] Re: Regressions in type_traits and lexical_cast (was: Re: Questionabout boost::function)
by Douglas Gregor 15 Mar '03

15 Mar '03

On Saturday 15 March 2003 06:03 pm, Andreas Huber wrote: > [ ... ] > > > Index: is_class.hpp > > =================================================================== > > RCS file: /cvsroot/boost/boost/boost/type_traits/is_class.hpp,v > > retrieving revision 1.5 > > retrieving revision 1.6 > > diff -r1.5 -r1.6 > > 84a85 > > > > > ::boost::type_traits::ice_not< ::boost::is_function<T>::value > > >:: > > >::value > > > > John? > > This is due to the missing comma I reported in > > http://lists.boost.org/MailArchives/boost/msg45211.php > > Regards, > > Andreas Ah, yes it is. I've applied the patch (actually, Daniel Frey's version: no need to duplicate code) on mainline and the RC_1_30_0 branch. Doug

2 1