An Example Program

• I have written an example program based upon the previous material
  • I wanted to.
  • To give you a framework for exploring algorithms
  • To introduce some programming techniques which might be useful this semester.
• If you don't know some of the concepts involved, and don't understand after I have talked about them, please ask.
• The Boost library
  • Boost is an organization which is building a set of standard, portable, reviewed c++ extension libraries.
  • Many of the libraries have been incorporated into the new C++ standard (C++11)
  • Scott Meyer's Effective C++, point 55, Familiarize yourself with Boost.
  • There are many useful libraries available.
  • We will look at a few of them.
• Random numbers
  • Random numbers are problematic.
  • The default random number generator in C/C++ is less than outstanding
  • Sculpting and limiting ranges have always been a problem.
  • The boost library provides a replacement.
  • It is more complex than srand, rand
  • But far more complete.
  • And a bit more difficult to use.
  • The system consists of random number generators
    • There are quite a few of these.
    • Many, such as the Mersenne Twister are based upon research papers.
    • If the names of the generators don't ring any bell and you have no idea which generator to use, it is reasonable to employ mt19937 for a start: It is fast and has acceptable quality.
    • The constructor will take a seed.
    • Or you can call the seed function.

      #include <boost/random/mersenne_twister.hpp>
      
      ...
      int main() {
          boost::mt19937 rng(time(NULL));
          boost::mt19937 rng2();
      
          rng2.seed(time(NULL));
      
      
      	     

  • Random number generators are tied to a distribution function
    • A distribution function will transform the underlying distribution of a random number generator to a desired distribution.
    • There are quite a few of these as well.
    • We are most likely interested in uniform integer distribution.
    • I will use uniform_int, the constructor takes a lower and upper bound.

      #include <boost/random/uniform_int.hpp>
      ...
              boost::uniform_int<> dist(1,1<<30);
      
      	    

  • Finally, a variate generator pairs a random number generator with a distribution.
    • This is what is actually used to generate the random numbers.

    • #include <boost/random/variate_generator.hpp>
      
      ...
      	boost::variate_generator<boost::mt19937&, boost::uniform_int<> > bigDie(rng,dist);
      
      	cout << bigDie() << endl;
      	     

  • A sample program
  • There is much more, but this will do for now.
• Time
  • Time has always been troublesome
  • The boost date_time library helps with this.

  • #include <boost/date_time/posix_time/posix_time.hpp>
    	 

  • We will only be using a small portion of this library.
  • Specifically we are interested in the microsecond clock.
  • This provides us with the most accurate clock available on the system.

  • boost::posix_time::ptime startTime, endTime;
    startTime = boost::posix_time::microsec_clock::local_time();
    
    	 

  • An example program
  • There is much more, but this will do for now.
• Passing a function as an parameter.
  • This is really useful if we wish to compare the capabilities of implementations.
  • We can set up a function to "test" a routine.
  • Pass different routines into it, and test each the same way.
  • This requires that the routines have the same function signature.
  • A example program
• The code.
• Points:
  • Command line arguments.
  • Passing a function, and a standard test routine.
• And a Makefile