// -*- mode: c++; mode: visual-line; mode: flyspell; fill-column: 100000 -*- /*************************************************************************** * doc/tutorial_vector_billing.dox * * Part of the STXXL. See http://stxxl.sourceforge.net * * Copyright (C) 2006 Roman Dementiev * * Distributed under the Boost Software License, Version 1.0. * (See accompanying file LICENSE_1_0.txt or copy at * http://www.boost.org/LICENSE_1_0.txt) **************************************************************************/ namespace stxxl { /** \page tutorial_vector_billing A Billing System for Phone Calls (stxxl::vector and stxxl::sort) \author Roman Dementiev (2006) The intended audience of this tutorial are developers or researchers who develop applications or implement algorithms processing large data sets which do not fit into the main memory of a computer. They must have basic knowledge in the theory of external memory computing and have working knowledge of C++ and an experience with programming using STL. Familiarity with key concepts of generic programming and C++ template mechanism is assumed. Let us start with a toy but pretty relevant problem: the phone call billing problem. You are given a sequence of event records. Each record has a time stamp (time when the event had happened), type of event ('call begin' or 'call end'), the callers number, and the destination number. The event sequence is time-ordered. Your task is to generate a bill for each subscriber that includes cost of all her calls. The solution is uncomplicated: sort the records by the callers number. Since the sort brings all records of a subscriber together, we \a scan the sorted result computing and summing up the costs of all calls of a particular subscriber. The phone companies record up to 300 million transactions per day. AT&T billing system Gecko \cite BillingLarge has to process databases with about 60 billion records, occupying 2.6 terabytes. Certainly this volume can not be sorted in the main memory of a single computer (Except may be in the main memory of an expensive supercomputer.) Therefore we need to sort those huge data sets out-of-memory. Now we show how STXXL can be useful here, since it can handle large volumes I/O efficiently. # STL Code If you are familiar with STL your the main function of bill generation program will probably look like this: \code int main(int argc, char * argv[]) { if(argc < 4) // check if all parameters are given { // in the command line print_usage(argv[0]); return 0; } // open file with the event log std::fstream in(argv[1], std::ios::in); // create a vector of log entries to read in std::vector v; // read the input file and push the records // into the vector std::copy(std::istream_iterator(in), std::istream_iterator(), std::back_inserter(v)); // sort records by callers number std::sort(v.begin(), v.end(), SortByCaller()); // open bill file for output std::fstream out(argv[3], std::ios::out); // scan the vector and output bills std::for_each(v.begin(), v.end(), ProduceBill(out)); return 0; } \endcode To complete the code we need to define the log entry data type \c LogEntry, input operator \c >> for \c LogEntry, comparison functor \c SortByCaller, unary functor \c ProduceBills used for computing bills, and the \c print_usage function. \snippet examples/algo/phonebills.cpp prolog # Going Large -- Use STXXL In order to make the program I/O efficient we will replace the STL internal memory data structures and algorithms by their STXXL counterparts. The changes are marked with \c //! \code #include //! include STXXL headers // the rest of the code remains the same int main(int argc, char * argv[]) { if(argc < 4) // check if all parameters are given { // in the command line print_usage(argv[0]); return 0; } // open file with the event log std::fstream in(argv[1], std::ios::in); // create a vector of log entries to read in stxxl::vector v; //! use stxxl::vector instead of std::vector // read the input file and push the records // into the vector std::copy(std::istream_iterator(in), std::istream_iterator(), std::back_inserter(v)); // bound the main memory consumption by M // during sorting const unsigned M = atol(argv[2])*1024*1024; //! calculated memory limit M // sort records by callers number stxxl::sort(v.begin(), v.end(), SortByCaller(), M); //! use stxxl::sort instead of std::sort // open bill file for output std::fstream out(argv[3], std::ios::out); // scan the vector and output bills // the last parameter tells how many buffers // to use for overlapping I/O and computation stxxl::for_each(v.begin(), v.end(), ProduceBill(out), 2); //! use stxxl::for_each instead of std::for_each return 0; } \endcode As you note the changes are minimal. Only the namespaces and some memory specific parameters had to be changed. See \ref examples/algo/phonebills.cpp for the full source code. The example program is automatically compiled when building STXXL, refer to \ref install on how to build programs with STXXL. The program \ref examples/algo/phonebills_genlog.cpp can be used to generate logs for processing with the phonebills example. Do not forget to configure you external memory space in file .stxxl. See \ref config. \example examples/algo/phonebills.cpp This example code is explain in \ref tutorial_vector_billing \example examples/algo/phonebills_genlog.cpp This example code is explain in \ref tutorial_vector_billing */ } // namespace stxxl