> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec start
get_properties CPU time : 3.06708 secs
emplace set sort CPU time : 0.903617 secs
write stdout CPU time : 1.23459 secs
total CPU time : 5.20544 secs
total wall clock time : 5 secs
> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec start
get_properties CPU time : 3.4832 secs
emplace set sort CPU time : 1.10209 secs
write stdout CPU time : 0.939805 secs
total CPU time : 5.52519 secs
total wall clock time : 6 secs
> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec start
get_properties CPU time : 3.42605 secs
emplace set sort CPU time : 0.916222 secs
write stdout CPU time : 1.00049 secs
total CPU time : 5.343 secs
total wall clock time : 5 secs
> diff f.tmp vec.tmp
####
> For some reason, short string version is faster when compiled with clang++
> (while long string version is not)
> clang++ -o llil2vec -std=c++11 -Wall -O3 llil2vec.cpp
> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec (fixed string length=6) start
get_properties CPU time : 1.87217 secs
emplace set sort CPU time : 0.589238 secs
write stdout CPU time : 0.842179 secs
total CPU time : 3.30369 secs
total wall clock time : 3 secs
> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec (fixed string length=6) start
get_properties CPU time : 1.95909 secs
emplace set sort CPU time : 0.610479 secs
write stdout CPU time : 0.959859 secs
total CPU time : 3.52965 secs
total wall clock time : 4 secs
> ./llil2vec big1.txt big2.txt big3.txt >f.tmp
llil2vec (fixed string length=6) start
get_properties CPU time : 1.86549 secs
emplace set sort CPU time : 0.608097 secs
write stdout CPU time : 0.857176 secs
total CPU time : 3.33087 secs
total wall clock time : 3 secs
> diff f.tmp vec.tmp
##
##
// llil2vec.cpp.
// Vector version of llil2grt.pl.
// g++ compile on Linux:
// g++ -o llil2vec -std=c++11 -Wall -O3 llil2vec.cpp
// This g++ command also works with mingw C++ compiler (https://sourceforge.net/projects/mingw-w64)
// that comes bundled with Strawberry Perl (C:\Strawberry\c\bin\g++.exe).
// Example run: llil2vec big1.txt big2.txt big3.txt >vec.tmp
// Experiment with fast_io.h (see [id://11149504] by [marioroy])
#if 0
#include
#else
#include
#endif
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
static_assert(sizeof(size_t) == sizeof(int64_t), "size_t too small, need a 64-bit compile");
// ----------------------------------------------------------------------------
// Crude hack to see Windows Private Bytes in Task Manager by sleeping at
// program end (see also sleep hack at end of main)
// #include
// #include
// ----------------------------------------------------------------------------
typedef long long llil_int_type;
// Note: all words in big1.txt, big2.txt, big3.txt are <= 6 chars in length
// To use (limited length) fixed length strings uncomment the next line
// big.txt max word length is 6
// long.txt max word length is 208
// Based on rough benchmarking, this short fixed string hack
// is only worth trying for MAX_STR_LEN_L up to about 64
// #define MAX_STR_LEN_L 6
#ifdef MAX_STR_LEN_L
using str_type = std::array;
#else
using str_type = std::string;
#endif
using str_int_type = std::pair;
using int_str_type = std::pair;
using vec_str_int_type = std::vector;
using vec_int_str_type = std::vector;
using set_int_str_type = std::set;
// Mimic the Perl get_properties subroutine ----------------------------
// Limit line length and use ANSI C functions to try to boost performance
#define MAX_LINE_LEN_L 255
static void get_properties(
int nfiles, // in: the number of input files
char* fname[], // in: the input file names
vec_int_str_type& vec_ret) // out: a vector of properties
{
FILE* fh;
char line[MAX_LINE_LEN_L+1];
char* word;
llil_int_type count;
for (int i = 0; i < nfiles; ++i) {
fh = ::fopen(fname[i], "r");
if (fh == NULL) {
std::cerr << "Error opening '" << fname[i] << "' : errno=" << errno << "\n";
continue;
}
while ( ::fgets(line, MAX_LINE_LEN_L, fh) != NULL ) {
word = ::strtok(line, "\t");
count = ::atoll( ::strtok(NULL, "\n") );
#ifdef MAX_STR_LEN_L
// str_type fixword { { '\0', '\0', '\0', '\0', '\0', '\0', '\0' } };
str_type fixword;
::memset( fixword.data(), '\0', MAX_STR_LEN_L+1 );
::memcpy( fixword.data(), word, strlen(word) );
vec_ret.emplace_back( -count, fixword );
#else
vec_ret.emplace_back( -count, word );
#endif
}
::fclose(fh);
}
// Needs to be sorted by word for later sum of adjacent count fields to work
std::sort( vec_ret.begin(), vec_ret.end(),
[](const int_str_type& left, const int_str_type& right) { return left.second < right.second; }
);
}
// ---------------------------------------------------------------------
int main(int argc, char* argv[])
{
if (argc < 2) {
std::cerr << "usage: llil2vec file1 file2 ... >out.txt\n";
return 1;
}
#ifdef MAX_STR_LEN_L
std::cerr << "llil2vec (fixed string length=" << MAX_STR_LEN_L << ") start\n";
#else
std::cerr << "llil2vec start\n";
#endif
time_t tstart1 = ::time(NULL);
clock_t cstart1 = ::clock();
// Create the vector of properties
vec_int_str_type vec_ret;
get_properties(argc - 1, &argv[1], vec_ret);
clock_t cend1 = ::clock();
double ctaken1 = (double) (cend1 - cstart1) / (double)CLOCKS_PER_SEC;
std::cerr << "get_properties CPU time : " << ctaken1 << " secs\n";
clock_t cstart2 = ::clock();
// To avoid calling sort(), create an inverted std::set container
// Note: negative count gives desired ordering
set_int_str_type myset;
auto it = vec_ret.cbegin();
str_type name_last = it->second;
llil_int_type count = it->first;
for (++it; it != vec_ret.cend(); ++it) {
if ( it->second == name_last ) {
count += it->first;
}
else {
myset.emplace_hint( myset.end(), count, name_last );
name_last = it->second;
count = it->first;
}
}
myset.emplace_hint( myset.end(), count, name_last );
clock_t cend2s = ::clock();
// Output the (already sorted) std::set - no sort() function required
// Note: fix up negative count via -n.first
#ifdef MAX_STR_LEN_L
// If name is not NULL-terminated:
// for ( auto const& n : myset ) ::printf( "%.*s\t%lld\n", MAX_STR_LEN_L, n.second.data(), -n.first );
// If name is NULL-terminated:
// for ( auto const& n : myset ) ::printf( "%s\t%lld\n", n.second.data(), -n.first );
for ( auto const& n : myset ) std::cout << n.second.data() << '\t' << -n.first << '\n';
#else
// for ( auto const& n : myset ) ::printf( "%s\t%lld\n", n.second.c_str(), -n.first );
for ( auto const& n : myset ) std::cout << n.second << '\t' << -n.first << '\n';
#endif
clock_t cend2 = ::clock();
time_t tend2 = ::time(NULL);
long ttaken = static_cast(::difftime(tend2, tstart1) + 0.5);
double ctaken = (double) (cend2 - cstart1) / (double)CLOCKS_PER_SEC;
double ctaken2s = (double) (cend2s - cstart2) / (double)CLOCKS_PER_SEC;
double ctaken2o = (double) (cend2 - cend2s) / (double)CLOCKS_PER_SEC;
std::cerr << "emplace set sort CPU time : " << ctaken2s << " secs\n";
std::cerr << "write stdout CPU time : " << ctaken2o << " secs\n";
std::cerr << "total CPU time : " << ctaken << " secs\n";
std::cerr << "total wall clock time : " << ttaken << " secs\n";
// Hack to see Private Bytes in Windows Task Manager (uncomment next line so process doesn't exit too quickly)
// std::this_thread::sleep_for(std::chrono::milliseconds(90000000));
return 0;
}
##
##
// llil3vec.cpp.
// Vector version of llil2grt.pl.
// g++ compile on Linux:
// g++ -o llil3vec -std=c++20 -Wall -O3 llil3vec.cpp
// g++ -o llil3vec -std=c++20 -fopenmp -Wall -O3 llil3vec.cpp
// This g++ command also works with mingw C++ compiler (https://sourceforge.net/projects/mingw-w64)
// that comes bundled with Strawberry Perl (C:\Strawberry\c\bin\g++.exe).
// Example run: llil3vec big1.txt big2.txt big3.txt >vec.tmp
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#if defined(_OPENMP)
#include
#include
#else
#include
#include
#endif
#include
#include
#include
static_assert(sizeof(size_t) == sizeof(int64_t), "size_t too small, need a 64-bit compile");
// ----------------------------------------------------------------------------
// Crude hack to see Windows Private Bytes in Task Manager by sleeping at
// program end (see also sleep hack at end of main)
// #include
// #include
// ----------------------------------------------------------------------------
typedef long long llil_int_type;
// Note: all words in big1.txt, big2.txt, big3.txt are <= 6 chars in length
// To use (limited length) fixed length strings uncomment the next line
// big.txt max word length is 6
// long.txt max word length is 208
// Based on rough benchmarking, this short fixed string hack
// is only worth trying for MAX_STR_LEN_L up to about 64
#define MAX_STR_LEN_L 6
#ifdef MAX_STR_LEN_L
using str_type = std::array;
#else
using str_type = std::string;
#endif
using str_int_type = std::pair;
using int_str_type = std::pair;
using vec_str_int_type = std::vector;
using vec_int_str_type = std::vector;
using set_int_str_type = std::set;
// Mimic the Perl get_properties subroutine ----------------------------
// Limit line length and use ANSI C functions to try to boost performance
#define MAX_LINE_LEN_L 255
static void get_properties(
int nfiles, // in: the number of input files
char* fname[], // in: the input file names
vec_int_str_type& vec_ret) // out: a vector of properties
{
FILE* fh;
char line[MAX_LINE_LEN_L+1];
char* found;
llil_int_type count;
for (int i = 0; i < nfiles; ++i) {
fh = ::fopen(fname[i], "r");
if (fh == NULL) {
std::cerr << "Error opening '" << fname[i] << "' : errno=" << errno << "\n";
continue;
}
while ( ::fgets(line, MAX_LINE_LEN_L, fh) != NULL ) {
found = ::strchr(line, '\t');
count = ::atoll( &line[found - line + 1] );
line[found - line] = '\0'; // word
#ifdef MAX_STR_LEN_L
str_type fixword { { '\0', '\0', '\0', '\0', '\0', '\0' } };
::memcpy( fixword.data(), line, found - line );
vec_ret.emplace_back( -count, fixword );
#else
vec_ret.emplace_back( -count, line );
#endif
}
::fclose(fh);
}
// Needs to be sorted by word for later sum of adjacent count fields to work
#if defined(_OPENMP)
__gnu_parallel::sort(
#else
std::sort(
// std::sort(std::execution::par_unseq,
#endif
vec_ret.begin(), vec_ret.end(),
[](const int_str_type& left, const int_str_type& right) { return left.second < right.second; }
);
}
// ---------------------------------------------------------------------
int main(int argc, char* argv[])
{
if (argc < 2) {
std::cerr << "usage: llil3vec file1 file2 ... >out.txt\n";
return 1;
}
#ifdef MAX_STR_LEN_L
std::cerr << "llil3vec (fixed string length=" << MAX_STR_LEN_L << ") start\n";
#else
std::cerr << "llil3vec start\n";
#endif
time_t tstart1 = ::time(NULL);
clock_t cstart1 = ::clock();
// Create the vector of properties
vec_int_str_type vec_ret;
get_properties(argc - 1, &argv[1], vec_ret);
clock_t cend1 = ::clock();
double ctaken1 = (double) (cend1 - cstart1) / (double)CLOCKS_PER_SEC;
std::cerr << "get_properties CPU time : " << ctaken1 << " secs\n";
clock_t cstart2 = ::clock();
// To avoid calling sort(), create an inverted std::set container
// Note: negative count gives desired ordering
set_int_str_type myset;
auto it = vec_ret.cbegin();
str_type name_last = it->second;
llil_int_type count = it->first;
for (++it; it != vec_ret.cend(); ++it) {
if ( it->second == name_last ) {
count += it->first;
}
else {
myset.emplace_hint( myset.end(), count, name_last );
name_last = it->second;
count = it->first;
}
}
myset.emplace_hint( myset.end(), count, name_last );
clock_t cend2s = ::clock();
// Output the (already sorted) std::set - no sort() function required
// Note: fix up negative count via -n.first
#ifdef MAX_STR_LEN_L
// If name is not NULL-terminated:
for ( auto const& n : myset ) ::printf( "%.*s\t%lld\n", MAX_STR_LEN_L, n.second.data(), -n.first );
// If name is NULL-terminated:
// for ( auto const& n : myset ) ::printf( "%s\t%lld\n", n.second.data(), -n.first );
// for ( auto const& n : myset ) std::cout << n.second.data() << '\t' << -n.first << '\n';
#else
// Can try printf vs std::cout to see which is faster
for ( auto const& n : myset ) ::printf( "%s\t%lld\n", n.second.c_str(), -n.first );
// for ( auto const& n : myset ) std::cout << n.second << '\t' << -n.first << '\n';
#endif
clock_t cend2 = ::clock();
time_t tend2 = ::time(NULL);
long ttaken = static_cast(::difftime(tend2, tstart1) + 0.5);
double ctaken = (double) (cend2 - cstart1) / (double)CLOCKS_PER_SEC;
double ctaken2s = (double) (cend2s - cstart2) / (double)CLOCKS_PER_SEC;
double ctaken2o = (double) (cend2 - cend2s) / (double)CLOCKS_PER_SEC;
std::cerr << "emplace set sort CPU time : " << ctaken2s << " secs\n";
std::cerr << "write stdout CPU time : " << ctaken2o << " secs\n";
std::cerr << "total CPU time : " << ctaken << " secs\n";
std::cerr << "total wall clock time : " << ttaken << " secs\n";
// Hack to see Private Bytes in Windows Task Manager (uncomment next line so process doesn't exit too quickly)
// std::this_thread::sleep_for(std::chrono::milliseconds(90000000));
return 0;
}
##
##
> g++ -o llil3vec -std=c++20 -Wall -O3 llil3vec.cpp
> time ./llil3vec big1.txt big2.txt big3.txt >f.tmp
llil3vec (fixed string length=6) start
get_properties CPU time : 1.47895 secs
emplace set sort CPU time : 0.592783 secs
write stdout CPU time : 0.832071 secs
total CPU time : 2.90392 secs
total wall clock time : 3 secs
real 0m3.217s
user 0m2.806s
sys 0m0.412s
> diff f.tmp vec.tmp
> g++ -o llil3vec -std=c++20 -fopenmp -Wall -O3 llil3vec.cpp
> time ./llil3vec big1.txt big2.txt big3.txt >f.tmp
llil3vec (fixed string length=6) start
get_properties CPU time : 2.5809 secs
emplace set sort CPU time : 0.793355 secs
write stdout CPU time : 0.860855 secs
total CPU time : 4.23521 secs
total wall clock time : 3 secs
real 0m2.673s
user 0m4.073s
sys 0m0.465s
> diff f.tmp vec.tmp
##
##
// llil3grt.cpp.
// Inspired by llilgrt.pl: improve sort performance via a negative count.
// g++ compile on Linux:
// g++ -o llil3grt -std=c++20 -Wall -O3 llil3grt.cpp
// g++ -o llil3grt -std=c++20 -fopenmp -Wall -O3 llil3grt.cpp
// This g++ command also works with mingw C++ compiler (https://sourceforge.net/projects/mingw-w64)
// that comes bundled with Strawberry Perl (C:\Strawberry\c\bin\g++.exe).
// Example run: llil3grt tt1.txt tt2.txt tt3.txt >out.txt
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
##
// llil3m.cpp. C++ 11 version of Perl llil.pl.
// C++ version inspired by Mario's famous dualvar two-sort Perl solution.
// g++ compile on Linux:
// g++ -o llil3m -std=c++20 -Wall -O3 llil3m.cpp
// g++ -o llil3m -std=c++20 -fopenmp -Wall -O3 llil3m.cpp
// This g++ command also works with mingw C++ compiler (https://sourceforge.net/projects/mingw-w64)
// that comes bundled with Strawberry Perl (C:\Strawberry\c\bin\g++.exe).
// Example run: llil3m tt1.txt tt2.txt tt3.txt >out.txt
// Uncomment next line to mimic marioroy two sort trick (needs stable sort)
#define MARIOROY_TWO_SORT_TRICK_L 1
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
#include
##
llil3grt (fixed string length=6) start
- openmp version
get_properties CPU time : 2.34487 secs
emplace set sort CPU time : 0.942098 secs
write stdout CPU time : 0.855157 secs
total CPU time : 4.14223 secs
total wall clock time : 4 secs
real 0m4.752s
user 0m4.241s
sys 0m0.511s
##
##
$ time ./llil3m big1.txt big2.txt big3.txt >f.tmp
llil3m (fixed string length=6) start
- openmp version
get_properties CPU time : 2.40089 secs
vector copy CPU time : 0.544294 secs
vector stable sort CPU time : 5.84981 secs
write stdout CPU time : 0.805509 secs
total CPU time : 9.6006 secs
total wall clock : 4 secs
real 0m4.713s
user 0m9.238s
sys 0m0.679s
##
##
llil3m (fixed string length=6) start
- openmp version
get_properties CPU time : 2.3031 secs
vector copy CPU time : 0.544613 secs
vector sort CPU time : 2.64047 secs
write stdout CPU time : 0.791186 secs
total CPU time : 6.27946 secs
total wall clock : 4 secs
real 0m4.182s
user 0m6.128s
sys 0m0.473s
##
##
$ time ./llil3m big1.txt big2.txt big3.txt >f.tmp
llil3m (fixed string length=6) start
get_properties CPU time : 2.28072 secs
vector copy CPU time : 0.471632 secs
vector stable sort CPU time : 0.735042 secs
write stdout CPU time : 0.67514 secs
total CPU time : 4.16263 secs
total wall clock : 4 secs
real 0m4.470s
user 0m4.159s
sys 0m0.311s
$ time ./llil3m big1.txt big2.txt big3.txt >f.tmp
llil3m (fixed string length=6) start
get_properties CPU time : 2.30618 secs
vector copy CPU time : 0.473185 secs
vector sort CPU time : 1.13081 secs
write stdout CPU time : 0.668702 secs
total CPU time : 4.57897 secs
total wall clock : 4 secs
real 0m4.889s
user 0m4.558s
sys 0m0.331s
$ time ./llil3vec big1.txt big2.txt big3.txt >f.tmp
llil3vec (fixed string length=6) start
get_properties CPU time : 1.46864 secs
emplace set sort CPU time : 0.630914 secs
write stdout CPU time : 0.847912 secs
total CPU time : 2.9476 secs
total wall clock time : 3 secs
real 0m3.233s
user 0m2.852s
sys 0m0.381s
$ time ./llil3grt big1.txt big2.txt big3.txt >f.tmp
llil3grt (fixed string length=6) start
get_properties CPU time : 2.34418 secs
emplace set sort CPU time : 0.901415 secs
write stdout CPU time : 0.90025 secs
total CPU time : 4.14595 secs
total wall clock time : 5 secs
real 0m4.784s
user 0m4.232s
sys 0m0.551s