performance measurements

Each table row shows performance measurements for this C++ g++ program with a particular command-line input value N.

 N  CPU secs Elapsed secs Memory KB Code B ≈ CPU Load
250,0000.520.5311,3521252  2% 2% 0% 100%
2,500,0002.672.6822,1401252  1% 1% 3% 100%
25,000,00024.0324.04138,9601252  1% 1% 0% 100%

Read the ↓ make, command line, and program output logs to see how this program was run.

Read k-nucleotide benchmark to see what this program should do.

 notes

gcc version 4.8.2 (Ubuntu 4.8.2-19ubuntu1)

 k-nucleotide C++ g++ #3 program source code

/* The Computer Language Benchmarks Game

   http://benchmarksgame.alioth.debian.org/



   Contributed by Branimir Maksimovic

*/

// g++ 4.8.x bug, compile with: -Wl,--no-as-needed option 


#include <iostream>

#include <iomanip>

#include <cstdint>

#include <cstdio>

#include <string>

#include <cstring>

#include <algorithm>

#include <map>

#include <ext/pb_ds/assoc_container.hpp>

#include <future>

#include <unistd.h>


unsigned char tonum[256],tochar[4];
static void init()
{
   tonum['A'] = 0;
   tonum['C'] = 1;
   tonum['T'] = 2;
   tonum['G'] = 3;
   tochar[0] = 'A';
   tochar[1] = 'C';
   tochar[2] = 'T';
   tochar[3] = 'G';
}

struct T{
   T(const std::string& s = std::string())
   :data(0),size(s.size())
   {
      reset(s,0,s.size());
   }
   void reset(const std::string& s,unsigned beg,unsigned end)
   {
      size = end-beg;
      data = 0;
      for(unsigned i = beg; i != end; ++i)
      {
         data <<= 2;
         data |= tonum[unsigned(s[i])];
      }
   }
   bool operator<(const T& in)const
   {
      return data < in.data;
   }
   bool operator==(const T& in)const
   {
      return data == in.data;
   }
   operator std::string()const
   {
      std::string tmp;
      uint64_t tmp1 = data;
      for(unsigned i = 0;i!=size;++i)
      {
         tmp+=tochar[tmp1 & 3];
         tmp1 >>= 2;
      }
      std::reverse(tmp.begin(),tmp.end());
      return tmp;
   }
   struct hash{
   uint64_t operator()(const T& t)const{ return t.data; }
   };
   uint64_t data;
   unsigned char size;
};

__gnu_pbds::cc_hash_table<T,unsigned,T::hash>
calculate(const std::string& input,unsigned size, unsigned beg=0,unsigned incr=1)
{
   __gnu_pbds::cc_hash_table<T,unsigned,T::hash> frequencies;
   T tmp;
   for (unsigned i = beg, i_end = input.size() + 1 - size; i < i_end; i+=incr)
   {
     tmp.reset(input,i,i+size);
      ++frequencies[tmp];
   }
   return frequencies;
}

__gnu_pbds::cc_hash_table<T,unsigned,T::hash>
tcalculate(const std::string& input,unsigned size)
{
   unsigned N = sysconf (_SC_NPROCESSORS_ONLN);

   std::future<__gnu_pbds::cc_hash_table<T,unsigned,T::hash>> ft[N];
   for(unsigned i = 0; i<N;++i)
      ft[i] = std::async(std::launch::async,calculate,std::ref(input),size,i,N);

   auto frequencies = ft[0].get();

   for(unsigned i = 1 ; i<N; ++i)
      for(auto& j : ft[i].get())
      {
         frequencies[j.first]+=j.second;
      }
   return frequencies;
}

void write_frequencies(const std::string & input, unsigned size)
{
   unsigned sum = input.size() + 1 - size;
   auto frequencies = tcalculate(input,size);
   std::map<unsigned, std::string,std::greater<unsigned>> freq;
   for(auto& i: frequencies)
   {
      freq.insert(std::make_pair(i.second,i.first));
   }
   for(auto& i : freq)
      std::cout << i.second << ' ' << (sum ? double(100 * i.first) / sum : 0.0) << '\n';
   std::cout << '\n';
}

void write_count(const std::string & input, const std::string& string)
{
   unsigned size = string.size();
   auto frequencies = tcalculate(input,size);

   std::cout << frequencies[string] << '\t' << string << '\n';
}

int main()
{
   init();
   std::string input;
   char buffer[256];
   while (fgets(buffer,100,stdin) && memcmp(">THREE",buffer,6)!=0);
   while (fgets(buffer,100,stdin) && buffer[0] != '>')
   {
      if (buffer[0] != ';')
      {
         input.append(buffer,strlen(buffer)-1);
      }
   }
   std::transform(input.begin(),input.end(),input.begin(),::toupper);

   std::cout << std::setprecision(3) << std::setiosflags(std::ios::fixed);
   write_frequencies(input,1);
   write_frequencies(input,2);
   write_count(input, "GGT");
   write_count(input, "GGTA");
   write_count(input, "GGTATT");
   write_count(input, "GGTATTTTAATT");
   write_count(input, "GGTATTTTAATTTATAGT");
}

 make, command-line, and program output logs

Thu, 24 Apr 2014 02:08:22 GMT

MAKE:
/usr/bin/g++ -c -pipe -O3 -fomit-frame-pointer -march=native  -std=c++0x knucleotide.gpp-3.c++ -o knucleotide.gpp-3.c++.o &&  \
        /usr/bin/g++ knucleotide.gpp-3.c++.o -o knucleotide.gpp-3.gpp_run -Wl,--no-as-needed -lpthread 
rm knucleotide.gpp-3.c++
3.02s to complete and log all make actions

COMMAND LINE:
./knucleotide.gpp-3.gpp_run 0 < knucleotide-input25000000.txt

PROGRAM OUTPUT:
A 30.295
T 30.151
C 19.800
G 19.754

AA 9.177
TA 9.132
AT 9.131
TT 9.091
CA 6.002
AC 6.001
AG 5.987
GA 5.984
CT 5.971
TC 5.971
GT 5.957
TG 5.956
CC 3.917
GC 3.911
CG 3.909
GG 3.902

1471758	GGT
446535	GGTA
47336	GGTATT
893	GGTATTTTAATT
893	GGTATTTTAATTTATAGT

Revised BSD license

  Home   Conclusions   License   Play