Last active
August 29, 2015 14:19
-
-
Save TalissonBento/2be6fe35e66a8732244e to your computer and use it in GitHub Desktop.
Parallel Accomulator
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#include <stdio.h> | |
#include <functional> | |
#include <memory> | |
#include <thread> | |
#include <string> | |
#include <vector> | |
#include <algorithm> | |
#include <utility> | |
#include <numeric> | |
#include <time.h> | |
using namespace std; | |
template<typename Iterator, typename T> | |
struct Accomulator_Block | |
{ | |
void operator()(Iterator first, Iterator last, T& result) | |
{ | |
result = std::accumulate(first, last, result); | |
} | |
}; | |
template<typename Iterator, typename T> | |
T Parallel_accomulate(Iterator first, Iterator last, T init) | |
{ | |
unsigned long const length = distance(first, last); | |
if (!length) return init; | |
unsigned short const Min_per_Thread = 25; | |
unsigned long const Max_threads = (length + Min_per_Thread - 1) / Min_per_Thread; //Ensure the the max threads for little length | |
unsigned long const hardware_threads = thread::hardware_concurrency(); | |
unsigned short const num_threads = std::min( hardware_threads!=0? hardware_threads : 2, Max_threads) ; | |
unsigned long const block_size = length / num_threads; | |
vector<T> th_results(num_threads); | |
vector<std::thread> threads(num_threads-1); | |
Iterator it_start; | |
Iterator it_end = first; | |
for (unsigned short i = 0; i < (num_threads - 1); ++i) | |
{ | |
it_start = it_end; | |
advance(it_end, block_size); | |
threads[i] = std::thread( Accomulator_Block<Iterator, T>(), it_start, it_end, ref(th_results[i]) ); | |
} | |
std::for_each(threads.begin(), threads.end(), mem_fn(&std::thread::join)); //Wait for jobs | |
Accomulator_Block<Iterator, T>()( it_end, last, ref(th_results[num_threads - 1]) ); | |
return accumulate(th_results.begin(), th_results.end(), init); | |
}; | |
int main(int argc, char* argv[]) | |
{ | |
clock_t t; | |
printf("Hardware core: %d\n", thread::hardware_concurrency()); | |
vector<unsigned long> data(100000000, 10); | |
t = clock(); | |
int total = Parallel_accomulate(data.begin(), data.end(), 1); | |
t = clock() - t; | |
printf("Total: %d in %.3f\n", total, (float)(t / CLOCKS_PER_SEC)); | |
total = 1; | |
t = clock(); | |
// total = accumulate(data.begin(), data.end(), total);// Faster | |
/* | |
int size = data.size(); | |
for (int it = 0; it < size; ++it) //Faster | |
{ | |
total += data[it]; | |
} | |
*/ | |
for (auto it = data.begin(); it != data.end(); it++) //Slower | |
{ | |
total += *it; | |
} | |
t = clock() - t; | |
printf("Total: %d in %.3f\n", total, (float)(t / CLOCKS_PER_SEC)); | |
getc(stdin); | |
return 0; | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment