This repository provides the source code for a fast parallel in-place merge algorithm, which is related to the submitted paper "On the improvement of the in-place merge algorithm parallelization".
The content of the directory is the following:
- LICENSE
- main.cpp : the source code of the implementation (including the tests and benchmarks)
- README.md : this file
- tocompile.sh : a simple command to compile the main.cpp file
- torun.sh : the command to run the benchmark
- results/ : the results shown in the paper
$ cat tocompile.sh
g++ -DNDEBUG -std=c++11 -O3 -mtune=native -march=native main.cpp -o test.exe -fopenmpThe binary accept a option.
$ ./test.exe
[Help] ./test.exe [-reorder;-merge;-time;-find]-reorder, -merge will run some tests to validate the results.
-find will run some tests to validate the results in finding the pivots.
-time will run the benchmarks and it is advised to execute them while binding the thread.