add multithread-bfs

[suncanghuai]

• implement multi-thread bfs
• add test cases for multi-thread bfs(in BFSTest.cpp)
• add benchmark for multi-thread bfs(in BFS_BM.cpp)

I actually implement a much more simple multithread-bfs with OpenMP at first, which is very similar to the pseudocode of the paper from the corresponding issue.
But there are two problems with that version of implementation:

• pseudo-multi-thread (parameter num_threads = 1) has a performance advantage over true multi-thread in each benchmark test, even when a graph is large.
• it forces users to load the OpenMP package when they compile their projects

Finally, I change the code to a version with std::threads using condition_variable to synchronize threads and do some code optimization. Although the result of the benchmark is still not good as I expected, at least in the CitHepPh graph true multi-thread performs better than pseudo-multi-thread.

[ghost]

👇 Click on the image for a new way to code review

• Make big changes easier — review code in small groups of related files

• Know where to start — see the whole change at a glance

• Take a code tour — explore the change with an interactive tour

• Make comments and review — all fully sync’ed with github

  Try it now!

Legend

[codecov-commenter]

Codecov Report

Merging #252 (7a9212d) into master (588f64f) will decrease coverage by 0.11%.
The diff coverage is 97.82%.

@@            Coverage Diff             @@
##           master     #252      +/-   ##
==========================================
- Coverage   96.90%   96.79%   -0.12%     
==========================================
  Files          47       47              
  Lines        5500     5740     +240     
==========================================
+ Hits         5330     5556     +226     
- Misses        170      184      +14     

Flag	Coverage Δ
unittests	96.79% <97.82%> (-0.12%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
include/Graph/Graph.hpp	93.59% <94.93%> (+0.12%)	⬆️
test/BFSTest.cpp	99.61% <99.33%> (-0.39%)	⬇️
include/Partitioning/EdgeBalancedVertexCut.hpp	74.50% <0.00%> (-5.89%)	⬇️
include/Partitioning/GreedyVertexCut.hpp	84.69% <0.00%> (-5.11%)	⬇️
include/Partitioning/EBV.hpp	83.58% <0.00%> (-4.48%)	⬇️
include/Partitioning/WeightBalancedLibra.hpp	77.77% <0.00%> (-0.80%)	⬇️
include/Partitioning/CoordinatedPartitionState.hpp	75.96% <0.00%> (+1.21%)	⬆️
include/Partitioning/Partitioner.hpp	100.00% <0.00%> (+2.04%)	⬆️
include/Partitioning/HDRF.hpp	81.31% <0.00%> (+2.19%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[ZigRazor]

Before I can accept this PR I think you should:

• Give to this function an ifdef condition ( also in the cmake files ), in this way the library can be used with/without this functionality and OpenMP library;
• Report the result in the readme, and explain the current problem of performances, in this way the users can choose better and as reminder for next study;
• If possible correct some issue of the static code analysis ( this is optional )

Thank you in advance @suncanghuai

[suncanghuai]

@ZigRazor perhaps I haven't described it clearly. I have changed the way multithread is implemented from “omp parallel” to std::thread and std::condition_variable, the version I apply to be merged is without OpenMP. The reason I talk about "omp parallel" is to explain why I didn't implement multithread-bfs exactly following the pseudocode of the paper. Besides, some benchmark results really should be added in the readme file so that users could know when to use multithread-bfs instead of bfs. And I will look into issues of static analysis, but may not fix them all.

[ZigRazor]

@suncanghuai Ok, i misunderstand, now is clear.
Good work and Thank you

[suncanghuai]

@ZigRazor I'm gonna add something to readme file and fix some issus of static analysis in this MR,but you have merged it.So I suppose I have to put them in my next MR.

[ZigRazor]

Yes, thank you

[ZigRazor]

@suncanghuai given you contributions and your follow to the project I think you can be upgraded as collaborator of the project, what do you think about? You want to become collaborator?

[suncanghuai]

@suncanghuai given you contributions and your follow to the project I think you can be upgraded as collaborator of the project, what do you think about? You want to become collaborator?

Sure, I'm glad to work for this project as a collaborator.