The code of algorithms

• Peeling: the state-of-the-art sequential D-core decomposition algorithm;
• AC: the distributed anchored coreness-based D-core decomposition algorithm, and we parallelize it by using multi-threads;
• SC: the state-of-the-art distributed skyline coreness-based D-core decomposition algorithm, and we parallelize it by using multi-threads;
• ParPeel: our proposed parallel D-core decomposition algorithm, which is depicted in Algorithm 3;
• ParPeel+: our proposed parallel D-core decomposition algorithm in Algorithm 3 with shell-based pruning strategy in Section 5.1;
• Shell-PDC: our proposed parallel D-core decomposition algorithm, listed in Algorithm 5.

Compiling and Running

Compiling the program

g++ -std=c++17 -fopenmp ${fileDirname}/*.cpp -o ${FILE_NAME}

For example:

g++ -std=c++17 -fopenmp ./src/*.cpp -o dcore

Running the program:

./${FILE_NAME} -t ${THREAD} -f ${GRAPH_FILE} -a ${ALGORITHM}

For example:

./dcore \
-t 64 \
-f ./materials/em.txt \
-a 6 

For simply, we can run the program via a script ./run.py.

Input format

• THREAD: The number of threads.

• GRAPH_FILE: The first line is consited of # of nodes and # of directed edges in the graph, which is denoted as following.

${NODES} ${EDGE}

Remains line represents a directed edge from node u to node v, which is presented as following.

${u} ${v}

• ALGORITHM: The id of algorithm will be running, which the id is presented as following.

id	algorithms
1	Peeling
2	AC
3	SC
4	ParPeel
5	ParPeel+
6	Shell-PDC

Other

• The ten datasets used in paper are availabe from:

Name	Abbr.	Source
Email-EuAll	EM	https://snap.stanford.edu/data/index.htm
Slashdot	SD
Amazon	AM
Pokec	PO
Live Journal	LJ
Enwiki-2013	EW	https://law.di.unimi.it/index.php
Hollywood	HW
Webbase	WB
IT-2004	IT
UK-2007	UK

• We give an example of Email-EuAll in ./materials/, and running time of all algorithms in ./result/EM/.