marconms/dynamic-closure

This is a library for efficiently computing the closure of a changing graph.

Dynamic Closure

Dynamic Closure is a library for efficiently computing the closure of a changing graph.

Requirements

Name	Version
cmake	^3.5
graphblas	^7
gcc	^13

To compile the library for release:

mkdir ./Release && cd ./Release
cmake -DCMAKE_BUILD_TYPE=Release ..
make

Once the library has been compiled, the binaries will be available in the ./Release/dynamic-closure folder.

Executing Benchmarks

Benchmarks are available in the ./Release/benchmarks folder, once the library has been compiled for release.

Available benchmarks:

• generalized-floyd-warshall
• decrease-idempotent
• increase-column
• increase-row

Preloaded datasets:

• Bai/rw5151
• Lucifora/cell1
• Pajek/patents_main
• vanHeukelum//cage10

To run all benchmarks, please execute the ./Release/benchmarks/run-benchmarks binary. Benchmarks can also be executed indipendently.

Parameters

Parameter	Description	Required
-f	The path of the dataset to use.	Yes
-o	The path of the output directory.	No
-n	The number of updates to generate.	No
-s	A predefined closure to utilize for dynamic closure benchmarks. If this parameter is not specified, the closure will be computed before the benchmarks.	No

Examples (in the context of ./Release/benchmarks):

./generalized-floyd-warshall -f ../assets/cage10.mtx -o ../out

./run-benchmarks -f ../assets/cage10.mtx -o ../out -s ../out/generalized-floyd-warshall_cage-10/F.mtx

./increase-column -f ../assets/cage10.mtx -o ../out -n 1000 -s ../out/generalized-floyd-warshall_cage-10/F.mtx

./increase-row -f ../assets/cage10.mtx -o ../out -n 1000 -s ../out/generalized-floyd-warshall_cage-10/F.mtx

./decrease-idempotent -f ../assets/cage10.mtx -o ../out -n 1000 -s ../out/generalized-floyd-warshall_cage-10/F.mtx