source: proiecte/pgraph/Readme.txt @ 157

1.Prim

To build the input file for the parallel version of Prim and the serial version
with a cost matrix, compile and run agenerate1.c .It will build a complete graph
with random values for the costs of the edges.
To build the input file for the serial version with adjacency lists first
compile and run agenerate1.c and after that agenerate2.c .The input file will
represent a graph equivalent with the one represented by the file built with
agenerate1.c .

2.Dijkstra

To build the input file for the parallel version of Dijkstra and
the serial version with a cost matrix, compile and run buildR.cpp.
It will build a complete graph with random cost edges with values between
1 and 100.
To build the input file for the serial version with adjacency lists, first
run and compile buildR.cpp and after that generate2.c.It will build an input
file that represents the same complete graph as the one created with buildR.cpp.

3.Bellman-Ford

For Bellman-Ford I recommend using as input file the grader_test10.in file because
it has the greatest number of nodes and edges and shows clearly that the
parallel version with OpenMP is inferior to the serial one because of the
thread creation and destruction overhead.You just have to rename the grader_test10.in file
as bell.in.The aproximate times for this input file were:
-For the serial version 0.5 seconds
-The parallel version 2 seconds

4.Roy-Floyd

To build the input file for the Roy-Floyd algorithms just compile and run the buildRoy.cpp file.
It will build an input file that represents a complete graph with random cost edges between 1 and 100.