Parallel Conjugate Gradient method using MPI
make clean removes the compiled executable cg.
make cg compiles the code with no funny business.
make CFLAGS=-DVERIFY cg and max error is printed when cg runs.
make CFLAGS=-DDEBUG cg and lots of debug info is printed when cg runs.
Including all matrices (if the matrix dimension is no larger than 20) and
comparison to the sequential max error. Takes much longer than other versions.
NOTE: If you compile with one of the flags, you will need to run
make clean to recompile, unless the code has been changed.
mpirun -np <p> ./cg <N>, where <p> is the desired number of processes
and <N> is the problem size - i.e. the N in the NxN matrix generated.