As you have noticed in my Scalable Parallel C++ Conjugate Gradient
Linear System Solver Library i am not using a Scaling technic nor
a conditioning technic, because if you have noticed
in the scaling technic on the paper above, you have to multiply
two times a matrix by a matrix and this is expenssive,
the conditioning technic is expenssive also, and because also
my Scalable Parallel C++ Conjugate Gradient Linear System Solver
Library is scalable on NUMA architecture.
So hope you will be happy with my scalable library.
You can download my Scalable Parallel C++ Conjugate Gradient Linear
System Solver Library from: