Newbie to MKL/Pardiso. I was running my first program yesterday. Some quick questions on the right hand side B and solution X:
- is it correct that both B and X have to be allocated in every MPI process, and B must be set in every process also ?
- I am actually trying to get inverse of matrix, so the B is an identity matrix. But it seems that there is no way to specify B in a sparse format, in order to save memory. Is it true ?
thank you very much
allocation B and X will depends on iparm. For example, in the case if iparm==2, The solution vector, A, and RHS elements are distributed between processes in same manner. See documentation for more details.