Alloy Calculations on Massively Parallel Computers

We discuss ways that first principles LDA methods can be adapted to take advantage of the computational power of massively parallel supercomputers. We discuss the need for new theoretical and computational strategies that circumvent the N 3 problem associated with standard LDA implementations limiting their extention to large systems. We propose a new, spatially local, multiple scattering theory algorithm that offers the possibility of producing a scalable O(N) method. As a first step, we use the method to calculate the self-consistent charge density for a large system.