A NEW DIRECT MP2 GRADIENT ALGORITHM WITH IMPLEMENTATION ON A MASSIVELY-PARALLEL COMPUTER

A new direct second-order Moller-Plesset gradient algorithm is present ed. By avoiding generation of molecular orbital integrals with three v irtual indices, the memory requirement for a calculation with n basis functions and o occupied orbitals is reduced to on(2), and overall com putational savings are obtained. Tile algorithm hss been implemented w ith distributed data on a massively parallel computer, the Intel Parag on. The parallelization scheme provides high parallel efficiency, and the performance of the algorithm is illustrated for calculations runni ng on up to 64 processors.