The performance of the density functional theory (DFT) methods with different gradient corrections as an approach for the computation of transition metal complexes has been evaluated. As a test, the structures, binding energies, and vibrational frequencies of a series of binary transition metal carbonyl complexes were calculated. Comparison with previous studies shows that the gradient correction significantly improves the performance of the DFT schemes, and that the results obtained generally match the quality of the data obtained from coupled cluster and pair functional methods.

This paper presents a comparison between density functional theory local density approximation (LDA) and Hartree–Fock approximation (HFA) calculations of dipole moments, polarizabilities, and first hyperpolarizabilities, using ‘‘comparable’’ basis sets, in order to assess the relative quality of the LDA and the HFA for calculating these properties. Specifically, calculations were done using basis sets of roughly double or triple zeta plus polarization quality, with and without added field-induced polarization (FIP) functions, for the seven small molecules H2, N2, CO, CH4, NH3, H2O, and HF, using the HFA option in the program HONDO8 and the LDA options in the programs DMol and deMon. For the calculations without FIP functions, the results from HONDO8 HFA and deMon LDA, both of which use Gaussian basis sets, are very similar, while DMol, which uses a LDA numerical atomic orbital basis set, gives substantially better results. Adding FIP functions does much to alleviate these observed basis set artifacts and improves agreement with experiment. With FIP functions, the results from the two sets of LDA calculations (deMon and DMol) are very similar to each other, but differ markedly from the HFA results, and the LDA results are in significantly better agreement with experiment. This is particularly true for the hyperpolarizabilities. This appears to be the first detailed study of DFT calculations of molecular first hyperpolarizabilities. We note that closer attention to numerical details of the finite field calculation of β⇊ is necessary than would usually be the case with traditional ab initio methods. A proof that the Hellmann–Feynman theorem holds for Kohn–Sham calculations is included in the Appendix.

The computer industry has a problem. As Moore's law marches on, it will be exploited to double cores, not frequencies. But all those cores, growing to 8, 16 and beyond over the next several years, are of little value without parallel software. Where will this come from? With few exceptions, only graduate students and other strange people write parallel software. Even for numerically intensive applications, where parallel algorithms are well understood, professional software engineers almost never write parallel software. Somehow we need to (1) design many core systems programmers can actually use and (2) provide programmers with parallel programming environments that work. The good news is we have 25+ years of history in the HPC space to guide us. The bad news is that few people are paying attention to this experience. This talk looks at the history of parallel computing to develop a set of anecdotal rules to follow as we create manycore systems and their programming environments. A common theme is that just about every mistake we could make has already been made by someone. So rather than reinvent these mistakes, let's learn from the past and "do it right this time".

mDensity functional theory (DFT) was used to study reactions involving small molecules. Relative energies of isomers and transition structures of diazene, formaldehyde, and methylenimine were determined using various DFT functionals and results were compared with MP2 and MP4 calculations. DFT reaction barriers were found to be consistently lower.For some reactions, such as OH + H, + H,O + H, gradient-corrected functionals predict very low or nonexistent barriers. The hybrid Hartree-Fock-DFT adiabatic connection method (ACM) often provides much better results in such cases. The performance of several density functionals, including ACM, was tested in calculations on over 100 atomization, hydrogenation, bond dissociation, and isodesmic reactions. The ACM functional provides consistently better geometries and reaction energetics than does any other functional studied. In cases where both HF and gradient-corrected DFT methods underestimate bond distances, the ACM geometries may be inferior to those predicted by gradient-corrected DFT methods.

We describe the implementation of the mesh-based first-principles density functional code DMol on nCUBE 2 parallel computers. The numerical mesh nature of DMol makes it naturally suited for a massively parallel computational environment. Our parallelization strategy consists of a domain decomposition of mesh points. This evenly distributes mesh points to all available processors and leads to a substantial computational speedup with limited communication overhead and good node balancing. To achieve better performance and circumvent memory storage limitations, the torus wrap method is used to distribute both the Hamiltonian and overlap matrices, and a parallel matrix diagonalization routine is employed to calculate eigenvalues and eigenvectors. Benchmark calculations on a 128-node nCUBE 2 are presented. Wiley & Sons, Inc. 0 1995 by John properties of molecules, clusters, solids, and surfaces,',' Compared to traditional quantum chemistry methods (such as Hartree-Fock), density functional theory in the local density approximation (LDA)3,4 offers computational advantages due to its approximately third-power dependence on the number of atoms and basis orbital^.^ It is also reasonably accurate because electron correlation is included in the formalism. Such calculations are still computationally intensive and require sub-