Conferences & workshops: International conference on computer design '98

Daniel Pryor, Andreas Kuehlmann, Trevor J. Smedley, David R. Kincaid, Anne C. Elster
1998 IEEE Computational Science & Engineering  
INGVAR EBBSJO ANUPAM MADHUK~R Materials and devices with microstructures on the nanometer scale are revolutionizing technology, but until recently simulation at this scale has been problematic. Developments in parallel computing are now allowing atomistic simulation using multiresolution algorithms, such as fast multipole methods. With these algorithms, researchers may soon be able to simulate applications up to one billion atoms. emiconductor feature sizes are expected to shrink to 70
more » ... s in the next 10 years in an effort to increase processing speed and memory density.' In roughly the same timeframe, engineers will be developing nanostructwed materials, materials with nanometer grain sizes (one nanometer is one billionth of a meter). Nanostructured materials promise to dramatically increase material strength and durability.* At feature sizes smaller than 100 nm, however, conventional theoretical modeling and computer simulations, supplemented with atomistic models. With conventional approaches, such as the finite-element method based on linear elasticity, modelers cannot include the effects of defects, surfaces, and other microstructures-all of which fundamentally affect the material's performance. Atomistic simulations, in particular those based on molecular dynamics, capture these microstructural effects as well as highly nonhomogeneous stresses a t the atomic scale. These stresses are at the heart of many advanced semiconductor technologies. such a s those based on continuum approaches, must be Simulations based on state-of-the-art molecular dynamics can involve up to 100 million atoms, but many researchers believe that petaflop computers will be built by the end of the next decade (one petaflop is 10" floating-point operations per second). Such computers will make it possible to simulate nanostructures that involve 1OI2 atoms-with a total system size ofwell over one micron. This in turn means that modelers can simulate an entire device structure atomistically. 68
doi:10.1109/99.735897 fatcat:tuiuihcoqfgybb47geoio5mlc4