Figure 2 : HPC paradigms – current and future; Moore’s law and the slow-down due to the power wall.
A summary of historical developments in parallel and high-performance computing architectures is sketched in Table 3. A sketch of the early instruction for computation, including the concept of parallelism in computing, can be traced to Charles Babbage (see Table 3). Scientific computing has benefitted from the advances in chip architecture that led to the linear Moore’s law behavior for four decades from the 1980s-2010 (Figure 2, right). However, even during this golden age of increasing clock speeds and doubling of computational speed every 18 months in single-core architectures, high-performance computing broke the shackles of serial (and vectorized) computing to embrace parallel computing as a mainstream route to solve computational problems. The switch to parallel microprocessors is a game-changer in the history of computing [47] (see, http://www.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-183.html). The advances in parallel hardware and software have torpedoed the advances in multiphysics and multiresolution simulations. This convergence of high-performance computing and multiscale modeling has transformed parallel algorithms (see Table 2), which are the engines of multiphysics modeling.