Design and framework of reduced instruction set codes for scalable computations for nonlinear structural dynamics

摘要

vspace-3ptA general framework and avenues towards the design of a unified integrated computational technology for nonlinear structural dynamics encompassing a wide variety of new and unexplored, and existing time integration operators is now possible employing the so-called Reduced Instruction Set Codes (RISC) via a unified family of generalized integration operators [GInO] towards scalable computations on massively parallel computing platforms. Whilst the RISC paradigm has a critical impact on the scientific code design and development time and efforts, it simultaneously increases the functionality of the scientific codes by many folds by providing a variety of choices to the analyst. A unified scalable computational approach towards such a computational technology is desirable for large-scale structures and large processor counts employing a message-passing paradigm (using MPI), graph partitioning techniques, and Lagrange multiplier based domain decomposition methods. Here, the focus is on the scalability analysis conducted via an integrated unified technology for [GInO] with emphasis on the family of optimal non-dissipative and dissipative algorithms for structural dynamics in conjunction with large deformation, elastic, elastic-plastic dynamic response. For geometric nonlinearity a total Lagrangian formulation, and for material nonlinearity elasto-plastic formulations are employed. This is the first time that such a general framework and capability is plausible via a unified technology and the developments further enhance computational structural dynamics areas.