Related studies of incremental memory statistics varying with parallel system scale

In parallel computing, especially in the field of massively large scale problems, such as the complex large scale electromagnetic problems, the memory requirement is extremely huge. To understand the memory shifting pattern, when the system size alternating is crucial in optimizing or utilizing a parallel program. In this paper, an incremental memory statistic model, varying with system size, is proposed, which is also available to comprehending the entire memory requirement of a program. Based on the statistic model, the relationship of memory and time is also analyzed.