Performance Study of Non-blocking Collective Communication Implementations Toward Adaptive Selection
Tsuyoshi Okuma, Takeshi Nanri · 2014
Abstract—Non-blocking collective communication is a key technique for enabling overlap of collective communication and computation to achieve higher scalability on large scale parallel computers. There can be many types of implementation meth-ods for non-blocking collective communications. To sufficiently obtain effect of overlap, appropriate method should be chosen. Therefore, it is important to study characteristics of each method. In this paper, we explain a characteristic of each method of im-plementation and and evaluate the characteristic by performing the experiment using a simple test program. In addition to that, this paper proposes an implementation with lower overhead, and evaluates the performance. Keywords—non-blocking collective communication, overlapping communication and computation, large scale parallel computing