Accelerating Large-Scale Interconnection Network Simulation by Cellular Automata Concept

Abstract

State-of-the-art parallel systems employ a huge number of computing nodes that are connected by an interconnection network. An interconnection network (ICN) plays an important role in a parallel system, since it is responsible to communication capability. In general, an ICN shows non-linear phenomena in its communication performance, most of them are caused by congestion. Thus, designing a large-scale parallel system requires sufficient discussions through repetitive simulation runs. This causes another problem in simulating large-scale systems within a reasonable cost. This paper shows a promising solution by introducing the cellular automata concept, which is originated in our prior work. Assuming 2D-torus topologies for simplification of discussion, this paper discusses fundamental design of router functions in terms of cellular automata, data structure of packets, alternative modeling of a router function, and miscellaneous optimization. The proposed models have a good affinity to GPGPU technology and, as representative speed-up results, the GPU-based simulator accelerates simulation upto about 1264 times from sequential execution on a single CPU. Furthermore, since the proposed models are applicable in the shared memory model, multithread implementation of the proposed methods achieve about 162 times speed-ups at the maximum.

Genetic Node-Mapping Methods for Rapid Collective Communications

“…We implemented an evaluation platform for evaluating the node-mapping methods, which is extended from our interconnection network simulator based on the cellular automata principle [6], [7] as a simulation engine. As the nodemapping methods aim at reduction of the duration time in collective communication, the major evaluation function in the genetic operation is the duration time for each gene that represents the logical mapping result.…”

“…The largest problem in the evaluation cost is that the duration time is measured only by simulation. In this paper, we use our simulator that is based on the cellular-automata principle [7]. Although the simulator performs high-speed, the performance is not sufficient for the purpose of this paper, since, in largescale networks, simulation time becomes dominant in our method.…”

“…On the other hand, the duration time metric requires a simulation run for every mapping vector x L . Although the simulation performance is considerably accelerated [7], computational cost of a simulation run is much higher than that of avg.hop. Thus, we can expect that the computational costs in the genetic operation, specifically in the evaluation functions, is reduced, if the simulation run is omitted or reduced.…”

Enhancing Cup-Stacking Method for Collective Communication

“…We assume a simple (non-pipelined) router whose organization is illustrated in Fig. 2 ( [18]). In the router model, the arbitration task in a router consists of two steps: queue selection and virtual channel (VC) selection.…”

“…The queue selection step arbitrates the requests from input queues (in input ports) to a certain virtual channel (in an output port). In our router model [18], the central crossbar-switch connects between every input queue (in input ports) and every output virtual channel (in output ports).…”