Dynamic SMP clusters with communication on the fly

Abstract

The paper presents a comparative analysis of parallel implementation of the preconditioned conjugate gradient method with symmetric successive over relaxation preconditioner. Two parallel implementations are compared. The first one is a message passing version that turned out to be inefficient when executed on a typical cluster of workstations. The other one is an efficient version simulated on a novel architecture of dynamically reconfigurable shared memory clusters with a new paradigm of inter-processor communication called communication on the fly. The presented example shows high suitability of the proposed architecture for fine grain numerical computations what can be very useful in simulation of physical phenomena described as numerical problems for fine grain parallel execution.