Current processor allocation techniques for highly parallel systems are based on centralized front-end based algorithms. As a result, the applied strategies are restricted to static allocation, low parallelism and weak fault tolerance. To lift these restrictions we are investigating a distributed approach to the processor allocation problem in large mesh-connected multicomputers. A noncontiguous version of a distributed dynamic processor allocation strategy is proposed and studied in this paper as an alternative for parallel programming models that allow dynamic creation and deletion of tasks. Simulations compare the performance of the proposed dynamic strategy with the static counterpart and also with well-known centralized algorithms in such an environment with growing and shrinking processor demands. We also present the results of experiments on a Siemens hpcLine Primergy Server with 96 nodes that show dynamic allocation is feasible with current technologies.