M. Thor, A MULTI-THREAD APPROACH REDUCING PROGRAM EXECUTION TIME IN A HETEROGENEOUS RECONFIGURABLE MULTIPROCESSOR ARCHITECTURE, Journal of systems architecture, 43(1-5), 1997, pp. 143-153
A new approach leading to reduced user program execution time in a mul
ti-processor environment with dynamic reconfiguration capabilities is
proposed in this paper. The general idea is that a user program partit
ioned into specific parts called slices (based on a program graph repr
esentation) is executed on different heterogeneous, dynamically reconf
igurable multi-processor units (MPUs). Reduction of the program execut
ion time is achieved through finding the MPU with the shortest executi
on time for each program slice and appropriate control flow between ex
ecuting units. Two different methods are proposed: . Run-time reducing
a program execution time using multiple heterogeneous MPUs. All MPUs
simultaneously start execution of each program slice. This method is u
sed for real-time applications when analytical methods are impossible
to apply, program graphs activations are undeterministic, and/or exter
nal control signals influence the program execution paradigm. . Findin
g the optimal hardware topologies of the MPUs and allocating them prog
ram slices; next a reduced time program execution through look-ahead r
econfiguration and switching of control between the MPUs. The proposed
strategy and algorithms are described in details. The methods are com
pared and the scope of their applicability is discussed. Some indicati
ons for the future investigations are given in conclusions.