''Scalability of the FR-Mesh under Different Concurrent Write Rules''
J. Alberto Fernandez-Zepeda
Jerry L. Trahan
Ramachandran Vaidyanathan
Proc. World Multiconference on Systemics, Cybernetics
and Informatics, (Caracas, Venezuela, July 1997), vol. 1, pp. 437-444
Abstract:
Algorithmic scalability captures the ability of a small-sized
instance of a model to efficiently simulate a larger-sized
instance. Recently the authors have shown that algorithms
scale very well on the FR-Mesh, a restriction of the reconfigurable
mesh (R-Mesh). This result uses a small FR-Mesh to simulate
an arbitrary step of a larger FR-Mesh. This simulation assumes that
in both the simulated and simulating FR-Meshes all concurrent
writes are of the same value (Common write).
This paper extends the above FR-Mesh scalability simulation to
other write rules such as Priority and Collision. The results of
this paper permit a single R-Mesh algorithm to use scalability
results on different models. Our results also allow write rules
that are not easy to implement on buses (such as Priority)
to be used effectively as algorithm design tools.