The design and implementation of a concurrent circuit simulation program for multicomputers

The development and implementation of a concurrent circuit simulation program, CONCISE, for MOS circuits, which is targeted for multicomputers (that is, message-passing, concurrent computers), is described. The objective is to design a program that uses hundreds of processing nodes efficiently. Implementation issues which are of importance for the performance on a concurrent computer are discussed. The waveform relaxation method with full-window techniques and Jacobi iterations is used because of its simple scheduler and limited demands on the message-passing network of the multicomputer. To speed up convergence, time windows are placed at transitions in the input waveforms, and dynamic window splitting at the nonconvergence time point is used. Results of experiments performed on a 192-node multicomputer showing that CONCISE can efficiently exploit hundreds of nodes are presented. For large and medium-size problems the speedup is almost two orders of magnitude. It is shown that CONCISE can efficiently handle very large problems with 12000 circuit nodes. >

[1]  J. Gillis,et al.  Matrix Iterative Analysis , 1961 .

[2]  A. Jimenez,et al.  Algorithms for ASTAP--A network-analysis program , 1973 .

[3]  Randal E. Bryant,et al.  A Switch-Level Model and Simulator for MOS Digital Systems , 1984, IEEE Transactions on Computers.

[4]  Jacob K. White,et al.  Relaxation Techniques for the Simulation of VLSI Circuits , 1986 .

[5]  Jakov N. Seizovic,et al.  The Reactive Kernel , 1988 .

[6]  Charles L. Seitz,et al.  The cosmic cube , 1985, CACM.

[7]  Alberto L. Sangiovanni-Vincentelli,et al.  The Waveform Relaxation Method for Time-Domain Analysis of Large Scale Integrated Circuits , 1982, IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems.

[8]  S. Mattisson,et al.  Distributed Model Evaluation for the Waveform Relaxation Method , 1990, Proceedings of the Fifth Distributed Memory Computing Conference, 1990..

[9]  Steven M. Burns,et al.  The design of an asynchronous microprocessor , 1989, CARN.

[10]  Jeffrey T. Deutsch.  Algorithms and architecture for multiprocessor-based circuit simulation , 1985 .

[11]  Sven Mattisson,et al.  Hot Clock nMOS , 1985 .

[12]  G. Amdhal,et al.  Validity of the single processor approach to achieving large scale computing capabilities , 1967, AFIPS '67 (Spring).

[13]  R. E. Oakley,et al.  CASMOS-an accurate MOS model with geometry-dependent parameters: I , 1981 .

[14]  David William Smart.  Parallel Processing Techniques for the Simulation of MOS VLSI Circuits Using Waveform Relaxation , 1988 .

[15]  Anthony Skjellum.  Concurrent dynamic simulation: multicomputer algorithms research applied to ordinary differential-algebraic process systems in chemical engineering , 1990 .

[16]  Ibrahim N. Hajj,et al.  Parallel circuit simulation on supercomputers , 1989 .

[17]  Lawrence Snyder,et al.  An empirical study of on-chip parallelism , 1988, 25th ACM/IEEE, Design Automation Conference.Proceedings 1988..

[18]  Walter S. Scott,et al.  1986 VLSI Tools: Still More Works by the Original Artists , 1985 .

[19]  Sven Mattisson,et al.  Circuit partitioning and iteration scheme for waveform relaxation on multicomputers , 1989, IEEE International Symposium on Circuits and Systems,.

[20]  Albert E. Ruehli,et al.  Parallel waveform relaxation of circuits with global feedback loops , 1992, [1992] Proceedings 29th ACM/IEEE Design Automation Conference.

[21]  Hugo De Man,et al.  Acceleration of relaxation-based circuit simulation using a multiprocessor system , 1990, IEEE Trans. Comput. Aided Des. Integr. Circuits Syst..