Theoretical constraints on multidimensional retiming design techniques

Image signal processing depends on computation intensive programs, which include the repetition of sequences of operations coded as nested loops. An effective technique in increasing the computing performance of such applications is the design and use of Application Specific Integrated Circuits using loop transformation techniques, and in particular, multi-dimensional (MD) retiming. The MD-retiming method improves the instruction-level parallelism of uniform loops. While many have written about the multi-dimensional retiming technique, no results have been published on the possible limitations of its application. This paper presents an analysis of that technique and its constraints when applied to nested loops with known index bounds, such as those found in two and three dimensional image processing.

[1]  Keshab K. Parhi,et al.  Two-dimensional retiming [VLSI design] , 1999, IEEE Trans. Very Large Scale Integr. Syst..

[2]  Edwin Hsing-Mean Sha,et al.  Rotation Scheduling: A Loop Pipelining Algorithm , 1993, 30th ACM/IEEE Design Automation Conference.

[3]  Edwin Hsing-Mean Sha,et al.  Scheduling of uniform multidimensional systems under resource constraints , 1998, IEEE Trans. Very Large Scale Integr. Syst..

[4]  Edwin Hsing-Mean Sha,et al.  Synthesis of multi-dimensional applications in VHDL , 1996, Proceedings International Conference on Computer Design. VLSI in Computers and Processors.

[5]  Edwin Hsing-Mean Sha,et al.  Loop Pipelining for Scheduling Multi-Dimensional Systems via Rotation , 1994, 31st Design Automation Conference.

[6]  Leslie Lamport,et al.  The parallel execution of DO loops , 1974, CACM.

[7]  Lindley,et al.  Practical Image Processing in C , 1994 .

[8]  Keshab K. Parhi,et al.  Generalized ILP scheduling and allocation for high-level DSP synthesis , 1993, Proceedings of IEEE Custom Integrated Circuits Conference - CICC '93.

[9]  Keshab K. Parhi,et al.  Static Rate-Optimal Scheduling of Iterative Data-Flow Programs via Optimum Unfolding , 1991, IEEE Trans. Computers.

[10]  Ángel Sánchez,et al.  Application of multidimensional retiming and matroid theory to DSP algorithm parallelization , 1999, Proceedings 25th EUROMICRO Conference. Informatics: Theory and Practice for the New Millennium.

[11]  N.L. Passos,et al.  Push-up scheduling: optimal polynomial-time resource constrained scheduling for multi-dimensional applications , 1995, Proceedings of IEEE International Conference on Computer Aided Design (ICCAD).

[12]  Yves Robert,et al.  Constructive Methods for Scheduling Uniform Loop Nests , 1994, IEEE Trans. Parallel Distributed Syst..

[13]  Jang-Ping Sheu,et al.  On the Parallelism of Nested For-Loops Using Index Shift Method , 1990, ICPP.

[14]  Edwin Hsing-Mean Sha,et al.  Achieving Full Parallelism Using Multidimensional Retiming , 1996, IEEE Trans. Parallel Distributed Syst..

[15]  Edwin Hsing-Mean Sha,et al.  Static scheduling of uniform nested loops , 1993, [1993] Proceedings Seventh International Parallel Processing Symposium.

[16]  Monica S. Lam,et al.  A Loop Transformation Theory and an Algorithm to Maximize Parallelism , 1991, IEEE Trans. Parallel Distributed Syst..

[17]  Edwin Hsing-Mean Sha,et al.  Schedule-based multi-dimensional retiming on data flow graphs , 1994, Proceedings of 8th International Parallel Processing Symposium.

[18]  Keshab K. Parhi,et al.  Two-dimensional retiming with low memory requirements , 1996, 1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings.