Effective and efficient FPGA synthesis through functional decomposition based on information relationship measures

Artificial Intelligence in Logic Design

“…2001;J6zwiak and Chojnacki 2001). The experimental results and other arguments presented above and in (J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Vo1f 1995;J6zwiak 1995;J6zwiak and Chojnacki 2001) demonstrate that the double-beam search scheme is suitable for solving the circuit synthesis problems and its application results in very effective and efficient synthesis algorithms.…”

“…When tested on benchmarks, it efficiently produced very good results in all considered applications (J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Vol 1995;J6zwiak and Chojnacki 2001).…”

“…We applied with success the double-beam search scheme in numerous circuit synthesis methods and tools solving different kinds of synthesis problems (J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Volf 1995;J6zwiak 1995;J6zwiak et aI. 2001;J6zwiak and Chojnacki 2001). The experimental results and other arguments presented above and in (J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Vo1f 1995;J6zwiak 1995;J6zwiak and Chojnacki 2001) demonstrate that the double-beam search scheme is suitable for solving the circuit synthesis problems and its application results in very effective and efficient synthesis algorithms.…”

“…In circuit synthesis, it enables satisfaction of the interconnection constraints, input constraints of building blocks, and minimization of both the circuit active and interconnection area (J6zwiak and Konieczny 1996). Although our research of the minimal input support problem (MISP) is primarily motivated by applications in digital circuit synthesis (Jasinski et al 1989;J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Volf 1995;J6zwiak 1995;J6zwiak and Ong 1996;J6zwiak and Konieczny 1996;J6zwiak et al 1998;J6zwiak and Chojnacki 2001), solutions to this problem can also be applied in many other fields, such as decision systems, database design, pattern recognition, machine learning and data mining (for minimizing the number of attributes in relation tables) (Lin and Cercone 1997;Luba and Rybnik 1992;Luba and Rybnik 1992). QuickScan is a simple, but quite effective and extremely efficient problem-specific search algorithm for input support minimisation of discrete relations, functions and sequential machines.…”

“…In a number of constructive double-beam algorithms that we developed and tested (J6zwiak and Kolsteren 1991;J6zwiak 1992a;J6zwiak and Volf 1992;J6zwiak 1992b;J6zwiak and Volf 1995;J6zwiak and Chojnacki 2001), we implemented certain heuristic elaborations of the following general decision rule for selection of the construction operators and partial solutions: "at each stage of the search take a decision that has the highest chance (or certainty) of leading to the optimal solution according to the estimations given by the evaluation criteria; if there are more decisions of comparable certainty, try a number of them in paralle~ preferring the decisions that bring more information for deciding on the successive issues".…”