Multicycle Scheduling under Local Timing Constraints using Genetic Algorithms and Tabu Search

Multicycle scheduling, i.e., the possibility to schedule operations over several clock cycles, gives faster designs. In this paper a method for clock period optimization during multicycle scheduling is described. The proposed method takes into account a set of local timing constraints and selects the appropriate functional units (with different area and delay) that can be used to implement the operations. The objective of the algorithm is to find the smallest design fulfilling the specified local timing constraints. The combinatorial optimization problem is solved with an evolution program, i.e., a genetic algorithm working on complex data structures, and a tabu search algorithm. 1. Introduction High-level synthesis (HLS) of digital systems consists of transforming a behavioral description of a design into a register transfer level description [4] [8]. The HLS process can be divided into the three main subtasks operation scheduling, resource allocation, and resource binding. These th..