A New ILP Based Approach to Schedule and Bind Simultaneously

To construct complete design space for high-level synthesis and obtain globally optimized solution. An approach that fulfills operation scheduling and binding simultaneously is presented in this paper. The improved ASAP and ALAP scheduling algorithms are employed to compute the actually soonest and latest control steps with the condition of referring binding information. A basic model based on ILP is presented to solve the minimum resource occupance scheduling and binding under timing constraints. The means to solve minimum execution control steps scheduling and binding is also investigated. According to experiments, when scheduling fulfilled by this approach is accomplished, the binding results are obtained meanwhile. The executing time required by the approach is very low. And the feature of ILP has guaranteed the global optimization of results. Based on plenty of benchmark experiments, it can be concluded that the approach presented in this paper is effective and efficient.