Time constrained allocation and assignment techniques for high throughput signal processing

Abstract

A technique for the allocation of complex application specific datapaths is presented. The technique is especially suited for the synthesis of application specific architectures for high-throughput signal processing applications. Such applications comprise hierarchical compositions of nested loops and condition blocks. A minimum area set of datapaths is allocated and the available cycle budget is automatically distributed over the different blocks of the hierarchy in one global optimization process. Results for a number of real life examples are presented.<>