Constraint driven code selection for fixed-point DSPs

Abstract

Fixed-point DSPs are a class of embedded pro- cessors with highly irregular architectures. This irregu- larity makes it difficult to generate high-quality machine code from programming languages such as C. In this pa- per we present a novel constraint driven approach to code selection for irregular processor architectures, which pro- vides a twofold improvement of earlier work. First, it handles complete data flow graphs instead of trees and thereby generates better code in presence of common subexpressions. Second, the presented technique is not restricted to computation of a single solution, but it gen- erates alternative solutions. This feature enables the tight coupling of different code generation phases, resulting in better exploitation of instruction-level parallelism. Ex- perimental results indicate that our technique is capable of generating machine code that competes well with hand- written assembly code.