2002•International Conference on Acoustics, Speech, and Signal ProcessingRequires access

Code generation for the AT&T DSP32

K. Baudendistel, James H. McClellan

Abstract

The use of compiler generation tool sets (CGTSs) targeting programmable digital signal processing (DSP) microprocessors (PDSPs) is examined. Such tool sets allow separate construction of all compiler components: language-dependent front ends, internal optimization phases, and machine-dependent back ends for target code generation, including post-generation machine code optimization. In order to compare different CGTSs, standard C language front ends were used and machine-dependent back ends constructed for the AT&T DSP32 digital signal processor. The quality and efficiency of the machine code produced by two CGTS-based compilers are compared with that generated by the AT&T C compiler.>

About this research paper

What this paper is about

Why it matters

OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Key concepts: Compiler, Computer science, Code generation, Digital signal processing, Machine code, Programming language, Code (set theory), Dead code elimination