A Behavioural VHDL Synthesis System using Data Path Optimisation


Williams, A.C. (1997) A Behavioural VHDL Synthesis System using Data Path Optimisation. : Southampton University, Doctoral Thesis .

Download

Full text not available from this repository.

Description/Abstract

MOODS (Multiple Objective Optimisation in Data and control path synthesis) is a synthesis system which provides the ability to automatically optimise a design from a behavioural to a structural VHDL description. This thesis details two sets of enhancements made to the original system to improve the overall quality of the final hardware implementations obtained, and expand the range of the accepted VHDL subset. Whereas the original MOODS considered each functional unit in the target module library to be a purely combinational logic block, the �expanded modules� developed for this project provide a means of implementing sequential multi-cycle modules. These modules are defined as technology-independent templates, which are inline expanded into the internal design structure during synthesis. This enables inter-module optimisation to occur at the sub-module level, thus affording greater opportunities for unit sharing and module binding. The templates also facilitate the development of specialised interface modules. These enable the use of fixed timing I/O protocols for external interfacing, while maintaining maximum scheduling flexibility within the body of the behaviour. The second set of enhancements includes an improved implementation of behavioural VHDL as input to the system. This expands the previously limited subset to include such elements as signals, wait statements, concurrent processes, and functions and procedures. These are implemented according to the IEEE standard thereby preserving the computational effects of the VHDL simulation model. The final section of work involves the development and construction of an FPGA-based real-time audio-band spectrum analyser, synthesised within the MOODS environment. This design process provides valuable insights into the strengths and weaknesses of both MOODS and behavioural synthesis in general, serving as a firm foundation to guide future development of the system.

Item Type: Thesis (Doctoral)
Divisions: Faculty of Physical Sciences and Engineering > Electronics and Computer Science
ePrint ID: 251233
Date Deposited: 22 Oct 1999
Last Modified: 27 Mar 2014 19:52
Further Information:Google Scholar
URI: http://eprints.soton.ac.uk/id/eprint/251233

Actions (login required)

View Item View Item