Dynamic Memory Allocation within a Behavioural Synthesis System
Dynamic Memory Allocation within a Behavioural Synthesis System
MOODS (Multiple Objective Optimisation in Data and control path Synthesis) is a behavioural synthesis system tool that converts behavioural descriptions of users digital designs into synchronous structural representations. This thesis describes an enhancement to the original MOODS system that allows direct conversion of dynamic memory constructs within the source language into a fully structural design with run-time memory representation.
VHDL is used as the source language for design descriptions, and is capable of directly describing dynamic memory structures in the form of explicitly created structures as well as the more implicit dynamic memory requirements of procedural recursion. The VHDL compiler required extensive modification to handle the increased subset of the language at the behavioural level.
The conversion of explicit structure allocation requires a run-time system that is capable of storing the data represented by the dynamically allocated structures. This system is realised by a behavioural description of a heap management algorithm that is both space and speed efficient and interfaces with the users design directly via an automatically generated interface.
Procedural recursion is now synthesisable by MOODS from the inclusion of a dynamically modified call stack created again, automatically within the users design, which contains the storage for local variables and passed parameters declared within the subprograms.
Finally, two demonstration systems have been designed and synthesised with the enhanced system, with both designs displaying the use of dynamic memory allocation and the second design showing the use of procedural recursion within a synthesised hardware system.
University of Southampton
Milton, Daniel James David
0061f704-d8dd-4bd8-a0fb-63ead80d09d6
2002
Milton, Daniel James David
0061f704-d8dd-4bd8-a0fb-63ead80d09d6
Milton, Daniel James David
(2002)
Dynamic Memory Allocation within a Behavioural Synthesis System.
University of Southampton, Doctoral Thesis.
Record type:
Thesis
(Doctoral)
Abstract
MOODS (Multiple Objective Optimisation in Data and control path Synthesis) is a behavioural synthesis system tool that converts behavioural descriptions of users digital designs into synchronous structural representations. This thesis describes an enhancement to the original MOODS system that allows direct conversion of dynamic memory constructs within the source language into a fully structural design with run-time memory representation.
VHDL is used as the source language for design descriptions, and is capable of directly describing dynamic memory structures in the form of explicitly created structures as well as the more implicit dynamic memory requirements of procedural recursion. The VHDL compiler required extensive modification to handle the increased subset of the language at the behavioural level.
The conversion of explicit structure allocation requires a run-time system that is capable of storing the data represented by the dynamically allocated structures. This system is realised by a behavioural description of a heap management algorithm that is both space and speed efficient and interfaces with the users design directly via an automatically generated interface.
Procedural recursion is now synthesisable by MOODS from the inclusion of a dynamically modified call stack created again, automatically within the users design, which contains the storage for local variables and passed parameters declared within the subprograms.
Finally, two demonstration systems have been designed and synthesised with the enhanced system, with both designs displaying the use of dynamic memory allocation and the second design showing the use of procedural recursion within a synthesised hardware system.
Text
835859.pdf
- Version of Record
More information
Published date: 2002
Identifiers
Local EPrints ID: 464559
URI: http://eprints.soton.ac.uk/id/eprint/464559
PURE UUID: a50537c4-113a-45fd-9401-7d2b7a39c6d6
Catalogue record
Date deposited: 04 Jul 2022 23:46
Last modified: 16 Mar 2024 19:36
Export record
Contributors
Author:
Daniel James David Milton
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics