On this CD:
Guide
Glossary
Papers
Talks
Team Levo


Online:
(Levo central)

(Levo Visualizer)

Guide to Levo CD Version 1.0

All content on this CD Copyright 2002 by S. Langford, A.K. Uht, and the respective paper authors and/or the publishers.

Introduction to Levo

Levo is the University of Rhode Island and Northeastern University high ILP model machine realizing resource flow execution, DEE, data speculation, rampant speculation and automatic full predication. It does this primarily via a modification of the classic Tomasulo algorithm. Levo uses similar but more advanced Active Stations instead of Tomasulo’s reservation stations. The major modification, existing throughout Levo, is the use of time tags to identify the nominal execution order of the instructions in the Execution Window. These tags are used to implicitly enforce data, predicate and memory dependencies. Patents have been applied for. Please contact Prof. Gus Uht at URI if you are interested in licensing the technology.


The Levo Visualizer

The Levo Visualizer is a cycle by cycle graphhical visualization of the operation of the Levo microarchitecture. The simulation may be operated forward or backwards in time, with random access to any simulated cycle.

The latest version of the Levo Visualizer can be downloaded here. This is beta software. It is mainly working, but be warned - there are bugs.


Contents of this CD

This CD is an information resource for the Levo project. We include relevant published and unpublished papers, as well as links to online resources, especially the LevoVis interactive Levo simulation and visualization tool. Links to key personnel’s and labs’ home pages are provided. A brief Glossary is also included.

The major links are on the left of this screen.


Portable Document Format (PDF)

Much of the documentation on this CD is in Adobe 3.0 compatible PDF. Please click the Adobe logo below to get a copy of the free Adobe Acrobat Reader, if you don’t already have it.

Some of the documentation is in Postscript format; please go to the Ghostview site if you need a Postscript viewer.


Acknowledgements

The Levo project has been funded in part by the USA National Science Foundation through grants MIP-9708183 and EIA-9729839, the Intel Corporation and the URI Office of the Provost. Software donations have been provided by the Mentor Graphics and Xilinx Corporations.


February 6, 2002 | Gus Uht | uht@ele.uri.edu