ETH Zurich
  
ARCS'07: Architecture of Computing Systems
Swiss Federal Institute of Technology (ETH) Zurich, Switzerland
March 12-15, 2007
   ARCS 2007

Tutorials Monday March 12, 2007

The tutorials take place in room ETZ F76.1 in the ETZ building.
  • Haohuan Fu, Oskar Mencer and Wayne Luk, Imperial College London:
    Optimizing Hardware Function Evaluation
    Details
  • Mateusz Majer, Diana Göhringer, Josef Angermeier and Jürgen Teich, University of Erlangen-Nuremberg:
    The Erlangen Slot Machine: A run-time reconfigurable FPGA-based computer for partially reconfigurable applications
    Details

Schedule

Time Program
08:30-10:30 The Erlangen Slot Machine, part I
10:30-11:00 Coffee Break
11:00-13:00 The Erlangen Slot Machine, part II
13:00-14:00 Lunch
14:00-15:30 Optimizing Hardware Function Evaluation, part I
15:30-16:00 Coffee Break
16:00-17:30 Optimizing Hardware Function Evaluation, part II

Workshops Thursday March 15, 2007

The workshops take place in the ETZ building in rooms ETZ E6 (Dynamically reconfigurable systems) and ETZ E8 (Dependability and Fault Tolerance).
  • A. Koch, Technical University Darmstadt and Ch. Hochberger, Technical University Dresden:
    Dynamically Reconfigurable Systems
    Details
  • K.E. Grosspietsch, Fraunhofergesellschaft St. Augustin:
    Dependability and Fault Tolerance
    Details

Detailed Information

Tutorial: Optimizing Hardware Function Evaluation

Speakers: Haohuan Fu, Oskar Mencer, Wayne Luk
Location: room ETZ F76.1, ETZ building

Abstract

Many applications in multimedia, communications and finance involve mathematical functions, such as exponential and trigonometric functions. A good understanding of hardware function evaluation enables selection and development of architectures for specific requirements; such architectures typically make use of polynomials, tables, shift-and-add, and various other techniques. The large number of available methods leaves users with the task of deciding when to use which method.

In this tutorial, we first present a methodology and an automated system to decide which method to use, given range, precision, space, and time considerations. We show how to select the best function evaluation hardware for a given function, accuracy requirements, technology mapping and optimization metrics, such as area, throughput, and latency. Function evaluation f(x) typically consists of range reduction, and the actual evaluation on a small convenient interval such as [0, pi/2] for sin(x).

We outline the impact of hardware function evaluation with range reduction for a given range and precision of x and f(x) on area and speed. An automated bit-width optimization technique for minimizing the size of operators in hardware datapaths is also proposed. We illustrate design space exploration for various fixed-point functions such as sin(x) and log(x) accurate to one unit in the last place using MATLAB and ASC, A Stream Compiler for Field-Programmable Gate Arrays (FPGAs).

As part of the function evaluation optimization challenge, we extract bit-width optimization as an easily identifiable subproblem with a wide variety of possible solutions. Bit-width analysis can be done statically or dynamically. We show one method for static bit-width analysis based on affine arithmetic, and one method for dynamic bit-width analysis based on automatic differentiation of sets of expressions.

The application of the methods presented in this tutorial can be used to optimize hardware implementation at the architecture level, the arithmetic level, and the bit level.

Speaker Biographies

  • Oskar Mencer

    Oskar Mencer is an EPSRC Advanced Research Fellow and Senior Lecturer at the Department of Computing at Imperial College London. Prior to joining Imperial, Oskar was a Member of Technical Staff in the Computing Sciences Center at Bell Labs in Murray Hill. He got his PhD in Electrical Engineering in the Computer Architecture and Arithmetic Group at Stanford University, and a B.Sc degree from the Technion, Israel Institute of Technology. More information at: http://www.doc.ic.ac.uk/~oskar

  • Wayne Luk

    Wayne Luk, Professor of Computer Engineering at Imperial College London, obtained his doctorate from Oxford University. He received Best Paper Awards at International Conference on Field-Programmable Logic and Applications in 2004, and at IEEE International Conference on Field-Programmable Technology in 2005. His research interests include theory and practice of customizing hardware and software for specific application domains, such as multimedia, communications, and finance. More information at: http://www.doc.ic.ac.uk/~wl

Tutorial: The Erlangen Slot Machine: A run-time reconfigurable FPGA-based computer for partially reconfigurable applications

Speakers: Mateusz Majer, Diana Göhringer, Josef Angermeier and Jürgen Teich
Location: room ETZ F76.1, ETZ building

Abstract

In this tutorial, we introduce an FPGA-based reconfigurable platform called Erlangen Slot Machine (ESM) and tools supporting the development of partially reconfigurable modules and their communication. The main advantages of this platform in comparison to other platforms are on one hand the possibility for each module to access peripherals, such as video or audio in-and outputs, independent from its placement through a programmable crossbar, and on the other hand local SRAM banks for individual modules. With these advantages, the implementation of run-time reconfigurable partial modules is easier and an unrestricted relocation of modules on the device is possible.

The ESM consists of two boards called MotherBoard and BabyBoard. The actual MotherBoard is designed for Multimedia applications. All peripherals together with a PowerPC and an FPGA for the crossbar are implemented on this board. The BabyBoard is application independent and can therefore be mounted on different MotherBoards, depending on the needs of the user. The main FPGA, in which the reconfigurable applications run in so-called slots, is located on the BabyBoard. In another FPGA next to it, the Reconfiguration Manager is located, that reconfigures and relocates the modules on the main FPGA. For the reconfiguration, the user just has to access the PowerPC on the MotherBoard via the ESM-Shell, which then sends a command to the Reconfiguration Manager. The PowerPC also is used to reprogram the crossbar. A tool called SlotComposer helps the user to generate partial bitstreams, by first inserting communication structures required for partial reconfigurable designs, such as bus-macros, into VHDL files. Secondly, scripts are generated to start the synthesis and to call the EAPR-Flow tool offered by Xilinx. During this tutorial we will demonstrate the functionality and advantages of the ESM as a platform for partially reconfigurable applications.

The tutorial will include the following:

  • Introduction to platform and tools
  • Partially reconfigurable module development using SlotComposer
  • Reconfiguration Management
  • Development of a reconfigurable video-filter application

Speaker Biographies

  • Mateusz Majer received his diploma degree (Dipl.-Ing.) in Electrical Engineering from Darmstadt University of Technology, Germany, in October 2003. He is now in the final year of his PhD studies at the Hardware/Software Co-Design Chair of the University of Erlangen-Nuremberg.

  • Diana Göhringer received her diploma degree (Dipl.-Ing.) in Electrical Engineering and Information Technology from the University of Karlsruhe(TH), Germany, in January 2006. She is now working as a PhD student at the Hardware/Software Co-Design Chair of the University of Erlangen-Nuremberg.

  • Josef Angermeier received his diploma degree (Dipl.-Inf.) in Computer Science from University of Erlangen-Nuremberg, Germany, in December 2005. He is now working as a PhD student at the Hardware/Software Co-Design Chair of the University of Erlangen-Nuremberg.

  • Jürgen Teich received his masters degree (Dipl.-Ing.) in 1989 from the University of Kaiserslautern (with honors). From 1989 to 1993, he was PhD student at the University of Saarland, Saarbrücken, Germany from where he received his PhD degree (summa cum laude). His PhD thesis entitled "A Compiler for Application-Specific Processor Arrays" summarizes his work on extending techniques for mapping computation intensive algorithms onto dedicated VLSI processor arrays. In 1994, Dr. Teich joined the DSP design group of Prof. E. A. Lee and D.G. Messerschmitt in the Department of Electrical Engineering and Computer Sciences (EECS) at UC Berkeley where he was working in the Ptolemy project (PostDoc). From 1995-1998, he held a position at Institute of Computer Engineering and Communications Networks Laboratory (TIK) at ETH Zürich, Switzlerland, finishing his habilitation entitled "Synthesis and Optimization of Digital Hardware/ Software Systems" in 1996. From 1998-2002, he was full professor in the Electrical Engineering and Information Technology department of the University of Paderborn, Germany, holding a chair in Computer Engineering. In Paderborn he also worked in two Collaborative Research Centers sponsored by the German Science Foundation (DFG). Since 2003, he is appointed full professor in the Computer Science Institute of the University Erlangen-Nuremberg, holding the new chair in Hardware-Software-Co-Design.

    Mr. Teich has been a member of multiple program committees of well-known conferences such as the DATE (Design, Automation, and Test in Europe) as well as editor of several books. Prof. Teich coordinates the German priority program 1148 (DFG) on reconfigurable computing. Since 2004, Prof. Teich is also elected reviewer of the German Research Foundation (DFG) for the area of Computer Architectures and Embedded Systems. His special interests are massive parallelism, embedded systems, hardware/software codesign, and computer architecture.

  • Workshop on "Dynamically Reconfigurable Systems (DRS)"

    Chairs: A. Koch, Technical University Darmstadt and Ch. Hochberger, Technical University Dresden
    Location: room ETZ E6, ETZ building

    Schedule

    Time Program
    9:00-10:00 Keynote 1: Dr. Juanjo Noguera, Xilinx Research Labs Europe: Application-driven research in partial reconfiguration abstract
    10:00-10:30 Coffee Break
    Architectures and Paradigms
    10:30-11:00 Yann Thoma, Andres Upegui, Andres Perez-Uribe, and Eduardo Sanchez (ReDS, HES-SO/HEIG-VD, Yverdon): Self-Replication Mechanism by Means of Self-Reconfiguration
    11:00-11:30 Zied Marrakchi, Hayder Mrabet, Christian Masson, Habib Mehrez (Univesité Pierre et Marie Curie, Paris): Unifying Mesh and MFPGA architectures to improve performances
    11:30-12:00 Peter Zipf, Lei Liu, Zdravko Bozakov, Manfred Glesner (Darmstadt University of Technology): Design Optimisations for a Cellular Automata Model with Programmable Interconnect Structure
    Middleware
    12:00-12:30 Alessandro Cilardo, Luigi Coppolino, Nicola Mazzocca (University of Naples Federico II): A framework for the Design of Distributed Reconfigurable Embedded Systems
    12:30-13:00 Axel Schneider, Joachim Knäblein, Bernd Müller, Marcel Putsche (Alcatel-Lucent Germany), Sebastian Goller, Uwe Pross, Ulrich Heinkel (University of Technology Chemnitz): Ethernet based in-service reconfiguration of SoCs in telecommunication networks
    13:00-14:00 Lunch Break
    14:00-15:00 Keynote 2: Dr. Mladen Berekovic, IMEC: Reconfigurable Architectures in Embedded System-on-Chips
    15:00-15:30 Coffee Break
    Tools
    15:30-16:00 Sebastian Lange, Martin Middendorf (University of Leipzig): Online Strategies for the Reconfiguration of Two-Level Reconfigurable Architectures
    16:00-16:30 Markus Rullmann, Sebastian Siegel, Renate Merker (University of Technology Dresden), Julio A. Oliveira Filho, Thomas Schweizer, Tobias Oppold, Wolfgang Rosenstiel (University of Tübingen): Efficient Mapping and Functional Verification of Parallel Algorithms on a Multi-Context Reconfigurable Architecture
    16:30-17:00 Sándor P. Fekete, Jan C. van der Veen (Braunschweig University of Technology), Josef Angermeier, Diana Göhringer, Mateusz Majer, Jürgen Teich (University of Erlangen-Nuremberg): Scheduling and Communication-Aware Mapping of HW/SW Modules for Dynamically and Partially Reconfigurable SoC Architectures

    Keynotes

    • Dr. Juanjo Noguera, Xilinx Research Labs Europe: Application-driven research in partial reconfiguration

      Partial Reconfiguration (PR) is a unique feature of Xilinx FPGA's that allows the reconfiguration of a part of the device while the rest of the FPGA continues operating. This approach, where the application functionality is time-multiplexed on the FPGA has been widely addressed by many researchers in academia, proposing multiple architectures and techniques for hardware virtualization but without a clear focus on the application perspective. The main goal of this talk is to propose a change in the current approach. Thus, we believe that applications or application domains, and not architectures, should drive the research in partial reconfiguration. Using this application-focussed view of partial reconfiguration, we introduce the need for high-level domain-specific design tools that enable the users to specify the behaviour of the application and how the system should react to changes in its environment. Several examples using the networking domain will be provided.

    Workshop on "Dependability and Fault Tolerance"

    Chair: K.E. Grosspietsch, Fraunhofergesellschaft St. Augustin
    Location: room ETZ E8, ETZ building

    Schedule

    Time Program
    09:00-09:10 Welcome
    09:10-10:00 Invited Talk: K. Echtle (University of Duisburg-Essen): Fault-Tolerant Communication in Safety-Relevant Automotive Applications
    10:00-10:30 Coffee Break
    Hardware Fault Tolerance
    10:30-11:00 R. Kothe, H.T. Vierhaus (University of Technology Cottbus): Repair Functions and Redundancy Management for Bus Structures
    11:00-11:30 B. Fechner, J. Keller (Fernuniveristät Hagen): Compression-free Checksum-based Fault-Detection Schemes for Pipelined Processors
    Dependable Mechatronic Systems
    11:30-12:00 F. Mösch, M. Litza, A. El Sayed Auf, B. Jakimovski, E. Maehle (University of Lübeck), W. Brockmann (University of Osnabrück): Organic Fault-Tolerant Controller for the Walking Robot OSCAR
    12:00-12:30 R. Agarwal (Envirotech Instruments Pvt. Ltd, New Delhi), K.-E. Grosspietsch (Fraunhofer-Gesellschaft, St. Augustin): Fault Tolerance for Autonomous Robots by Means of Adaptive Filters
    12:30-13:00 D. Dötz, H. Hoffmann, D. Scherer, C. Trinitis, M. Walter (Technische Universität München): On the Automation of Incremental Metal Forming Processes
    13:00-14:00 Lunch Break
    Modeling and Simulation of Dependable Systems
    14:00-14:30 P. Limbourg, H.-D. Kochs, K. Echtle, I. Eusgeld (University of Duisburg-Essen) Reliability Prediction in Systems with Correlated Component Failures — An Approach Using Copulas
    14:30-15:00 M. Walter, C. Trinitis (Technische Universität München): Automatic Generation of State-Based Dependability Models: from Availability to Safety
    14:30-15:00 P. Gawkowski, J. Sosnowski (Warsaw University of Technology): Experiences with Software Implemented Fault Injection
    15:00-15:30 F. Belli, C.J. Budnik (Universität Paderborn): Elementare Mutations-Operatoren zur Generierung von Testfällen anhand von Statecharts
    16:00-16:30 Coffee Break
    Modeling and Simulation of Dependable Systems
    16:30-17:00 F. Saglietti, N. Oster, F. Pinte (University Erlangen-Nuremberg): Interface Coverage Criteria Supporting Model-Based Integration Testing
    17:00-17:30 F. Belli, M. Linschulte (Universitüt Paderborn), I. Schieferdecker (Technische Universität Berlin): Ereignisorientiertes Testen Web-basierter Systeme — Verfeinerung des holistischen Ansatzes und eine Fallstudie