Full download text book at textbookfull.

com

Architecture of Computing Systems ARCS

2020 33rd International Conference
Aachen Germany May 25 28 2020
Proceedings André Brinkmann
DOWLOAD HERE

https://textbookfull.com/product/architecture-of-
computing-systems-arcs-2020-33rd-international-
conference-aachen-germany-
may-25-28-2020-proceedings-andre-brinkmann/

DOWLOAD NOW

Download more textbook from textbookfull.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Architecture of Computing Systems ARCS 2014 27th

International Conference Lübeck Germany February 25 28
2014 Proceedings 1st Edition Erik Maehle

https://textbookfull.com/product/architecture-of-computing-
systems-arcs-2014-27th-international-conference-lubeck-germany-
february-25-28-2014-proceedings-1st-edition-erik-maehle/

Architecture of Computing Systems ARCS 2016 29th

International Conference Nuremberg Germany April 4 7
2016 Proceedings 1st Edition Frank Hannig

https://textbookfull.com/product/architecture-of-computing-
systems-arcs-2016-29th-international-conference-nuremberg-
germany-april-4-7-2016-proceedings-1st-edition-frank-hannig/

High Performance Computing: 35th International

Conference, ISC High Performance 2020, Frankfurt/Main,
Germany, June 22–25, 2020, Proceedings Ponnuswamy
Sadayappan
https://textbookfull.com/product/high-performance-computing-35th-
international-conference-isc-high-performance-2020-frankfurt-
main-germany-june-22-25-2020-proceedings-ponnuswamy-sadayappan/

Computational Methods in Systems Biology 18th

International Conference CMSB 2020 Konstanz Germany
September 23 25 2020 Proceedings Alessandro Abate

https://textbookfull.com/product/computational-methods-in-
systems-biology-18th-international-conference-cmsb-2020-konstanz-
germany-september-23-25-2020-proceedings-alessandro-abate/
Cooperative Design, Visualization, and Engineering:
17th International Conference, CDVE 2020, Bangkok,
Thailand, October 25–28, 2020, Proceedings Yuhua Luo

https://textbookfull.com/product/cooperative-design-
visualization-and-engineering-17th-international-conference-
cdve-2020-bangkok-thailand-october-25-28-2020-proceedings-yuhua-
luo/

Proceedings of International Conference on Frontiers in

Computing and Systems COMSYS 2020 Debotosh
Bhattacharjee

https://textbookfull.com/product/proceedings-of-international-
conference-on-frontiers-in-computing-and-systems-
comsys-2020-debotosh-bhattacharjee/

High Performance Computing: ISC High Performance 2020

International Workshops, Frankfurt, Germany, June
21–25, 2020, Revised Selected Papers Heike Jagode

https://textbookfull.com/product/high-performance-computing-isc-
high-performance-2020-international-workshops-frankfurt-germany-
june-21-25-2020-revised-selected-papers-heike-jagode/

Euro Par 2020 Parallel Processing 26th International

Conference on Parallel and Distributed Computing Warsaw
Poland August 24 28 2020 Proceedings Maciej Malawski

https://textbookfull.com/product/euro-par-2020-parallel-
processing-26th-international-conference-on-parallel-and-
distributed-computing-warsaw-poland-
august-24-28-2020-proceedings-maciej-malawski/

Foundations of Intelligent Systems 25th International

Symposium ISMIS 2020 Graz Austria September 23 25 2020
Proceedings Denis Helic

https://textbookfull.com/product/foundations-of-intelligent-
systems-25th-international-symposium-ismis-2020-graz-austria-
september-23-25-2020-proceedings-denis-helic/
 André Brinkmann · Wolfgang Karl ·
Stefan Lankes · Sven Tomforde ·
Thilo Pionteck · Carsten Trinitis (Eds.)

Architecture of
LNCS 12155

Computing Systems –
ARCS 2020
33rd International Conference
Aachen, Germany, May 25–28, 2020
Proceedings
Lecture Notes in Computer Science 12155

Founding Editors
Gerhard Goos
Karlsruhe Institute of Technology, Karlsruhe, Germany
Juris Hartmanis
Cornell University, Ithaca, NY, USA

Editorial Board Members

Elisa Bertino
Purdue University, West Lafayette, IN, USA
Wen Gao
Peking University, Beijing, China
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Gerhard Woeginger
RWTH Aachen, Aachen, Germany
Moti Yung
Columbia University, New York, NY, USA
More information about this series at http://www.springer.com/series/7407
André Brinkmann Wolfgang Karl
• •

Stefan Lankes Sven Tomforde

• •

Thilo Pionteck Carsten Trinitis (Eds.)

•

Architecture of
Computing Systems –
ARCS 2020
33rd International Conference
Aachen, Germany, May 25–28, 2020
Proceedings

123
Editors
André Brinkmann Wolfgang Karl
Johannes Gutenberg University of Mainz Karlsruhe Institute of Technology
Mainz, Germany Karlsruhe, Germany
Stefan Lankes Sven Tomforde
RWTH Aachen University Christian-Albrecht University of Kiel
Aachen, Germany Kiel, Germany
Thilo Pionteck Carsten Trinitis
Otto-von-Guericke University Magdeburg Technical University of Munich
Magdeburg, Germany Garching b. München, Germany

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-030-52793-8 ISBN 978-3-030-52794-5 (eBook)
https://doi.org/10.1007/978-3-030-52794-5
LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

© Springer Nature Switzerland AG 2020

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, expressed or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
 Preface

The 33rd edition of the International Conference on Computer Architecture (ARCS

2020) was planned to be held at RWTH Aachen University, one of the oldest technical
universities in Germany and a member of the German TU9-association of technical
universities. The conference would have taken place during May 25–28, 2020, and
therefore in the middle of the 150th anniversary celebrations of RWTH Aachen.
The coronavirus disease COVID-19, nevertheless, completely changed the way of
living in Germany, Europe, and all over the world in 2020. The physical execution of
nearly all scientific conferences had to be stopped, as borders were closed and all
bigger events had to be canceled to confine the spread of the disease and therefore to
save lives. This of course also hit ARCS 2020 and the Organizing Committee made the
decision to cancel the networking event for our community in the exciting city of
Aachen. Due to the extraordinary situation only these conference proceedings were
published. In addition, the organization team offered the possibility to publish videos
of the contributions in order to foster the dissemination of the results.
The ARCS conference series has over 30 years of tradition reporting leading-edge
research in computer architecture and operating systems. ARCS addresses the complete
spectrum from fully integrated, self-powered embedded systems up to
high-performance computing systems and provides a platform covering newly
emerging and cross-cutting topics, such as autonomous and ubiquitous systems,
reconfigurable computing and acceleration, neural networks and artificial intelligence,
as well as outlooks on future topics such as post-Moore architectures and organic
computing.
The focus of the 2020 conference was set on concepts and tools for incorporating
self-adaptation and self-organization mechanisms in high-performance computing
systems. This includes upcoming approaches for runtime modifications at various
abstraction levels, ranging from hardware changes to goal changes and their impact on
architectures, technologies, and languages. The conference was organized by the
special interest group on “Architecture of Computing Systems” of the GI (Gesellschaft
für Informatik e. V.) and ITG (Informationstechnische Gesellschaft im VDE).
ARCS 2020 attracted 33 submissions from authors in 15 countries worldwide,
including China, Colombia, South Korea, Venezuela, and the USA. Each submission
was reviewed by a diverse and dedicated Program Committee. 3 submissions received
3 qualified reviews and the remaining 30 submissions got 4 or even 5 reviews, leading
to a total of 116 reviews.
The Program Committee selected 12 submissions to be published in the proceed-
ings, which corresponds to a 37% paper acceptance rate. The accepted papers cover a
variety of topics from the ARCS core domains, including scheduling, near-data pro-
cessing, HW architectures, or the application of transactional memory.
vi Preface

ARCS has a long tradition of hosting associated workshops. The following three
workshops were organized in conjunction with the main conference this year and we
decided to include the accepted papers within the conference proceedings:
– 14th GI/ITG Workshop on Parallel Systems and Algorithms (PASA 2020)
– 6th FORMUS3IC Workshop
– Third Workshop on Computer Architectures in Space (CompSpace 2020)
We thank the many individuals who contributed to ARCS 2020, in particular the
members of the Program Committee and all the additional external reviewers for their
time and effort in carefully reviewing and judging the submissions. We further thank all
authors for submitting their work to ARCS and presenting accepted papers as video
contributions. The workshops were organized and coordinated by Carsten Trinitis, the
proceedings were compiled by Thilo Pionteck, publicity was managed by Lena Oden,
and the website was maintained by Markus Hoffmann. Thanks to all these individuals
and all the many other people who helped in the organization of ARCS 2020.

May 2020 André Brinkmann

Wolfgang Karl
Stefan Lankes
Sven Tomforde
 Organization

General Chairs
Stefan Lankes RWTH Aachen University, Germany
Wolfgang Karl Karlsruhe Institute of Technology, Germany

Program Chairs
Sven Tomforde University of Kiel, Germany
André Brinkmann Johannes Gutenberg University Mainz, Germany

Workshop and Tutorial Chair

Carsten Trinitis Technical University of Munich, Germany

Publicity Chair
Lena Oden FernUniversität in Hagen, Germany

Publication Chair
Thilo Pionteck Otto von Guericke University Magdeburg, Germany

Web Chair
Markus Hoffmann Karlsruhe Institute of Technology, Germany

Program Committee
Mladen Berekovic Universität zu Lübeck, Germany
Jürgen Brehm Leibnitz Universität Hannover, Germany
André Brinkmann Johannes Gutenberg University Mainz, Germany
Uwe Brinkschulte Goethe-Universität Frankfurt am Main, Germany
João M.P. Cardoso Universidade do Porto, Portugal
Thomas Carle Institut de Recherche en Informatique de Toulouse,
France
Ahmed El-Mahdy Egypt-Japan University of Science and Technology,
Egypt
Lukas Esterle Aarhus University, Denmark
Dietmar Fey Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Giorgis Georgakoudis Lawrence Livermore National Laboratory, USA
viii Organization

Roberto Giorgi University of Siena, Italy

Daniel Gracia-Pérez Thales Research and Technology, France
Christian Gruhl University of Kassel, Germany
Jan Haase Universität zu Lübeck, Germany
Jörg Hähner University of Augsburg, Germany
Heiko Hamann Universität zu Lübeck, Germany
Andreas Herkersdorf Technical University of Munich, Germany
Christian Hochberger Technische Universität Darmstadt, Germany
Gert Jervan Tallinn University of Technology, Estonia
Wolfgang Karl Karlsruhe Institute of Technology, Germany
Jörg Keller FernUniversität in Hagen, Germany
Dirk Koch The University of Manchester, UK
Hana Kubátová FIT CTU, Czech Republic
Stefan Lankes RWTH Aachen University, Germany
Erik Maehle Universität zu Lübeck, Germany
Lena Oden FernUniversität in Hagen, Germany
Alex Orailoglu University of California, San Diego, USA
Thilo Pionteck Otto von Guericke University Magdeburg, Germany
Mario Porrmann Osnabrück University, Germany
Jürgen Mottok Ostbayerische Technische Hochschule Regensburg,
Germany
Reza Salkhordeh Johannes Gutenberg University Mainz, Germany
Toshinori Sato Fukuoka University, Japan
Wolfgang Friedrich-Alexander-Universität Erlangen-Nürnberg,
Schröder-Preikschat Germany
Martin Schulz Technical University of Munich, Germany
Leonel Sousa Universidade de Lisboa, Portugal
Olaf Spinczyk Osnabrück University, Germany
Benno Stabernack Fraunhofer Institute for Telecommunications,
Heinrich Hertz Institute, Germany
Walter Stechele Technical University of Munich, Germany
Anthony Stein University of Augsburg, Germany
Tim Süss University of Applied Science Fulda, Germany
Djamshid Tavangarian Universität Rostock, Germany
Jürgen Teich Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Sven Tomforde University of Kassel, Germany
Carsten Trinitis Technical University of Munich, Germany
Theo Ungerer University of Augsburg, Germany
Hans Vandierendonck Queen’s University Belfast, UK
Daniel Versick NORDAKADEMIE - Hochschule der Wirtschaft,
Germany
Stephane Vialle CentraleSupelec and UMI GT-CNRS 2958, France
Klaus Waldschmidt Goethe-Universität Frankfurt am Main, Germany
Dominic Wist BIOTRONIC Berlin, Germany
Stephan Wong Delft University of Technology, The Netherlands
 Organization ix

Additional Reviewers
Christoph Borchert Osnabrück University, Germany
Peter Brand Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Hugues Cassé Institut de Recherche en Informatique de Toulouse,
France
Boris Dreyer Technische Universität Darmstadt, Germany
Jorge Echavarria Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Farnam Khalili Maybodi University of Siena, Italy
Kevin Kremer Johannes Gutenberg University Mainz, Germany
Leon Li University of California, San Diego, USA
Diogo Marques Universidade de Lisboa, Portugal
Rolf Meyer Universität zu Lübeck, Germany
Nuno Neves INESC-ID Lisboa, Portugal
Elbruz Ozen University of California, San Diego, USA
Marco Procaccini University of Siena, Italy
Oskar Pusz Leibnitz Universität Hannover, Germany
Paul Renaud-Goud Institut de Recherche en Informatique de Toulouse,
France
Amin Sahebi University of Siena, Italy
Frederic Schimmelpfennig Johannes Gutenberg University Mainz, Germany
Ladislav Steffko University of California, San Diego, USA
Ramon Wirsch Technische Universität Darmstadt, Germany
Mahdi Zahedi Delft University of Technology, The Netherlands

6th FORMUS3IC Workshop

Program Committee
Dietmar Fey Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Jens Harnisch Infineon AG, Germany
Martin Hobelsberger Hochschule München, Germany
Jose Daniel Gracia Sanchez Universidad Carlos III de Madrid, Spain
Ralph Mader Continental AG, Germany
Vaclav Matousek University of West Bohemia, Czech Republic
Avi Mendelson Technion University, Israel
Jürgen Mottok Ostbayerische Technische Hochschule Regensburg,
Germany
Tobias Schüle SIEMENS AG, Germany
Andreas Sailer Vector, Germany
Christian Siemers Clausthal University of Technology, Germany
Michael Wong Codeplay, UK
x Organization

Third Workshop on Computer Architectures in Space (CompSpace)

Program Committee
Carsten Trinitis Technical University of Munich, Germany
Sebastian Rückerl Technical University of Munich, Germany
Nicolas Appel Technical University of Munich, Germany

14th Workshop on Parallel Systems and Algorithms (PASA)

Program Committee
Martin Dietzfelbinger Technische Universität Ilmenau, Germany
Steffen Christgau Zuse Institute Berlin, Germany
Andreas Döring IBM Zürich, Switzerland
Norbert Eicker Jülich Supercomputing Centre, Germany
Thomas Fahringer Universität Innsbruck, Austria
Dietmar Fey Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Rolf Hoffmann Technische Universität Darmstadt, Germany
Klaus Jansen Christian-Albrechts-Universität zu Kiel, Germany
Ben Juurlink Technische Universität Berlin, Germany
Wolfgang Karl Karlsruhe Institute of Technology, Germany
Jörg Keller Fernuniversität in Hagen, Germany
Stefan Lankes RWTH Aachen University, Germany
Christian Lengauer University of Passau, Germany
Erik Maehle Universität zu Lübeck, Germany
Ulrich Margull Technische Hochschule Ingolstadt, Germany
Ernst W. Mayr Technical University of Munich, Germany
Ulrich Meyer Goethe-Universität Frankfurt am Main, Germany
Friedhelm Meyer Universität Paderborn, Germany
auf der Heide
Juergen Mottok Ostbayerische Technische Hochschule Regensburg,
Germany
Wolfgang Nagel Technische Universität Dresden, Germany
Michael Philippsen Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
Harald Räcke Technical University of Munich, Germany
Karl Dieter Reinartz Hochstädt, Germany
Christian Scheideler Universität Paderborn, Germany
Bettina Schnor Universität Potsdam, Germany
Martin Schulz Technical University of Munich, Germany
Uwe Schwiegelshohn Technische Universität Dortmund, Germany
Peter Sobe Hochschule fr Technik und Wirtschaft Dresden,
Germany
Carsten Trinitis Technical University of Munich, Germany
Rolf Wanka Friedrich-Alexander-Universität Erlangen-Nürnberg,
Germany
 Contents

Main Concerence

Approximate Data Dependence Profiling Based on Abstract Interval

and Congruent Domains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Mostafa Abbas, Rasha Omar, Ahmed El-Mahdy, and Erven Rohou

Evaluating Dynamic Task Scheduling with Priorities and Adaptive Aging

in a Task-Based Runtime System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Thomas Becker and Tobias Schüle

An Architecture for Solving the Eigenvalue Problem

on Embedded FPGAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Alwyn Burger, Patrick Urban, Jayson Boubin, and Gregor Schiele

ECC Memory for Fault Tolerant RISC-V Processors . . . . . . . . . . . . . . . . . . 44

Alexander Dörflinger, Yejun Guan, Sören Michalik, Sönke Michalik,
Jamin Naghmouchi, and Harald Michalik

3D Optimisation of Software Application Mappings

on Heterogeneous MPSoCs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Gereon Führ, Ahmed Hallawa, Rainer Leupers, Gerd Ascheid,
and Juan Fernando Eusse

Towards a Priority-Based Task Distribution Strategy for an Artificial

Hormone System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Eric Hutter and Uwe Brinkschulte

He..ro DB: A Concept for Parallel Data Processing

on Heterogeneous Hardware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Michael Müller, Thomas Leich, Thilo Pionteck, Gunter Saake,
Jens Teubner, and Olaf Spinczyk

Investigating Transactional Memory for High Performance

Embedded Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Christian Piatka, Rico Amslinger, Florian Haas, Sebastian Weis,
Sebastian Altmeyer, and Theo Ungerer

X-CEL: A Method to Estimate Near-Memory Acceleration Potential

in Tile-Based MPSoCs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Sven Rheindt, Andreas Fried, Oliver Lenke, Lars Nolte, Temur Sabirov,
Tim Twardzik, Thomas Wild, and Andreas Herkersdorf
xii Contents

Engineering an Optimized Instruction Set Architecture

for AMIDAR Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Alexander Schwarz and Christian Hochberger

Scaling Logic Locking Schemes to Multi-module Hardware Designs . . . . . . . 138

Dominik Šišejković, Farhad Merchant, Lennart M. Reimann,
Rainer Leupers, and Sascha Kegreiß

Exploration of Power Domain Partitioning with Concurrent Task Mapping

and Scheduling for Application-Specific Multi-core SoCs. . . . . . . . . . . . . . . 153
Bo Wang, Aneek Imtiaz, Joachim Falk, Michael Glaß, and Jürgen Teich

FORMUS3IC Workshop

Scalable, Decentralized Battery Management System Based

on Self-organizing Nodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Andrea Reindl, Hans Meier, and Michael Niemetz

Security Improvements by Separating the Cryptographic Protocol

from the Network Stack onto a Multi-MCU Architecture . . . . . . . . . . . . . . . 185
Tobias Frauenschläger, Sebastian Renner, and Jürgen Mottok

Equally Distributed Bus-Communication Access Rights for Inter MCU

Communication Using Multimaster SPI . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Manuel Dentgen, Sebastian Renner, and Jürgen Mottok

Workshop on Computer Architectures in Space (CompSpace)

On the Evaluation of SEU Effects on AXI Interconnect Within AP-SoCs. . . . 215

Corrado De Sio, Sarah Azimi, and Luca Sterpone

Satellite Onboard Data Reduction Using a Risc-V Core Inside

an RTG4-Based Data Processing Pipeline . . . . . . . . . . . . . . . . . . . . . . . . . 228
Gasper Skvarc Bozic, Thomas Unterlinner, Tanja Eraerds, Sabine Ott,
and Markus Plattner

Workshop on Parallel Systems and Algorithms (PASA)

Accelerating Real-Time Applications with Predictable Work-Stealing . . . . . . 241

Florian Fritz, Michael Schmid, and Jürgen Mottok

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257

Main Concerence
 Approximate Data Dependence Profiling
Based on Abstract Interval and Congruent
Domains

Mostafa Abbas1(B) , Rasha Omar1(B) , Ahmed El-Mahdy1,3(B) ,

and Erven Rohou2(B)
1
Egypt-Japan University of Science and Technology (E-JUST), Alexandria, Egypt
{mostafa.abbas,rasha.omar,ahmed.elmahdy}@ejust.edu.eg
2
Univ Rennes, Inria, CNRS, IRISA, Rennes, France
[email protected]
3
Alexandria University, Alexandria, Egypt

Abstract. Although parallel processing is mainstream, existing pro-

grams are often serial, and usually re-engineering cost is high. Data
dependence profiling allows for automatically assessing parallelisation
potential; Yet, data dependence profiling is notoriously slow and requires
large memory, as it generally requires keeping track of each memory
access. This paper considers employing a simple abstract single-trace
analysis method using simple interval and congruent modulo domains to
track dependencies at lower time and memory costs. The method gathers
and abstracts the set of all memory reference addresses for each static
memory access instruction. This method removes the need for keeping
a large shadow memory and only requires a single pair-wise analysis
pass to detect dependencies among memory instructions through simple
intersection operations. Moreover, the combination of interval and con-
gruent domains improves precision when compared with only using an
interval domain representation, mainly when the data is not accessed in
a dense access pattern. We further improve precision through partition-
ing memory space into blocks, where references in each block abstracted
independently. An initial performance study is conducted on SPEC CPU-
2006 benchmark programs and polyhedral benchmark suite. Results show
that the method reduces execution time overhead by 1.4× for polyhedral
and 10.7× for SPEC2006 on average; and significantly reduces memory
by 109780× and 6981× for polyhedral and SPEC2006 respectively; the
method has an average precision of 99.05% and 61.37% for polyhedral
and SPEC respectively. Using memory partitioning resulted in improv-
ing mean precision to be 82.25% and decreasing memory reduction to be
47× for SPEC2006 suite.

Keywords: Data dependence profiling · Dynamic binary analysis ·

Congruent domains · Interval domains

A. El-Mahdy—On-leave from Alexandria University.

c Springer Nature Switzerland AG 2020
A. Brinkmann et al. (Eds.): ARCS 2020, LNCS 12155, pp. 3–16, 2020.
https://doi.org/10.1007/978-3-030-52794-5_1
4 M. Abbas et al.

1 Introduction
Data dependence profiling is an essential step towards deciding on parallelis-
ing loops, especially for heterogeneous parallel platforms, where the cost of re-
engineering originally serial programs is high. However, obtaining the profiles
is generally a costly operation as it requires instrumenting all memory access
instructions; hence it suffers from significant memory consumption and runtime
overhead (i.e., separate records for every accessed memory address). Therefore,
profilers rely on approximation methods, trading-off accuracy with analysis over-
head. One typical method is sampling the execution trace, where only a portion
of the trace is analysed [5,23]. However, sampling is prone to missing some depen-
dence arcs and losing the most recent relationships. Other profilers tackle the
runtime overhead by parallelising runtime analysis [10,17,24].
This paper considers performing a form of abstract analysis for all memory
references over the whole program. The analysis borrows from the well-known
abstract interpretation static analysis method, which is used to generate an
abstract collective program state for all possible execution traces [1,6,18]. Here,
we specialise the method to analyse only one execution trace (profiling trace).
Moreover, we do not perform abstract computations or interpretations over the
abstract state; instead, we rely on the underlying execution system to generate
the current state and use a corresponding abstract operation (Union) to gather
and approximate the state. In other words, the main aim is to generate an
abstract single-trace semantics, from the current execution trace.
The advantage of this approach is that we dramatically reduce the memory
size required for the analysis, as there is no need to have a shadow memory,
which is typical in profilers. Moreover, we only perform pair-wise operations on
the static memory access instructions after gathering their trace state, to identify
data-dependence, thus also significantly reducing profiling time. The trade-off
here is precision or false dependencies, while the sensitivity or true dependencies
are never missed.
In this paper, for each dynamic memory access operation, we use a corre-
sponding abstract operation that joins the current memory access address with
the current abstracted set of seen addresses for the corresponding memory access
instruction, thus generating an abstract single-trace semantic for all static mem-
ory instructions. The abstraction uses a composite abstract domain, consisting
of interval and congruent domains. The former provides an approximation of the
covered range, and the latter provides information about the access pattern. We
partition the memory address space into blocks and associate a value from the
composite domain for each block, to further improve precision.
This paper has the following contributions:

– Utilise abstract interval and congruent domains to approximate the accessed

memory location for each static memory access instruction.
– Conduct an initial performance investigation using polyhedral benchmark and
SPEC CPU-2006 benchmark suites using the Pin system [14].
 Approximate Data Dependence Profiling 5

– Utilise the gathered semantics to detect pair-wise memory data dependencies

(RAW, WAR, and WAW) from binary files at runtime for all static memory
access instructions.

This paper is organised as follows: Sect. 2 discusses related work. Section 3

provides an overview of abstract interpretation and defines the interval and con-
gruent domains and their corresponding operations. Section 4 discusses our pro-
filing algorithm. Section 5 presents our initial results. Finally, Sect. 6 concludes
our paper and discusses future work.

2 Related Work

2.1 Static Analysis

Abstract interpretation is a static analysis method that generates a collective

program semantics at each program point. The method relies on generating
abstract equations that generate the collective program semantics at each pro-
gram point. Abstract interpretation is not generally used in the context of data
dependence analysis for imperative programs (functional programs have in-depth
treatment, e.g., [3]); nevertheless, there are attempts in this direction [8,16].
Ricci [20] is an early attempt to consider dependence analysis by combining
interval and bisection domains; the latter is used to maintain relations among
pairs of variables. The technique is applied at the source level on some kernels.
Moreover, the use of the bisection domain helped to reduce the interval width,
whereas in our approach we consider the density of the elements in the interval,
as well conducting the analysis on the binary level; however, we do not generate
collective semantics for all possible traces; we only consider the profiling trace,
which is typical for profilers.
The use of the congruence domain in abstract interpretation is described
in detail in Bydge’s master thesis [4]. This paper adopts the same definitions
and operations for that domain. However, the thesis does not consider program
parallelisation.
Tzolovski [22] has discussed possible data dependence abstractions, which
included direction and distance vectors as defined by Maydan et al. [15]. In
this paper, we explore an initial practical implementation of simple dependence
detection and leave more elaborate possible abstractions for future work.
SecondWrite [11] is a static binary rewriter tool. It analyses the dependence
in regular loops in binary programs and workout to rewrite a parallelised binary
version. SecondWrite translates the x86 binary input to the intermediate format
of the LLVM Compiler and then uses the x86 back-end LLVM to write the output
binary. They make use of the LLVM IR rich infrastructure, such as control-
flow analysis, data-flow analysis, and optimisation passes to generate a parallel
alternative version according to which loop can be partitioned.
6 M. Abbas et al.

2.2 Dynamic Analysis

Li et al. [13] introduce a profiler for the sequential and parallel program based
on LLVM. They tackle the runtime overhead by parallelising the profiler. The
profiler records memory accesses using signatures to achieve efficiency in space,
(as an approximate representation, concept from transactional memory [21])
rather than instrument every memory access. The serial profiler has a 190×
slowdown on average for NAS benchmarks and a 191× slowdown on average
for Starbench programs and consumed memory up to 7.856 GB. By using lock-
free parallel design, the runtime overhead reduces to 78× for NAS and 93× for
Starbench. By using a signature with 100 million slots, the memory consumption
reduces to 649 MB (NAS) and 1390 MB (Starbench), with accuracy less than
0.4% false-positive rate and less than 0.1% false-negative rate.
Chen et al. [5] implement a data dependence profiling tool on top of the Intel’s
Open Research Compiler (ORC) to provide information about the dynamic
behaviour of data dependence in programs, mainly for nested loops.
They study two approximation methods: shadow memory and sampling tech-
niques, as a trade-off to mitigate both space and runtime overhead. The tool
speculatively ignores dependence edges (between the source and the sink) that
have low probability. The threshold value of this low probability can be deter-
mined by the overhead of a data mis-speculation process according to the target
machine. The slowdown ranges from 16% to 167% on SPEC CPU2000 bench-
marks compared to the original execution time by using a sampling rate of 0.0001
to 0.1, respectively, and with a precision ranging from 30% to 10% in missing
dependence edges.
Vanka et al. [23] implement a set-based profiling approach coupled with
software signatures. The key insight is that set-level tracking provides a bet-
ter trade-off between accuracy and performance. At compile time, they identify
the essential dependence relationships according to a specific optimisation (i.e.
speculative code motion) for profiling at runtime. The profiling analysis working
on sets, figuring out set’s relationships dependence, rather than working with
pair-wise dependence relationships. The set-based profiler is implemented as an
IR level pass in LLVM and applied to SPEC2000 benchmarks for presenting
results. They achieved a slowdown 2.97× with the accuracy range from 0.15 and
0.17, measured by normalised average Euclidean distance.
Norouzi et al. [19] implement an extension of DiscoPoP data dependence pro-
filer that uses a hybrid (static and dynamic) approach in reporting the existence
of data dependence in the polyhedral loops. The static dependence analysis part
excludes the detected dependent memory access instructions in the annotated
area by PLUTO (an auto-parallelising compiler for polyhedral loops) and hence
excludes them in the dynamic analysis. Finally, static and dynamic dependencies
are merged in an appropriate way to be used later in suitable parallelisation dis-
covery tools. It is clear that if no polyhedral loops are detected in a program, the
hybrid approach turns into purely dynamic. By conducting experiments on Poly-
bench and NAS Parallel Benchmarks suits, they achieved a median profiling-time
reduction by 62% compared to DiscoPoP profiler framework.
 Approximate Data Dependence Profiling 7

Li et al. [12] introduce a profiling method based on repeatedly skipping mem-

ory operations in loops. They used DiscoPoP to implement the profiler. The
experiments on NAS Parallel Benchmarks and Starbench parallel benchmark
suite show a reduction on runtime overhead by 42.5% on average. The reduction
runtime mainly comes from the data dependence building phase, where updating
the shadow memory remains as the traditional way. The effect of this approach
results from the existence of sequencing memory access patterns in loops (e.g.
arrays), if not exists, the profiling suffers from extra runtime overhead due to
the extra conditions compared to traditional one.

3 Proposed Method Formulation

3.1 Gathering Single-Trace Semantics Dynamically
Our method differs from typical abstract interpretation in that we obtain the
abstract collective semantics for a single-trace of the program, instead of all
traces. Also, we rely on abstracting the collective semantics for only memory read
and write operations, not all program points. Moreover, instead of generating
abstract equations, and statically solve or interpret them, we rely on obtaining
actual read and write addresses values at runtime from the underlying execution
environment, and collect the obtained concrete values into collected abstract
semantics (i.e. all possible referenced address seen for each memory instruction).
In other words, we define a corresponding operation that collects the semantics.
Thus, for a memory access operation i reading or writing from/to address a, we
define the abstract collective semantics, Σ at this operation as:

Σi = Σi ∪ α({a}) (1)

Where Σ ∈ DA ; and DA is an abstract domain with partial lattice with

ordering relation ∪, moreover, α() is an abstraction function that abstracts the
current concrete partial collective semantic set into an abstract one. We also
adopt a composite abstract domain that consists of interval and congruence
domains. The congruent domain is helpful to represent stride memory access
patterns, and the interval domain considers the lower and upper limits of memory
access address. It is worth mentioning that abstract interpretation guarantees
sound analysis, where the obtained collective semantics is always a superset of
the concrete collective semantics.

3.2 The Interval Domain

The interval domain [7] is an abstract domain that can determine safe lower
and upper limits of program variables. The abstract interval domain, DI , is
defined as:
DI = {[a, b]}, ∀a ≤ b ∈ Z (2)
The ∪ and ∩ operators are defined as:
8 M. Abbas et al.

[a, b] ∪ [a , b ] = [min(a, a ), max(b, b )] (3)


  [max(a, a ), min(b, b )] if max(a, a ) ≤ min(b, b )
[a, b] ∩ [a , b ] = (4)
φ otherwise

3.3 The Congruence Domain

The congruence domain [9] consists of abstract values denoted, aZ + b, Where
b ∈ Z and a ∈ N. We will call a the modulo and b the remainder. A congruence
relation (c, b) is defined as c ≡ b mod a. The set of all C such that c ∈ C and
c ≡ b mod a, is C = {aZ + b : ∀Z ∈ Z}.
Thus, we define the abstract congruent, DCG , domain as:

DCG = {aZ + b}, ∀a, b ∈ N (5)

The ∪ and ∩ operators are defined as:

(aZ + b) ∪ (a Z + b ) = gcd{a, a , |b − b |}Z + min{b, b } (6)


 lcm{a, a }Z + b if b ≡ b
 mod gcd{a, a }
(aZ + b) ∩ (a Z + b ) = (7)
φ otherwise

Where b ≡ b mod a and also b ≡ b mod a .

4 Profiling Framework
4.1 Pin Framework
Pin [14] is a framework for dynamic binary instrumentation framework. Simi-
lar to other frameworks, users can observe the running code, detect intensive
functions and loops, monitor parameters, and modify the code while it runs.
Pin framework provides an API to let users build custom tools called Pintools,
which in turn dynamically instruments the compiled binary files in the user space
application. By inserting an appropriate runtime analysis routine for a kind of
instructions, we can understand the behaviour of a given binary program.
Our profiler inserts instrumentation code dynamically into the binary code
for each memory read/write operation; which is mainly callbacks to the corre-
sponding runtime analysis routines. The profiler can operate on three different
modes. The first one performs conventional profiling, where the second one per-
forms a comprehensive (i.e. pairwise method) profiling. Both the first and second
modes can be considered as a different perspective of ground truth for the under-
lying data dependence, as it provides an exhaustive, accurate data dependence
Another random document with
no related content on Scribd:
that meritorious nation, the French?”
“Surely, Sir Count, your name is your guarantee,” I rejoined. “At least
I have always understood it to be so.”
“In that particular you are doubtless correct, my dear Don Miguel,”
said the Count of Nullepart, “at least that is when I travel in Spain.
But now we are over the French border I rejoice in a better.”
I inquired his further title with some surprise.
“Upon the curious soil of France,” said the Count of Nullepart, “I go
by the name of Señor Fulano or Mr. What-you-will.”
“I protest, Sir Count, I do not understand this matter at all.”
“I pray you seek not to do so, my dear,” said the Count of Nullepart.
“It is only that I choose to have it so as becomes a free born citizen
of the world.”
I could get no further enlightenment from the Count of Nullepart,
and no sooner had we crossed into France than my mystification was
increased. At the inn at the first town we came to, of which I forget
the name, the Count of Nullepart declared solemnly that we must
speak for him on all occasions, for by a singular mischance he had
entirely mislaid the use of the French tongue. And, further, he
assured us that this grave calamity had had the unprecedented
consequence of stimulating in the highest degree the growth of his
chin hairs. Indeed, this growth was so remarkable that even upon
our first day in France he had acquired quite a large beard.
Instead, then, of the gay, sprightly, and handsome Count of
Nullepart, an admired member of the French nobility, with whom we
had come upon our journey from Toledo, we had for a companion
upon French soil one Señor Fulano, a staid, sober, and bearded
citizen, who claimed cousinship with the burgomaster of the town of
El Dorado, a place of which I had never heard the name.
“I protest, Sir Count,” said I, “there is no such place as El Dorado in
the length and breadth of our peninsula.”
This caused the Count of Nullepart and Sir Richard Pendragon a vast
amount of mirth, and I heard the latter declare that even his
preposterous horse Melanto was chuckling furiously.
“The truth is, good Master Fulano,” said the Englishman, “these
youthful Spaniardoes have so little fantasy as a trussed fowl.
Personally I ascribe their heaviness to the dryness of their climate
and the rough quality of their wines.”
“That is the root of the matter, doubtless,” said the Count of
Nullepart in a most execrable and rustic Spanish which you would
think a gentleman would be careful not to use.
Be this as it may, from the moment we crossed into France, and
during the whole time of our sojourn in that unprofitable country,
the Count of Nullepart, or Señor Fulano as he would have us call
him, had no French at all. Whenever he had occasion to speak he
used Spanish of a most rustic and barbarous sort.
Much as I disliked the country of France, I disliked the people, their
cookery, their manner of speaking, and their extremely foreign ways
even more. As I had small skill in their language, and the Count of
Nullepart had so mysteriously laid aside that which he could claim,
we had greatly to depend upon Sir Richard Pendragon’s knowledge
and adroitness for the least of our necessities. And to allow a due to
the devil—as my countrymen express it—it must be said that the
well-being of three travellers in a foreign country could not have
been in worthier hands.
Sir Richard Pendragon’s use of the French tongue, which I doubt not
to polite ears must have been as unseemly as his use of Castilian,
was so vigorous and his eyes rolled so freely, the name of God and
his Evil Adversary were so constantly upon his lips, and his hand was
so seldom off the hilt of his sword, that the French innkeepers vied
with each other in doing his behests, almost before he had been put
to the inconvenience of making them known. I cannot remember—
although on several occasions he has informed me of the number—
how many temporal kings of whom Sir Richard Pendragon claimed
kinship and acquaintancy, but at least he wore their manners in such
wise as to know how to be obeyed. Full many an innkeeper have I
seen turn pale at his utterance of the word “Sapristi!” And so
surprised were some of them to find themselves alive by the time he
quitted their houses that they forgot to ask him for the score, or
perhaps it was that they feared to do so. At least I know that in
several instances they must have gone unpaid had not the Señor
Fulano thrown them a silver dollar.
I had no favourable impression of this country of France. I suppose
it is a pleasant country; at least I have met those who allow it to be
so; but to the eye of a true Iberian it seemed to lack colour,
politeness, and originality. Besides, as soon as we came to the first
place out of Spain, of which, as I have said, I forget the name, it
came on to rain; and during the whole time we were in this
unfortunate land, which could not have been less than thirty days, it
continued to do so. I know not whether the inhabitants of the
country subsist upon frogs, as was said of them by Sir Richard
Pendragon, but if he spoke truly it was doubtless in obedience to the
dispensation of the good God, for their favourite food was
continually to be seen swimming in the pools that lay in the middle
of their roads.
I suppose it was about the thirty-sixth day of our long and arduous
journey that we came into Paris. It was nightfall when we reached
the capital of madam’s nephew, the famous King Louis. It had been
raining all that day, and the day before that, and it was still raining,
and we were covered with mud as high as our cheek bones. Our
cloaks were soaked through and through and were running over
with water. Further we were hungry and fatigued and in a
desperately evil humour; yet instead of entering the first inn we
came to within the precincts of the city, Sir Richard Pendragon would
have us repair to the auberge of the Compas d’Or, hard by to what is
called the Sorbonne, which Sir Richard pronounced “Sawbones” and
said was the same that in London was called the College of
Surgeons.
This auberge of the Compas d’Or—I have no curiosity to learn what
the name would mean in pure Castilian, but they would tell you
perhaps at Salamanca—was, according to Sir Richard Pendragon, the
best inn at Paris. Indeed, it was a trait I had observed in him that no
matter how hungry or weary or out of humour he might be,
whenever he came to a town or city where there was more than one
inn from which to choose, and some places through which we
passed kept quite a number, he would select the one which had the
best food, the best wine, the best corner in which to sup, and the
best chamber in which to sleep. It was due, he said, to the blood of
kings that its board and bed should be princely.
Thus when we came in this pouring wet night to the auberge of the
Compas d’Or, and we had seen to it that our honest horses were
cared for worthily by the ostlers of this great inn, we entered a large
and comfortable room. And no sooner had we made our appearance
in it than Sir Richard Pendragon’s mode of address occasioned some
surprise to the company we found there.
 CHAPTER XXIV
OF SIR RICHARD PENDRAGON’S PASSAGES
WITH THE GENTLEMEN OF THE KING’S GUARD

The large room was half full of a distinguished company. Many of the
persons there assembled wore a handsome and imposing livery;
others wore an equally handsome and imposing garb of peace.
These gentlemen were engaged in playing at the cards and throwing
the dice, and all were men whose air was lofty. Immediately we had
come into their midst, the proprietor of the auberge—I can see him
at this moment, a little round fellow with a great belly—came
forward half nervously, half uncivilly, crying that we must withdraw
at once, as the apartment was reserved for the gentlemen and the
friends of the gentlemen of the King’s guard.
“Oho!” said Sir Richard Pendragon, in a voice that rose like a
trumpet, “these honest Parley-voos will not look askance on the
hereditary overlord of the Russ, his court chamberlain, and his
second minister. Do you assure them, good Beer-barrel, with our
compliments, that they will find us pleasant good fellows when we
have dried our doublets; but for this present we are cold and
fatigued and most infernally hungry.”
I know not whether the manner of this address, the matter of which
was communicated to me at my own request by the Count of
Nullepart, was an offence to the proud feelings of the gentlemen of
the King’s Guard, but one and all turned glances upon us of the
greatest amazement and austerity.
Sir Richard Pendragon, however, paid them not the slightest heed.
Observing a vacant chair beside a small table, he flung himself into
it, and ordered the keeper of the auberge, in the voice he was
accustomed to use to persons of that condition, to bring us wine and
victual.
“But, sir,” said the innkeeper—I am indebted to my worshipful friend
the Count of Nullepart for all that follows—“you and your friends
cannot remain in this apartment. As I have informed you already, it
is only for the use of the gentlemen and the friends of the
gentlemen of the King’s Guard.”
“Well, you French monkey,” said the English giant, rolling his eyes
fearfully, “you may choose for yourself. Either you obey me this
minute, or, as I am a Christian gentleman, I will cut off your ears.”
Swearing an oath that blenched the cheek of the innkeeper, and
scowling with the ferocity that never failed to cow all of this kidney,
Sir Richard Pendragon drew his sword with a flourish, made a
magnificent pass at the air, and stuck it at one pace from him in the
wooden floor.
I think I have never seen more amazement in the human
countenance than this action excited in all who witnessed it. At first
the onlookers seemed unwilling to believe their eyes. That any
human being should enter their presence and thus bear himself was
a thing they could not grasp. And then, when they came to realize
that the Englishman regarded their presence no more than he did
that of the innkeeper, a kind of pitying contempt came into their
faces.
Nevertheless, some little time went by ere they addressed Sir
Richard Pendragon. Conversing together in low tones, they appeared
to wait upon the good pleasure of one among them. Then they
called the landlord, who stood awaiting their commands, and gave
him certain instructions.
Upon receipt of these the keeper of the auberge approached Sir
Richard Pendragon, yet with a good deal of wariness, and said,
“Monsieur, I am instructed by the gentlemen of the King’s Guard to
inform you that, whoever you may be, your behaviour is intolerable.
But as you and your companions are clearly of a foreign nation, they
are loth to admonish you. Yet I am to inform you that if you do not
immediately put up your sword and withdraw from this apartment,
you will compel them to visit you according to your merit.”
Now, although the keeper of the auberge, having both right and
might at his elbow, had spoken with a well-considered civility which
is rare in his class, and the words that he had been instructed to use
were those of an admirable moderation, which in the circumstances
did honour to his patrons, they were not accepted by Sir Richard
Pendragon in a spirit of forbearance.
“Do you presume to outface a Pendragon, you French dog?” he
roared. “For a pint of sherris I would pull your neck.”
Speaking thus, the Englishman took up a cup half full of the wine
that was near to him, and flung it full at the head of the innkeeper.
In spite, however, of this new affront to their ambassador, the
gentlemen of the King’s Guard showed no disposition to hurry their
measures. Again they conversed among themselves; and then a
thin, tall man, with a visage exceeding melancholy, not, however, in
the king’s livery, yet attired in a dress of sober richness, rose slowly
from the table at which he had been playing at the cards. There was
something of majesty in his movements, and as he approached the
Count of Nullepart and myself with a cold air, his mien was worthy of
a cardinal.
“I would speak with you, my friends,” he said in a deep and musical
voice, yet the tone was such as he would have used to his lackeys.
The Count of Nullepart shook his head solemnly, as though he
understood not a word, and said in a rude Spanish, “I have not your
language, Señor Soldado.”
I had to make a similar confession, but, as I hope, in a purer idiom.
“Muy bien,” said this distinguished French gentleman, speaking in a
very tolerable Spanish that put the Count of Nullepart’s to shame
and compared not unfavourably with my own, “Very well, my
friends, a word in your ears. Your conduct is worthy of the highest
censure, but the gentlemen of the King’s Guard are not accustomed
to turn their hands against the canaille. All the same, they pray you
to have a care.”
Thus having spoken with a degree of insolent contempt that few
could have equalled, this Frenchman, and I am sure among his own
nation he must have taken rank as a great lord, turned his back
upon us with a high degree of disdain, and proceeded to regard Sir
Richard Pendragon. The English giant met him with a sleepy
indifference. Thereupon the Frenchman lowered his gaze to an
amused contempt, and withdrew Sir Richard Pendragon’s sword from
the floor.
After examining this weapon with a care that was only half curious
he gave his shoulders a shrug, after the foreign manner, and then
presented it to the Englishman by the hilt, saying, “Put up your
butter-cutter, Monsieur l’Epicier, and when you return into your
peninsula give an additional alms to the Virgin that you find yourself
with as whole a skin as that with which you went.”
Being addressed in this fashion, an odd change fell upon the
Englishman. As in the affair in the inn at Madrid, a kind of sinister
softness overtook him. Immediately he abated his voice into a
modest and humble accent which was quite unlike his previous
immoderation.
“I thank you, good Frenchman, for my poor tuck. It is an ancient
arm, I might say an heirloom; yet once on a day it held the rank of a
sword. At least, in that capacity was it given to an elderly forebear
by Edward the Black Prince, who in his day did some pretty work
among the French. And now, as you say, although it is an old thing,
it still serves to cut butter.”
Thereupon, in the presence of the whole room, which had
suspended its affairs entirely, Sir Richard Pendragon quietly laid the
flat part of the sword against one side of the Frenchman’s cheek and
then against the other.
 CHAPTER XXV
OF SIR RICHARD PENDRAGON’S DUELLO
WITH THE GALLANT FRENCHMAN

Any excitement that was likely to arise was checked by the

Frenchman’s action. With a dark and cold smile on his lips, he turned
to his friends and held up a slender white hand that was covered
with jewels, and besought them, almost in a tone of entreaty, to
display calmness.
Then with a most courteous apology to all who had suspended their
play, and remarking that it was plaguy unfortunate that he must
suspend his own when the cards had smiled upon him for the first
time in a long season, he ordered the landlord to have all the room’s
furniture drawn close to the wall.
While this was being done, the Count of Nullepart went to the
Englishman and addressed him privily.
“This fellow,” said he, “is the first swordsman in France. He is the
hero of a hundred duellos, and he is quite invincible.”
“Is he so, my dear?” said the English giant in his modest voice,
which seemed to feign alarm. “How pleasant it must be for the poor
soul to be invincible.”
Sir Richard Pendragon turned to me and said in a manner of
courtesy I had seldom heard him use, “Prithee, good Don Miguel,
oblige old honest Dickon by going into the stable yonder and
procuring that little rapier of Ferrara steel, which you will find
strapped to the saddle of the meritorious Melanto, who is now eating
his supper of oats like a good Christian horse.”
In obedience to this request, I went forth to the stable to procure
the rapier of Ferrara steel. Upon returning with it into the room, I
found that a goodly space had been cleared in the centre. Both
parties to the duello were standing therein stripped of their doublets.
The spectators, exceeding a score in number, were seated on settles
which were ranged close to the wall.
It was curious to observe the looks of mingled contempt, pity, and
derision of these persons when I approached the Englishman and
handed to him the Italian rapier. Some of them were unable to
repress their mirth. They laughed out loudly, as though my action
was the height of the ridiculous.
Before I had taken the chair that was offered to me, these
adversaries had crossed their swords. If I live to be an hundred
years old I shall never forget the battle that ensued. At the first
shock of steel against steel it was clear that each recognized in the
other a foremost swordsman of the age.
The knowledge did not induce fury nor any kind of excitement. It
rendered them calmer, more wary, and subtle than they would have
been otherwise. And the gentlemen present, each of whom, as the
Count of Nullepart informed me, was a master of the sword, began
soon to realize that one of their peers had come quite unexpectedly
into their midst.
It was not all at once that this was made clear to them. At first they
regarded the contest with smiles that were merely mocking and
incredulous. Naturally it seemed the extreme of presumption that
such a fellow, whose manners and appearance were so barbarous,
should venture to stand up with a delicate and slender Italian
weapon before the first swordsman of the time. But so soon as their
true blades had met, the company began to exchange significant
glances one with another, and in a very little while they realized that
this was no tyro who stood before them.
From the first it was beautiful play. Owing to his stature, it was
necessary that the Englishman should lose something of elegance in
the comparison with his inimitable adversary, but, stroke and
counterstroke, they were perfectly matched.
The spectators inclined to the opinion at first that Monsieur du
Bartas, for that was the name of their champion, the foremost in all
France, was not putting forth the whole of his art; but when
presently they came to perceive how easily his deftest strokes were
turned aside, they began to waver.
It was a long duello, yet it was one of which every phase was
memorable. These two wonderfully accomplished men began to
weave a spell upon their audience; and as their actions grew quicker
and the finer shades of their play declared themselves, the
spectators began to lean forward out of their seats, and with the
loud and ringing music of the steel was mingled “bravas” and all
kinds of applause.
As the combat proceeded the excitement grew more intense. The
spectators seemed not to breathe. For the first time in his career
their invincible champion stood in danger. No matter what the
cunning and the incomparable skill of his devices, it began to appear
that unless the unforeseen occurred to save him he went in danger
of his life.
It was then that the buzz of voices and the murmurs of applause
grew hushed, and soon the gay shouts, the sneering smiles, the
sarcasm of their commentary, yielded to a dead silence. The circle of
onlookers craned ever closer to the combatants, yet now not a word
was spoken; and upon the faces of many there was a mingled
surprise and consternation that they sought not to conceal. For the
countenance of the first swordsman in France was growing livid. The
sweat had crept upon his brow. Proud and brave man though he
was, he had begun to feel himself in the grip of a power beyond his
own.
As with amazing skill the Englishman parried stroke after stroke,
which were themselves the fine flowers of his adversary’s talent,
each one of which must have sufficed to place one less in genius out
of his life, I overheard a bewildered gentleman of the King’s Guard
say to a companion, “This fellow must be the Devil!”
All at once and quite suddenly there came the sound of bare steel
striking the ground. The celebrated Frenchman, the hero of a
hundred duellos, stood without a weapon before his adversary.
It was a moment I shall never forget. The sympathies of the Count
of Nullepart and myself were of course engaged upon the side of Sir
Richard Pendragon; but as this noble and imperious French
gentleman stood with head upheld and a look of disdain upon his
lips to receive the penalty of his failure, I think, in common with all
the other witnesses of this splendid encounter, the count and I
would have been only too eager to avert it could we have done so.
Yet Monsieur du Bartas looked not for mercy. He was known as one
who neither gave quarter nor expected to receive it. None,
therefore, looked for mercy for him. The Englishman had gained the
victory in fair fight; it was perfectly just that he should enjoy its
fruits. Such an expectation, however, merely shows how imperfect a
thing is the science of reason, and how simple it is to do less than
justice even to our friends. For none could have foreseen, least of all
the Count of Nullepart and myself, that Sir Richard Pendragon, one
of a rude and uncivilized nation, would stoop to pick up a fallen
sword and, with a bow that would have become the most
accomplished of courtiers, return it to his conquered adversary.
It was then that the silence of the gentlemen of the King’s Guard
yielded to expressions of pleasure. They crowded round the victor,
shook him by the hand, paid him most flattering addresses; and
nothing would content these Frenchmen, regaled as they had been
by great generosity and the highest skill, save that the Englishman
and also his companions should remain with them, drink their wine,
and partake of their hospitality.
Indeed, the rest of the evening was the most delectable it could
have been given to any travellers to spend. We were treated with
the most distinguished consideration by those who were not
accustomed to exercise it on a light pretext. Yet it soon became clear
that they had come to regard their guests as mysterious.
The French, as all travellers allow, are people of quick parts; and
while they feasted and flattered us, it was plain to these gentlemen
that we were other than we seemed. With all the courtesy possible,
for even the Englishman could doff his brusque manners when it
suited his humour, we declined to disclose a word of our embassy,
contenting ourselves merely with inquiring whether the King was at
Paris. It was our good fortune to learn that he was.
As the evening passed we felt ourselves to be the objects of a
particular scrutiny. Our new friends grew most curious concerning
us. After we had supped they asked us to take a hand at the cards.
My two companions accepted this invitation; but I refrained from it
because I was still very poor, and perhaps quite as conclusively
because my northern breeding induces the virtue of caution.
No sooner had Sir Richard Pendragon and the Count of Nullepart
begun to play than the interest of these gentlemen grew lively
indeed. This was not so much aroused by the demeanour of the
conqueror of M. du Bartas, but rather by that of the Count of
Nullepart, who, in spite of his long beard and rustical Spanish,
betrayed his true condition in unsuspected ways to those who were
themselves high-born.
First, they observed his white and shapely hands as they lay upon
the board. Again, there was the delicacy of his features, the natural
politeness of his gestures; and yet again, they could not fail to
detect the subtle and charming quality of the accent that lurked
beneath his assumption of a rustical brogue. Thus it fell out that
presently they came once more to confer among themselves, and
then one of these gentlemen said with a most profound respect to
the so-called Señor Fulano, “By my life, monseigneur, is not this a
——”
Before this gentleman could conclude his remark the Count of
Nullepart answered him in his charming natural speech and in the
French tongue, “My good Clery, you are seldom prudent in the
evening; and I am told it is entirely due to this misfortune of yours
that you still remain without advancement. All the world knows that
at Paris nowadays one must learn to see little and to speak less.”
No sooner had the gentlemen of the King’s Guard recovered from an
astonishment which at first seemed to overwhelm them than they
began to shout with laughter. For what reason they should thus view
the matter I cannot explain; but from that moment a more formal air
was imparted to the assembly. Something of ceremony declared
itself, and the manner of all present became perceptibly less easy.
Still this may in a measure have been due to the fact that these
gentlemen put forth remarkable and in some cases highly ludicrous
efforts to conduct their discourse in Castilian.
In one particular, however, our singular companion saw fit to rebuke
them. They persisted in bestowing upon him the title of
“Monseigneur,” so that presently he was moved to exclaim, “I pray
you remember, my good friends, that I am neither more nor less
than the worshipful Señor Fulano, worshipful kinsman to the
worshipful burgomaster of the worshipful town of El Dorado. Beyond
this I claim no title. And when the Señor Fulano comes among you
to-morrow at the Louvre, which a conspiracy on the part of fortune
has rendered necessary, I pray you not to call the honest person out
of his true degree.”
Laughter and surprise greeted this speech; yet a kind of respect was
paid to it, and during the rest of the evening they were careful to
heed this request.
These gentlemen sat at their gaming far into the night. The play was
high; gold pieces were numerous, being piled upon the table and
exchanged freely. Also they drank an immense quantity of a very
superior kind of red wine. Whatever the individual fortunes of the
players, and of these I cannot speak, there was at least one among
them who rose from the board considerably richer in the things of
this world than when he sat down at it. I allude to Sir Richard
Pendragon. Both the Count of Nullepart and myself were fain to
observe that, whenever it came to the Englishman’s turn to take the
dice in hand, quite as often as not he would have the singular good
fortune to cast the double six.
 CHAPTER XXVI
OF OUR APPEARANCE AT THE LOUVRE BEFORE
KING LOUIS

At noon the next day we set out upon our embassy to the King of
France. However, before so doing, at the instance of Sir Richard
Pendragon we repaired to a furrier’s shop in a little narrow street
behind the church called Notre Dame, which the Count of Nullepart
informed us was the first in the city. Here we purchased three
baldricks of an extraordinary brilliancy, trimmed with ermine.
To our surprise, Sir Richard Pendragon disbursed the sum necessary
to this magnificence, for his winnings of the previous night had been
considerable. Besides, as he declared, it was due to our mistress
“that the plenipotentiary-extraordinaire of the young queen’s majesty
should appear at Paris like a man of condition, and that the retinue
by whom he was accompanied should appear in the same guise,
because they had worldly minds at the French court, and it would be
easier to conduct state business if they went upon terms of
familiarity with the current mode.”
To this piece of wisdom the Count of Nullepart assented laughingly.
And when our leader came to put on his baldrick this mirth bubbled
up to a point, for Sir Richard Pendragon was fain to add to it a pair
of new shoes with large silver buckles and a handsome collar of lace.
I confess that in this I approved of the Englishman’s conduct. And I
think we both felt the Count of Nullepart’s laughter to be somewhat
ill judged and out of place. Because in a city like Paris, which in the
light of day is not unpleasing, and among such a people as the
French, whom travellers allow to have a savour of the civilized arts,
we deemed that a certain richness among men of birth was not only
expedient but necessary.
Therefore I put on my new baldrick trimmed with ermine, also my
new shoes with silver buckles and my fine collar of lace which also
had been given to us. And let me tell you, reader, that never in my
life have I felt myself to be attired more worthily, a little the plain
side of splendour. Even then I felt that I did not compare with the
leader of our embassy, who, as he said, “to remove any lingering
traces of the provinces,” added to the bedizenments of his person a
number of jewels which his good fortune of the previous night
enabled him to obtain; and, further, as a crown to the whole, a sort
of jewelled cockado that is worn by the potentates of Eastern climes.
To obtain a field for the display of our magnificence, of which I
believe the three of us were proudly and justly conscious, we
proceeded slowly, arm-in-arm, down the centre of the streets of
Paris; and of almost every second person that passed us Sir Richard
Pendragon inquired in a haughty voice of the way to the palace of
the King of France.
I suppose it was our high and martial looks in company of our
resplendent attire—I may say that Sir Richard Pendragon had chosen
scarlet for the colour of our baldricks, that they might contrast
elegantly with the bright yellow of his own—that soon began to
attract the notice of the Parisians. Ere long a number of these
curious persons were following in our wake. By the time we had
traversed the length of two streets something of a crowd had
collected upon our heels; and this circumstance appeared to afford
Sir Richard Pendragon a great deal of pleasure.
“These good souls can see we are on the way to King Lewie,” said
he. “I am perfectly sure they mistake me for the Emperor
Maximilian, although I have five inches the better of my old crony in
the matter of perpendicularity, and at least six in the matter of
circumference. Still they cannot be expected to be informed of it.
And prithee, good Don Miguel Jesus Maria de Sarda y Boegas, do
you observe how all eyes are for my Persian cockado? I doubt not it
looks very brave with court livery; and it will afford me not the least
surprise if King Lewie, who is a good fellow, comes to adopt it at the
court of France.”
Just as the English giant had concluded this speech, a little old
woman came up behind him and plucked him eagerly by the cloak.
“Good Master Tumbler,” said she, “if you will stand on your head I
will give you a groat; and if you will swallow your sword you shall
have a new franc piece.”
Sir Richard Pendragon plucked his cloak away fiercely from the old
creature and walked on with his head in the air, as though he had
not heard her. During the next moment, however, an unmannerly
urchin had thrown a cake of mud at the Persian cockado.
By the time we had come to the gates of the Louvre, the press was
so great that it had become difficult to proceed in it. Indeed,
according to the Count of Nullepart’s computation, and he seemed to
derive much pleasure from assessing it, it could not have been less
than a thousand persons.
To the astonishment of our leader, when we came before the gates
of the palace, the soldiers of the King’s Guard, who kept the royal
entrance, declined to allow us to pass. And when Sir Richard
Pendragon threatened peremptorily to cut off the ears of their
captain, the prospect of our gaining admittance did not seem to
improve. For some reason, which I cannot explain, the attitude of
the King’s Guard seemed greatly to please the mob that was
pressing around us.
Sir Richard Pendragon was fain to produce the cartel of our mistress
duly sealed and inscribed: “To Lewis Our Nephew in His Court at
Paris, by the hands of Our Good Servants.” Yet even this document
went without effect, if only for the reason, as the Count of Nullepart
assured us, that the captain of the King’s Guard was unable so much
as to decipher the superscription.
In the next minute there was almost a riot in the open street. The
English giant, seeming to detect cries of derision arising about him,
turned to the ever-increasing multitude and observing a low fellow
that was near him in the act of making an insulting grimace, he
made no more to do, but lifted him up bodily, and flung him like a
sack of flour upon the heads of the people.
Upon this, mud and stones began to fly past us. And a missile
having struck Sir Richard Pendragon upon the cheek, he drew his
sword and began to lay about him lustily with the flat of it.
Our situation was now one of great peril. Three persons, whatever
their valour, were powerless to defend themselves from a press of
this magnitude. I incline to think our fate would have been a sorry
one had not the mother-wit of the Count of Nullepart arranged our
deliverance. While the mob were surging angrily around us and
stones were flying about our ears, our companion spoke some words
in a low voice to the captain of the King’s Guard, and this time he
used the French tongue. The effect was like magic. The captain
instantly removed his plumed hat, and bowing very low, led us
through the gate and into the precincts of the palace, leaving his
company to deal in what sort it suited them with a mob that by now
was in no gentle humour.
Once within the walls of the palace, the Count of Nullepart dismissed
the King’s officer with a word of thanks; and then, under the count’s
own direction, we entered an exceeding large antechamber, which
was thronged with as fine a company as I have ever beheld. There
were priests of high learning and dignity, wearing their soutanes;
there were soldiers in bright doublets and shining armour; there
were austere and sombre-coated ministers; there were gay and
handsome courtiers in very modish and brilliant attire; and beyond
all else there was a number of beautiful ladies.
This fine company was talking very loudly and laughing very gaily at
the time we came into the room. But our entrance being a public
one, mainly owing to the manner in which Sir Richard Pendragon
clanked his spurs on the marble floor and the great voice in which
he conversed with the Count of Nullepart, the attention of all present
was immediately drawn upon us. Now I know not whether it was
due to the magnificence of our apparel or the pride of our bearing,
yet the lively talk and the gay mirth subsided in the most sudden
manner. Each person in the room seemed to turn to regard us with a
wonderment that scorned disguise; and then the silence was broken
by a titter from one of the fine ladies.
The court gallants who surrounded them were not slow to follow
their example.
The leader of our embassy, however, was not disconcerted in the
least by this public rudeness. Sir Richard Pendragon stroked his chin
with a disdain that appeared to amuse these courtiers the more; and
then, turning at his leisure to the richly attired gallant that was
nearest to him, he said in a voice like thunder, “Hi, you, sirrah, you
with a face like a monkey, do you go to the King your master, and do
you inform him that an embassy is come from Spain upon an affair
of delicacy.”
The youthful courtier placed his jewelled fingers on the hilt of his
sword. His unseasonable mirth was now changed to a look of
ferocious anger.
“Do you hear me, good jackanapes?” said the English giant in his
great insolent voice that surmounted everything and re-echoed to
the high ceiling upon which was a painting of Venus and Cupid.
“Mon Dieu!” cried the courtier, livid with passion, “I have a mind to
run you through the body, you canary-coloured barbarian!”
“A mind, did you say, good jackanapes?” said the Englishman, with a
roar of laughter. “Why, a thousand such poor dogs could not muster
a mind among you.”
By now all the persons in the room were gathered around us. The
grave among them were amazed; the young, and particularly those
that were female, shaken with mirth; and the rest in all degrees of
anger, incredulity, excitement, and a desire for diversion. Yet so
sorely incensed was this youthful gallant that I verily believe, the
place and the company notwithstanding, he would have been moved
to an act of open violence to avenge the insult that had been set
upon him, had there not stepped forth from the throng one who
bore every mark of dignity and high consideration.
“I ask your pardon, Monsieur Ambassador,” said he with a courtesy
that was very grave, “but if it is your desire to have an audience of
his majesty the King, will you have the good kindness to accompany
me into another room.”
“I am at your service, mounseer,” said the Englishman. “I will go
with you willingly. It will give one who carries the blood of kings
under his doublet a great deal of pleasure to escape out of this
kennel in which his cousin of France keeps his puppy dogs.”
Speaking thus, our leader threw a glance around him of great
effrontery, which ministered further to the amazement of those who
were present. He then followed this high officer of the court into
another room. The Count of Nullepart and myself accompanied him.
Here we found ourselves alone, which, considering Sir Richard
Pendragon’s present humour, I cannot help thinking was a fortunate
circumstance. The chamberlain withdrew in order to convey our
business to his royal master. No sooner had he done so than the
Count of Nullepart broke forth into an outburst of inextinguishable
laughter.
Sir Richard Pendragon viewed the Count of Nullepart’s demeanour
with a grave disdain. Further, he assured me privily, that “a man’s
nation could not hide itself when his foot was on his native soil.
Mounseer Nullepart was a good fellow enough, but there was no
mistaking his nationality.”
In so far as the Englishman deplored the Count of Nullepart’s levity I
was in accord with him. Yet, for my own part, having the sangre azul
of Spain in my veins, which is apt to insist that a courtly bearing is
beyond all things essential to him who would converse with the
great of the earth, I could not help but regret the manner in which
our leader had invaded the palace of the most Christian prince.
As we remained thus to await an audience of the King of France, I
began to fear dreadfully lest the leader of our embassy should
mislay his manners before the Sovereign. The walls of the room
were covered by mirrors; and as Sir Richard Pendragon stood before
each of them in turn, preening himself like a bird of bright plumage,
now with his bonnet on his head to judge the appearance of his
Persian cockado, now with it off to see how he seemed without it, I
grew sensible of a concern for the affronts our singular leader was
like to put upon the Father of his People.
Six times Sir Richard Pendragon put his bonnet on before the
mirrors, and six times he took it off again. He then sighed deeply,
and said, “Prithee, good Miguel, in how far would you consider that
Spain is a civilized nation?”
“Good Sir Richard Pendragon,” I said, “surely your question asks not
an answer. From the time of the Cid, as all the world knows, Spain
has been the most civilized country on the face of the earth.”
“I understand that perfectly, good Don,” said Sir Richard Pendragon.
“But making abatement for your native peninsularity, which in its
due place and season is commendable, I would ask you whether, in
my capacity of plenipotentiary-extraordinaire to a Spanish princess, I
might come before the King of France wearing my bonnet, because I
find this Persian cockado sets off my countenance in a very proper,
majestical, modish, yet not foppish manner.”
“Good Sir Richard Pendragon,” answered I, “I do conceive that one
who has the sangre azul of Spain in his veins may be allowed to
answer your question judicially. Nothing could less beseem a
representative of Spanish nobility than that wearing his bonnet he
should enter the presence of a Christian sovereign.”
This opinion caused Sir Richard’s face to fall.
“It could be done by the ambassador of the Ottoman Empire,” said
he, “and the Turks are religious-men. The representatives of
Morocco could do it also, and the Moors are a very ancient people.
And of course at Teheran it is the mode. And if this Louis, this frog-
eating French fellow, were mine old gossip Maximilian, whose
kingdom is four times the size of France, the thing could be done so
easily as you might count nine.”
“Good Sir Richard Pendragon,” I said gravely, “this act which you
contemplate would be a blot upon the fair fame of Spain, of which
these many years we have been so jealous.”
Now I think my demeanour must have convinced Sir Richard
Pendragon that my opinion was a just one, had not the Count of
Nullepart, who had laid aside his mirth to listen to our conversation,
interposed an opinion of his own. And that opinion, as I grieve to
inform the gentle reader, was far from agreeing with the one I
myself had given.
“Good Sir Richard Pendragon,” said the Count of Nullepart in his
most subtle and melodious accent, “it seems to my mind that these
parallels you have been learned enough to adduce from
Constantinople, Tangier, Teheran, and other centres of light are
extremely pregnant to this embassy. If the measure of civilization in
such places—and as you say, in those countries religion is not
unknown—would permit the diplomatic body to appear bonneted à
la Persie before a crowned Christian prince, it seems to me that you
have furnished the clearest reason why you should conform to their
usage.”
“You speak well, mounseer, you speak well,” said Sir Richard
Pendragon with a complacent air.
“My good friends,” said I, “I deplore the fact that these are not my
views. Let me assure you that the act you contemplate would be far
from the dignity of Spain.”
The Count of Nullepart, observing that I was exercised upon the
subject, was good enough to make a proviso.
“Perchance, good Sir Richard,” said he, “there is one formality we
should observe if we would enter the presence of majesty bonneted
à la Persie. We owe it to the dignity of France, I think, that we follow
the practice of Mohammedan countries. If we wear our bonnets, it
seems to me that we must remove our shoes.”
To this proposal Sir Richard Pendragon seemed loth to assent. The
Count of Nullepart, with great courtesy, appealed to my judgment.
Now I, although extremely reluctant to appear in my bonnet before
a great Christian sovereign, yet felt that if such a course was
imperative, the Count of Nullepart’s suggestion came from a quarter
where breeding was admired. So familiar was he with the temper of
courts, and so firmly did he adhere to the opinion that the removal
of our shoes was necessary if the leader of our embassy was
determined to wear his bonnet, that I gave my sanction to this
proposal. But it was not until we had had further controversy upon
the subject that Sir Richard Pendragon, still declining to remove his
bonnet, at last consented to take off his shoes.
“Perhaps,” said he, as he reluctantly removed them, “it will give
France a better notion of our breeding.”
However, when he had discarded them and he came to survey their
buckles, he grew discomposed in his mind. He had purchased them
expressly that morning, and very handsome and imposing did they
look.
“By my good soul,” he said, “I am not at all clear that silver buckles
do not make a better appearance than Persian cockadoes in the
palaces of the West.”
“It is a mere matter of taste, my dear Sir Richard,” said the Count of
Nullepart, smiling.
Yet the count had already followed the example of his leader, having
put on his bonnet and having doffed his shoes. I also had deemed it
necessary to do the same.
Therefore, when the grave French nobleman presently returned to
say that the most Christian King would see us in audience, he found
us seated in somewhat remarkable case.
 CHAPTER XXVII
OF OUR AUDIENCE OF THE MOST CHRISTIAN
KING

I think the Marquis de Contreville-Lancy—that, as we afterwards

learned, was the name of this gentleman—had some little surprise
when he saw in what fashion we were disposed for our audience of
the King, his master. Yet, if surprised he was, and I think, good
reader, in this instance a little of such an eminently natural feeling is
to be pardoned, he was far too grave and serious a nobleman to
display it unduly. Yet I feel sure he viewed our appearance not
without displeasure, and I believe it gave as high a relief to his
feelings as it did to my own, when Sir Richard Pendragon coming to
stand up, exclaimed, “This French marble is plaguy cold to the feet.
Upon my good soul! it is a kind of distemper to buy a pair of shoes
tricked with silver and then to walk barefoot before the king’s
majesty.”
“All the same, good Sir Richard,” said the Count of Nullepart, “it is
well known that the etiquette of the French court is very nice.”
In consequence of Sir Richard Pendragon’s new qualms, the Marquis
de Contreville-Lancy was taken into our counsels. And I feel bound
to state that the reference to this dignified nobleman proved highly
fortunate. He persuaded Sir Richard Pendragon not only to don his
shoes, but also to doff his bonnet. For he declared that any other
proceeding would gravely imperil our embassy.
This piece of whimsicality being thus happily adjusted, we repaired
in a wholly civilized mode to the presence of the first prince of the
age. I cannot tell you, reader, what were my feelings when for the
first time in my life, and at a period when I had barely attained to
the estate of manhood, I found myself within a few paces of so
august a personage.
Upon first coming into the presence of King Louis I could observe
very little, for a most singular haze rose before my eyes. When
afterwards I came to mention this phenomenon to the learned Count
of Nullepart, he said that all who entered the presence of majesty
were thus afflicted. It was a kind of exhalation, he said, which
embodied their divinity. At the time, however, I was not aware of this
interesting fact in natural philosophy. I only knew that there was
nothing in the apartment that I could descry at all clearly, yet,
understanding by a kind of instinct that my two companions were
bowing low, I followed their example.
When at last I could see the King more fully he was conversing with
Sir Richard Pendragon. The remarkable man who had come to lead
our embassy had the seemliness to conduct himself with a most
polite civility, of which I had scarcely suspected him to be capable.
After humbly saluting the hand of the monarch, he paid King Louis
some highly flattering addresses, and sinking to one knee—an act of
courtly homage that was so well performed that it must have been
the fruit of long practice—presented to the King the cartel of our
mistress.
While one of his ministers read the terms of the reference aloud to
the King, who, of course, was too great a personage to read it
himself, I was able to muster my wits sufficiently to mark the most
Christian prince. And, good reader, you will doubtless call in question
the veracity of my two eyes when I assert that the French King Louis
was a small, wizened, pock-marked man, with a face, as became the
embodiment of his nation, that was not unlike a frog’s. His hair was
red in colour; there was a marked cast in his right eye; the lids were
twisted and puckered in a most curious manner; his insignificant
person and particularly his puny hands were twitching constantly;
his voice was not agreeable; and I could not decide whether the
colour of his eyes was brown or grey or dark green—the Count of
Nullepart inclined to the latter opinion, Sir Richard Pendragon to the
former—yet, good reader, I assure you solemnly that,
notwithstanding all these disabilities, the French sovereign was every
inch a prince.
There were three or four of the King’s ministers present in the room,
with ruffs about their necks and short pointed beards. When they
had read madam’s communication, the King seemed puzzled to
know who his correspondent might be. It appeared that he had not
the faintest recollection of his aunt. And, as I conceived, somewhat
singularly, the ruler of France showed a livelier concern for this
relationship than for the demand that we had come to make upon
his friendship.
At the command of the King an enormous genealogical chart was
brought into the room. Being laid upon the table, it provoked the
greatest curiosity. One and all scrutinized it with diligence, King Louis
with an even shrewder regard than the rest. And presently he was
able to convince himself that in Spain there dwelt an aunt of his of
whom he had neither seen nor heard.
The King having duly established their kinship, proceeded to ply Sir
Richard Pendragon with some pertinent questions, which the
Englishman answered in the grave manner of a true ambassador.
“A proud and royal lady, good your majesty,” said Sir Richard
Pendragon, “every inch a queen. She hath loyal and shrewd
advisers, good your majesty, among them these famous gentlemen
and one who shall be nameless.”
The King desired that the Count of Nullepart and myself should be
presented to him. Sir Richard Pendragon undertook this office. With
an air of magnificence that nought could surpass he recited our titles
and our merit; and had we discovered anything to be lacking in the
character he gave us we must have been ambitious men. It seemed
that the most noble the Marquis of Fulano was the most
accomplished minister in Spain; while I was a great lord, and, in
spite of my tender years, by no means the least of its captains.
“And may I assure you, sire,” said this extraordinary envoy, “the
most excellent queen’s majesty hath a standing army of not less
than forty thousand men-at-arms.”
This revelation of the puissance of our mistress was not without its
effect upon the King and his ministers. At first they seemed to
extend only a languid interest to our business, but the manner in
which Sir Richard Pendragon conducted it began to stimulate their
attention.
Indeed, it must have taken very cold-blooded statesmen not to
derive interest from Sir Richard Pendragon’s presentment of the
case. Taking the genealogical chart as a kind of map of Spain, with
one finger he indicated the stronghold of the King of Castile, at fifty
leagues’ distance from Toledo, and with another traced his broad
dominion upon which our mistress was already marching with forty
thousand men-at-arms. He declared that it was the Countess Sylvia’s
plan to draw the King of Castile out of his stronghold by falling upon
and laying waste his unprotected lands; and then, while the Castilian
was fully engaged in defending his own, she proposed that an army
of her royal nephew’s should come up in the rear—quite
unexpectedly, for the nature of this present mission was known to
none—enter the unprotected stronghold of the former enemy of
King Louis, and seize it for France.
With such conviction and enthusiasm did Sir Richard Pendragon
expound a plan of war that was entirely of his own invention, that
King Louis, who at first had not listened very closely to a proposal
that was outside the sphere of his politics, began to nod his head in
approval. And finally the King of France was moved to hold animated
converse with his ministers.
After they had spoken together for some while, one of these
gentlemen directed Sir Richard Pendragon’s attention to the fact that
although ten thousand was the number of soldiers that he asked for,
according to the tenor of madam’s petition the number was no more
than four thousand. Thereupon Sir Richard Pendragon produced a
large pair of horn spectacles, adjusted them gravely, and after
scrutinizing the parchment very carefully, although there is reason to
believe that madam’s ambassador was able to read little of the
Spanish tongue, he gave it as his opinion that the four thousand was
an undoubted error of the scrivener’s, inasmuch—and he spoke very
truly in this particular—that the author of the proposal was himself.
This assurance being given, the King and his ministers again
conferred; and presently the audience was terminated by his Majesty
saying that this affair was of such moment that he desired twenty-
four hours in which to sit in council with his advisers. The King then
took leave of us, yet with a courteous request that the envoys of his
respected aunt—who, according to Sir Richard Pendragon, was a
learned and devout lady of mature years—should be lodged in the
palace during their stay at Paris, and further, that they should
engage themselves to dine with him that evening.
 CHAPTER XXVIII
OF FURTHER PASSAGES IN THE LOUVRE AT
PARIS

When madam’s three envoys came to find themselves in the private

apartments that had been given to them by the King of France, I
had no words in which to express my amazement at Sir Richard
Pendragon’s audacity. When I remembered that the Countess Sylvia
was scarcely more than a child, with a beggarly retinue of three
hundred men-at-arms, who would be wholly incapable of holding the
castle of Montesina against the Castilian host; and when beside this
dismal truth I set the dazzling story by which Sir Richard Pendragon
had cozened one of the first princes of his age, I did not know
whether it was not the bounden duty of a caballero of Spain to
repair to King Louis and confess the fact.
All the rest of that day this problem afflicted me sorely. In these
circumstances my natural guide was the Count of Nullepart, who
was an older head and a wiser; and one who, to judge by his
conversation, was not unacquainted with the things that concern
man’s higher nature. But when I mentioned to him my perplexity, his
only reply was to break out into laughter.
Finally, in my concern, I spoke of this matter to its author. He, with
his court gravity still upon him, heard me out very patiently, and
made answer with great solemnity.
“Most noble marquis,” said he, “you must forgive the personal
opinion of a good man, of a chief ornament of a shining age; but I
do not think you would use these questions, marquis, had you a
nicer familiarity with courts. Believe me, marquis, it is not the rule in
such elevated places to observe that slavishness to the sober verities
which at once betrays the mind of the provincial. I ask you, noble