Swann Perarnau

Assistant Computer Scientist
Mathematics and Computer Science Division
Argonne National Laboratory
9700 South Cass Avenue
Argonne, IL, 60439
USA

Email: swann at anl dot gov
Phone: (+1) 630-252-1676

I am an assistant computer scientist at at Argonne National Laboratory since September 2016. I was a postdoctoral appointee at Argonne for two year before, and a JSPS postdoctoral fellow at the RIKEN AICS laboratory, home of the K Computer, for 2.5 years before that. I completed my Ph.D in Computer Science in December 2011, under the supervision of Denis Trystram and Guillaume Huard, in the MOAIS research team of INRIA/LIG, Grenoble, France.

Ph.D Thesis (in French):
Environnements pour l’analyse expérimentale d’applications de calcul haute performance [ pdf ]

Research Activity

I am currently part of the Argo Project to build a Operating System and Runtime stack for Exascale machines. My work involves building introspective resource management across the entire stack, from controlling node resources (memory, cores) using containers to monitoring global resources (power, nodes) across the system.

Other Research Interests

High Performance Computing, parallel programming models and runtimes, many-core architectures, low-level performance analysis and optimizations, experimentation as a scientific activity and research domain of computer science.

Publications, Presentations

International, Peer-reviewed, Conferences/Workshops

Exploring Data Migration for Future Deep-Memory Many-Core Systems
Swann Perarnau, Judicael A. Zounmevo, Balazs Gerofi, Kamil Iskra, Pete Beckman
In IEEE Cluster Conference, 2016.
[ paper (pdf) | bib ]

Reducing Waste in Large Scale Systems Through Introspective Analysis
Leonardo Bautista-Gomez, Ana Gainaru, Swann Perarnau, Devesh Tiwari, Saurabh Gupta, Franck Cappello, Christian Engelmann, Marc Snir
In IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2016.
[ paper (pdf) | bib ]

A Container-Based Approach to OS Specialization for Exascale Computing
Judicael A. Zounmevo, Swann Perarnau, Kamil Iskra, Kazutomo Yoshii, Roberto Gioiosa, Brian C. Van Essen, Maya B. Gokhale, Edgar A. Leon
In International Workshop on Container Technologies and Container Clouds (WoC), 2015.
[ paper (pdf) | bib ]

Victim Selection and Distributed Work Stealing Performance: A Case Study
Swann Perarnau and Mitsuhisa Sato
In IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2014.
[ paper (pdf) | bib ]

Discovering Sector Cache Optimizations on the K Computer
Swann Perarnau and Mitsuhisa Sato
In Asia-Pacific Programming Languages and Compilers Workshop (APPLC), 2013.
[ paper (pdf) | bib | slides (pdf) ]

Controlling Cache Utilization of HPC Applications
Swann Perarnau, Marc Tchiboukdjian, and Guillaume Huard
In International Conference on Supercomputing (ICS), 2011.
[ paper (pdf) | bib ]

Krash: Reproducible CPU Load Generation on Many-core Machines
Swann Perarnau and Guillaume Huard
In IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2010.
[ paper (pdf) | bib | slides (pdf)

Random graph generation for scheduling simulations
Daniel Cordeiro, Grégory Mounié, Swann Perarnau, Denis Trystram, Jean-Marc Vincent, and Frédéric Wagner.
In International ICST Conference on Simulation Tools and Techniques (SIMUTools), 2010.
[ paper (pdf) | bib | slides (pdf) ]

Quantitative separation logic and programs with lists
Marius Bozga, Radu Iosif, and Swann Perarnau
In International Joint Conference on Automated Reasoning (IJCAR), 2008
[ bib ]

Posters, Short Papers in English

Systemwide Power Management with Argo
Dan Ellsworth, Tapasya Patki, Swann Perarnau, Pete Beckman et al
In High-Performance, Power-Aware Computing (HPPAC), 2016.
[ paper (pdf) | bib ]

Argo: An Exascale Operating System and Runtime
Swann Perarnau, Rinku Gupta, Pete Beckman et al
In International Conference for High Performance Computing, Networking, Storage and Analysis (SC), 2015.
[ paper (pdf) | bib | poster (pdf) ]

Distributed Monitoring and Management of Exascale Systems in the Argo Project
Swann Perarnau, Rajeev Thakur, Kamil Iskra, Ken Raffenetti, Franck Cappello, Rinku Gupta, Pete Beckman, Marc Snir, Henry Hoffmann, Martin Schulz, Barry Rountree
In IFIP International Conference on Distributed Applications and Interoperable Systems (DAIS), 2015.
[ paper (pdf) | bib | slides (pdf) ]

Weighted Distribution for Random Victim Selection in Distributed Work Stealing
Swann Perarnau and Mitsuhisa Sato
In HPC in Asia Workshop, International Supercomputing Conference (ISC), 2013.
[ poster (pdf) | bib ]

Sector Cache Optimizations for the K Computer
Swann Perarnau and Mitsuhisa Sato
In AICS International Symposium on Computer and Computational Sciences for Exascale Computing, 2013.
[ poster (pdf) ]

Krash: Reproducible CPU Load Generation on Many-Core Machines
Swann Perarnau and Guillaume Huard
In ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPOPP), Poster Session/Short Paper, 2010.
[ paper (pdf) | bib | poster (pdf)

Japanese Domestic Communications

Toward Automated Cache Partitioning for the K Computer
Swann Perarnau and Mitsuhisa Sato
IPSJ HPC SIG Technical Report/Workshop #136, 2012
[ paper (pdf) | bib | slides (pdf)

International Journals

Quantitative separation logic and programs with lists
Marius Bozga, Radu Iosif, and Swann Perarnau
Journal of Automated Reasoning, 45(2):131-156, 2010.
[ bib ]

Publication in French, Peer-Reviewed Conferences

Sensibilité des algorithmes d'ordonnancement
Jean-Noël Quintin and Swann Perarnau
In French RENPAR Conference, 2011
[ paper (pdf) | bib | slides (pdf) ]

Posters, Short Papers in French

GGen: Génération aléatoire de graphes pour l'ordonnancement
Swann Perarnau, Denis Trystram, and Jean-Marc Vincent
In French ROADEF Conference, Short Paper, 2011
[ paper (pdf) | bib | slides (pdf) ]

Comment rater la validation de votre algorithme d'ordonnancement
Daniel Cordeiro, Grégory Mounié, Swann Perarnau, Denis Trystram, Jean-Marc Vincent, and Frédéric Wagner
In French RENPAR Conference, Poster Session/Short Paper, 2009
[ paper (pdf) | bib | poster (pdf) ]

Side Projects

I have designed and implemented several tools used to analyze parallel applications experimentally, including:

KRASH: a CPU load generator that was the most reproducible and precise tool of its kind during several years, until the CPU bandwidth limitation feature of the Linux kernel became available.

CControl: a kernel module and its associated library to limit the cache of applications using page coloring. Can also be used to redistribute the cache among structures inside an application.

GGen: a collection of algorithms to generate directed acyclic graphs for the simulation of schedulers.

You can find some of my projects in my github repository.