The 1st Workshop on Communication Architecture for Scalable Systems
CASS 2011
To be held in conjunction with the
International Parallel and Distributed Processing Symposium (IPDPS 2011)

Anchorage, Alaska, USA, 16 May 2011
and in association with a special issue of Elsevier's
Journal of Parallel and Distributed Processing (JPDC)

JPDC cover

Workshop Organization

Co-chairs

José Flich (UPV)
Scott Pakin (LANL)
Craig Stunkel (IBM Research)

Publicity chair

Gaspar Mora (Intel)

Program Committee

Dennis Abts (Google)
Ahmad Afsahi (Queen's U.)
Gheorghe Almasi (IBM Research)
Pavan Balaji (ANL)
Davide Bertozzi (U. Ferrara)
Taisuke Boku (U. Tsukuba)
Darius Buntinas (ANL)
Marcello Coppola (STM)
Natalie Enright (U. Toronto)
Holger Fröning (U. Heidelberg)
Ada Gavrilovska (GIT)
Torsten Hoefler (NCSA)
John Kim (KAIST)
Jesper Larsson Träff (U. Vienna)
Raymond Namyst (U. Bordeaux & INRIA)
José Luis Sánchez (UCLM)
Federico Silla (UPV)
Tor Skeie (Simula)
Gregory Thorson (SGI)
Keith Underwood (Intel)
Eitan Zahavi (Mellanox)

Steering Committee

D.K. Panda (OSU)
José Duato (UPV)

Theme

High-speed communication is critical to all parts of an HPC system. On-chip networks for emerging many-core processors; point-to-point interconnects, which have replaced the system bus for intra-node communication; and system-wide networks, which form the backbone of any large-scale parallel system, all contribute to the construction of the world’s fastest computers. Numerous research groups in academia, industry, and government are currently investigating the issues involved in improving the speed, reliability, power consumption, and other characteristics of communication subsystems and seeking new ways to advance the state of the art in cluster communication.

The goal of this workshop is to bring together researchers working on improving communication at every level of the network hierarchy (on-chip, intra-node, and cross-cluster), thereby enabling the sharing and adaptation of ideas from what have traditionally been separate communities.

All researchers and practitioners working in the area of communication architectures for scalable systems are encouraged to submit a paper to the workshop.

Workshop Program

7:55 -

8:00

Welcome and workshop introduction

8:00 -

9:00

Keynote address

Speaker:	Matt Reilly, Institute for Defense Analyses/Center for Computing Sciences
Title:	Populating the Fabric Design Space

9:00 -

10:00

Session I

Topic:	Routing Performance
Chair:	Tor Skeie, Simula Research Laboratory
Improving Routing Performance in InfiniBand Networks with 2-Level Generalized Fat-tree Topologies Wickus Nienaber, Santosh Mahapatra, Xin Yuan A Unified Algorithm for both Randomized Deterministic and Adaptive Routing in Torus Networks Keith D. Underwood, Eric Borch

10:00 -

10:30

Break

10:30 -

12:00

Session II

Topic:	Communication Protocols
Chair:	Jesper Larsson Träff, University of Vienna
Profile Guided MPI Protocol Selection for Point-to-Point Communication Calls Aniruddha Marathe, David Lowenthal, Zheng Gu, Matthew Small, Xin Yuan Dynamic Time-Variant Connection Management for PGAS Models on InfiniBand Abhinav Vishnu, Manojkumar Krishnan, Pavan Balaji Audit: New Synchronization for the GET/PUT Protocol Atsushi Hori, Jinpil Lee, Mitsuhisa Sato

12:00 -

2:00

Lunch (on your own)

2:00 -

3:00

Session III

Topic:	Contention Reduction
Chair:	Pavan Balaji, Argonne National Laboratory
Efficient and Contention-Free Virtualisation of Fat-Trees Frank Olaf Sem-Jacobsen, Aashild G Solheim, Olav Lysne, Tor Skeie, Thomas J Sødring Fat-Trees Routing and Node Ordering Providing Non-Blocking MPI Global Collectives for Exascale Jobs Eitan Zahavi

3:00 -

3:30

Break

3:30 -

5:00

Session IV

Topic:	Collective Communication
Chair:	Torsten Hoefler, National Center for Supercomputing Applications
Optimizing MPI Collectives using Efficient Intra-node Communication Techniques over the Blue Gene/P Supercomputer Amith Mamidala, Ahmad Faraj, Sameer Kumar, Douglas Miller, Michael Blocksome, Thomas Gooding, Philip Heidelberger, Gabor Dozsa ConnectX-2 CORE-Direct Enabled Asynchronous Broadcast Collective Communications Manjunath Gorentla Venkata, Richard Graham, Joshua Ladd, Pavel Shamis, Ishai Rabinovitz, Vasily Filipov, Gilad Shainer Dodging Non-Uniform I/O Access in Hierarchical Collective Operations for Multicore Clusters Brice Goglin, Stéphanie Moreaud

5:00 -

5:05

Wrap up

Note 1: The CASS 2011 program page on EDAS provides an alternate formatting of the workshop program (and includes talk abstracts).

Note 2: Hyperlinked paper titles point to the slides presented at CASS 2011. The CASS 2011 presentation directory includes all of these in one place and in all submitted formats.

Registration and Hotel Information

Workshop registration is handled by the IPDPS 2011 conference. There is a single registration for the conference and all of its 18 workshops. Please visit the IPDPS 2011 Web page for registration and hotel information.

Additional Information

E-mail Contact

For more information on CASS 2011 or if you have any questions please contact the workshop organizers at the e-mail address listed at the bottom of this page.

Call for Papers

Although the submission deadline has long since passed, the following links will take you to the CASS 2011 CFP:

Links to CAC workshops (CASS's predecessor)