The ever-increasing data and processing requirements of applications
from various domains are constantly pushing for dramatic increases in
computational and storage capabilities. Today, we have
reached a point where computer systems’ growth cannot be addressed
anymore in an incremental way, due to the huge challenges lying ahead, in
particular scalability, energy barrier, data management, programmability,
Ultrascale computing systems (UCS) are envisioned as a
large-scale complex system joining parallel and distributed
computing systems, maybe located
at multiple sites, that cooperate to provide solutions to the
users. As a growth of two or three orders of magnitude of today’s
computing systems is expected, including systems with unprecedented
amounts of heterogeneous hardware, lines of source code, numbers
of users, and volumes of data, sustainability is critical to ensure the
feasibility of those systems. Due to those needs, currently there is an
emerging cross-domain interaction between high-performance in clouds or
the adoption of distributed programming paradigms, such as Map-Reduce,
in scientific applications, the cooperation between HPC and
distributed system communities still poses many challenges towards
building the ultrascale
systems of the future. Especially in unifying the services to deploy
sustainable applications portable to HPC systems, multi-clouds,
data centers, and big data.
TASUS workshop focuses on the software side, aiming at bringing
together researchers from academia and industry interested in the design,
implementation, and evaluation of services and system software
mechanisms to improve sustainability in ultrascale computing systems
with a holistic approach.
We are looking for original high quality research and position papers
on applications, services, and system software for sustainable ultrascale
systems. Topics of interest include:
- Existing and emerging designs to achieve sustainable ultrascale systems.
- High-level parallel programming tools and programmability
techniques to improve applications sustainability on ultrascale
platforms. (model driven, refactoring, dynamic code generation, unified
services, middlewares, …).
- Synergies among emerging programming models and run-times
from HPC, distributed systems, and big data communities to provide
sustainable execution models (increased productivity, transparency,
- New energy efficiency techniques for monitoring, analyzing, and
modeling ultrascale systems, including energy efficiency metrics for
multiple resources (computing, storage, networking) and sites.
- Eco-design of ultrascale components and applications, with
special emphasis on energy-aware software components that help
users to shape energy issues for their applications.
- Sustainable resilience and fault-tolerant mechanisms that can
cooperate throughout the whole software stack to handle errors.
- Fault tolerance techniques in partitioned global address space
(e.g. PGAS, MPI, hybrid) and federated cooperative environments.
- Data management optimization techniques through cross layer
adaptation of the I/O stack to provide global system information to
improve data locality.
- Enhanced data management lifecycle on scalable architectures
combining HPC and distributed computing (clouds and data centers).
- Experiences with applications, high-level algorithms, and services
amenable to ultrascale systems.
· Workshop papers due: May 30, 2014
· Workshop author notification: July 4, 2014
· Workshop early registration: July 25, 2014
· Workshop camera-ready papers due: October 3, 2014
Prof. Jesus Carretero. University Carlos III of Madrid. Spain.
Dr. Laurent Lefevre. INRIA, ENS of Lyon. France
Prof. Gudula Rünger. Technical University of Chemnitz. Germany.
Prof. Domenico Talia. Universitá della Callabria. Italy.