IEEE Computer Special issue on Irregular Applications
Full paper submission deadline: 1 February 2015
Publication date: August 2015
The broad class of irregular applications is characterized by unpredictable
memory access patterns, control structures, and/or network transfers. These
applications typically use pointers or linked list–based data structures
such as graphs, unbalanced trees, and unstructured grids. Their complex
behavior makes it difficult to fully exploit their significant latent
parallelism. In addition to performance concerns, dataset size presents a
challenge in emerging irregular applications because they often operate
on massive amounts of unstructured heterogeneous data that is
usually difficult to partition.
Current high-performance architectures rely on data locality as well as regular
computations, structured data, and easily partitionable datasets; consequently,
they do not cope well with the computational and data requirements of
irregular applications. Furthermore, scaling on current supercomputing machines
is problematic, because of limits associated with fine-grained communication and
synchronization. These applications exist in well established and emerging
fields such as: CAD; bioinformatics; semantic graph databases; machine
learning; analysis of social, transportation, communication, and other types of
networks; and computer security. Addressing the many system-related issues
posed by irregular applications on current and future system architectures
is critical to solving future scientific challenges.
This special issue seeks to explore solutions for supporting the efficient
design, development, and execution of irregular applications. Practical
and theoretical topics of interest include but are not limited to:
- Micro- and system-level architectures;
- Network and memory architectures;
- Many-core, hybrid, heterogeneous, and custom architectures
(tiled processors, GPUs, FPGAs);
- Modeling, evaluation, and characterization of architectures for
memory-intensive and irregular applications;
- Innovative algorithmic techniques;
- Combinatorial (graph) algorithms and their applications;
- Languages and programming models;
- Library and runtime support;
- Compiler and analysis techniques; and
- Case studies of irregular applications (for example,
semantic graph databases, data mining, security, bioinformatics).
Articles focused on approaches that span multiple levels of the stack
— ideally providing application-specific, end-to‐end solutions — are of
particular interest. Articles should provide context for their contributions
with respect to existing solutions as well as potential commercial impact.
Antonino Tumeo (email@example.com), PNNL
John Feo (firstname.lastname@example.org), PNNL
Paper submissions are due 1 February 2015.
Only technical articles describing previously unpublished, original,
state-of-the-art research, and not currently under review by a conference
or a journal will be considered. Articles should be understandable to a
broad audience of computer science and engineering professionals,
avoiding a focus on theory, mathematics, jargon, and abstract concepts.
All manuscripts are subject to peer review on both technical merit and
relevance to Computer’s readership. Accepted papers will be professionally
edited for content and style.
For author guidelines and information on how to submit a manuscript,