Past Groups
- AMD: We are
collaborating with AMD
Research to investigate how Chapel can take advantage of
the Heterogeneous
System Architecture (HSA).
- Argonne
National Laboratory: We are working with Pavan Balaji, Brian
Skjerven, and Wesley Bland to explore the possibility of mapping
Chapel's runtime communication interface to MPI-3's features for
one-sided communication.
- BSC
/ UPC: We were
pursuing a collaboration with Alejandro Duran of
the Nanos++
Project to implement Chapel's tasking layer using Nanos++
user-level tasks.
- Carnegie Mellon
University: We have worked with Franz Franchetti and Thom
Popovici to study native support for FFT computations in
Chapel and the suitability of Chapel as a portable
parallel back-end language for the SPIRAL program generation system.
- Colorado
State University: We have collaborated with Michelle Strout, Andy Stone, and Christopher Krieger to
look at parallel programming patterns and support for tilings in
Chapel.
- EPCC
/ U. Edinburgh: We
supported a study undertaken by Michele
Weiland and Thom Haddow to evaluate the performance of Chapel's
task-parallel features on a single locale.
- ETH
Zurich: Andreas
Fichtner and Alexey Gokhberg have explored the possiblity of
using Chapel for programming seismic wave propagation
applications.
- The George
Washington University: Tarek
El-Ghazawi, Engin
Kayraklioglu, Olivier
Serres, Ahmad
Anbar, and others are studying topics such as hardware support
for Chapel array accesses and system support for optimizing the
placement of Chapel's tasks and locales on a target system
architecture by exploiting hierarchical locality.
- Heriot-Watt
University: Konstantina Panagiotopoulou
is working on resilience in Chapel to transparently handle hard
failures in hardware via forward error recovery with task
adoption.
- Indiana
University: We are exploring extensions to Chapel's
object-oriented features to support better interfaces (a la C++
concepts) with Jeremy
Siek and Chris Wailes.
- Individual Researcher: Brian Guarraci has
been exploring how to use Chapel for ARM processors, search, and Natural
Language Processing applications such as a distributed word2vec.
- LLNL: We work
with Jeff
Keasler, Bert
Still, Rob
Neely, Riyaz
Haque, and David Richards to study
proxy applications in Chapel, such as
LULESH, CoMD and LCALS.
- LLNL/Rice University: We are
collaborating with Tom Epperly, and Shams
Imam on extending the Babel
system for HPC language interoperability to support Chapel.
- ORNL: We are
working with David Bernholdt and
Tiffany Mintz, who are exploring the notion of
"programmer-guided reliability" in Chapel as a means of
detecting and correcting silent data corruption in applications.
- ORNL: We work
with a number of scientists in the Computer Science
Research, Extreme
Scale Systems Center, and
Future Technologies
groups to evaluate Chapel for applications of interest such
as MADNESS.
-
ORNL /
Notre Dame:
We were involved in a study with
Srinivas Sridharan,
Jeffrey Vetter, and
Peter Kogge
to look at asynchronous software transactional memory (STM) on
distributed memory architectures in the context of Chapel.
- OSU: We
worked with Gagan Agrawal and
Bin Ren, to
look at using Chapel's user-defined reductions to target
data-intensive computations using their FREERIDE technology as a
back-end. This resulted in a paper at HIPS 2011 and a poster at
LCPC 2010.
- Sandia National
Laboratories: We are collaborating with Sandia to
implement Chapel's tasking layer and synchronization variables
to hardware in an efficient manner using QThreads' user-level
tasks and implementation of full-empty bit variables.
- Rice
University:
Vivek Sarkar, Akihiro
Hayashi, and Jisheng Zhao are investigating how LLVM-based
optimizations for PGAS programs can improve the performance of
Chapel programs.
- Rice
University:
Vivek
Sarkar, Sagnak Tasirlar, Shams Imam, and Philippe Charles
are looking into adding support for futures to Chapel as well
as general improvements to its task parallel features.
- University of
Delaware: Stephen Siegel,
Timothy Zirkel, and Timothy McClory are investigating the use
of model checking and symbolic execution to verify Chapel
programs.
- UIUC: We
worked with Vikram
Adve and Robert
Bocchino to study Software Transactional Memory (STM) concepts for
distributed memory architectures (see Software Transactional Memory
for Large-Scale Cluters from PPoPP'08 for our initial
results: paper, slides).
- UIUC: We are
working with David
Padua, Maria
Garzaran, and Albert Sidelnik to
investigate the use of Chapel to program GPU accelerators.
- University
of Malaga: We are working with Rafael
Asenjo, Maria
Angeles Navarro, Rafael Larrosa Jimenez, and Alberto Sanz on
parallel I/O, bulk-copy optimizations and redistribution, and
dynamic self-scheduling iterators.
- University of
Maryland:
Jeff
Hollingsworth, Ray Chen, and Hui Zhang are investigating how
Chapel can include the
Active Harmony auto-tuning library and also are looking into
data-centric performance analysis.
- University of
Tokyo, Taura
Lab: We have collaborated with Kenjiro Taura
and Jun
Nakashima on lightweight threading via MassiveThreads
and performance evaluation of Chapel idioms.
- University of
Oregon / Paratools Inc.:
With Sameer Shende
and Allen Malony, we
are exploring the possibility of using Tau to analyze the performance
of Chapel programs.
- UT Austin:
With Calvin Lin and
Karthik
Murthy we explored possible memory consistency models for use
in Chapel.
- Western
Washington University: Phil Nelson has been
exploring tools for comprehending the execution and performance of
Chapel programs, currently included in the release as chplvis.
