|
Ongoing Projects
|
The Innovative Computing Laboratory is involved in a number of projects related to high-performance computing: - Distributed Network Computing - SInRG: The Scalable Intracampus Research Grid (SInRG) project will deploy a research infrastructure on the University of Tennessee, Knoxville campus that mirrors the underlying technologies and the interdisciplinary research collaborations that are characteristic of the emerging national technology grid. SInRG's primary purpose is to provide a technological and organizational microcosm in which key research challenges underlying grid-based computing can be attacked with better communication and control than wide-area environments usually permit. The project is supported by a grant from the National Science Foundation. NetSolve: Network based computational server that allows users to access computational resources, such as hardware and software, distributed across the network. IJHPCA : The International Journal of High Performance Computing Applications provides a lively forum for the communications of original research papers and timely review articles on the use of supercomputers to solve complex modeling problems in a spectrum of disciplines. The emphasis will be on experiences with the use of supercomputers rather than on the exposition of computational results peculiar to a specific research topic. Software techniques that apply to classes of problems often cross disciplines; articles should focus on the exchange of such techniques, as well as present methods for analyzing, measuring and applying algorithms and solution schemes related to particular application areas. FT-MPI: Fault Tolerant MPI (FT-MPI) is a full 1.2 MPI specification implementation that provides process level fault tolerance at the MPI API level. FT-MPI is built upon the fault tolerant HARNESS runtime system. Open-MPI: GrADS: The goal of the Grid Application Development Software (GrADS) Project is to simplify distributed heterogeneous computing in the same way that the World Wide Web simplifed information sharing over the Internet. The GrADS project will explore the scientific and technical problems that must be solved to make grid application development and performance tuning for real applications an everyday practice. The University of Tennessee's efforts in the GrADS (Grid Application Deployment Software) projects are mainly related to the seamless deployment of numerical libraries over the grid. |
||
|
- Numerical
Linear Algebra - top
ScaLAPACK: A software library for performing dense and band linear algebra computations on distributed-memory message-passing MIMD computers and networks of workstations supporting PVM and/or MPI. Click here to mail the authors. Linear Algebra Software Survey: Information on freely available software for the solution of linear algebra problems. The interest is in software for high-performance computers that's available in source form on the web for solving problems in numerical linear algebra, specifically dense, sparse direct and iterative systems and sparse iterative eigenvalue problems. Self Adapting Numerical Algorithms (SANS): Self Adaptive Numerical Software (SANS) is a collaborative effort between different projects that deal with the optimization of software at different levels in relation to the execution environment. The SANS effort is intended to identify the commonalities of these different projects and help build a common framework on which these projects can possibly coexist. SALSA: SALSA is a software project that aims to assist applications in finding suitable linear and nonlinear system solvers based on analysis of the application-generated data. The SALSA system features heuristic decision making, based on a database of performance results that can tune the heuristics over time. Generic Code Optimization (GCO): Linear Algebra Fault Tolerant Algorithms (LAFTA) |
|||
|
- Software
Repositories - top
RIB (Repository In a Box): a toolkit for creating software repositories that can interoperate with other software repositories via the Internet. NetBuild: NetBuild is a suite of tools designed to aid users in making use of computational software libraries that are stored on the network, without needing to have those libraries preinstalled on each user's computer. Instead, NetBuild will determine which libraries are not installed, identify suitable versions of those libraries that are accessible from the network, download those libraries, and link them into the user's program. Active Netlib: As computationally intensive modeling and simulation become staples of scientific life across every domain and discipline, the difficulty of acquiring and sustaining the necessary expertise in scientific computing is becoming increasingly acute for the broad rank and file of students and professionals. While access to necessary computing and information technology has improved dramatically over the past decade, the efficient application of scientific computing techniques still requires specialized knowledge of numerical methods and their implementation in mathematical software libraries that many students, scientists and engineers, working beyond the already strenuous demands of their particular field, must struggle to achieve. Active-Netlib addresses this problem by creating an active collection of mathematical software for science and engineering education. The purpose of this collection is to provide the kind of rich, highly interactive, inquiry-based learning environment needed to enable students and application scientists to attain the confident mastery of numerical methods and software libraries their work in this new era requires. |
|||
|
- Performance
Evaluation and Benchmarking - top
Linpack Benchmark: How to get the report Click here for the postscript version of the report. Linpack Benchmark Frequently Asked Questions PAPI: PAPI aims to provide the tool designer and application engineer with a consistent interface and methodology for use of the performance counter hardware found in most major microprocessors. PAPI enables software engineers to see, in near real time, the relation between software performance and processor events. HPC Challenge: The HPC Challenge benchmark consists of basically seven benchmarks for measuring the performance of parallel systems. - Other
Projects - top
DOE SciDAC CScADS: The Center for Scalable Application Development Software (CScADS) was established in January 2007 as a partnership between Rice University, Argonne National Laboratory (ANL), the University of California at Berkeley, the University of Tennessee at Knoxville (UTK), and the University of Wisconsin at Madison. The mission of the center is to catalyze the development of software tools and libraries that enable applications to achieve scalable performance on the DOE's leadership computing platforms and increase the productivity of DOE computational scientists. DOE SciDAC PERI: The goal of this SciDAC Center is to develop a science for understanding performance of scientific applications on high-end computer systems, and develop engineering strategies for improving performance on these systems. The project is integrating several active efforts in the high performance computing community and is forging alliances with application scientists working on Department of Energy's Office of Science missions to ensure that the resulting techniques and tools are truly useful to end users. DOE SciDAC TOPS: TOPS is an integrated software infrastructure center (ISIC) focusing on developing, implementing, and supporting optimal or near optimal schemes for partial differential equation-based simulations and closely related tasks, including optimization of PDE-constrained systems, sensitivity analysis, eigenanalysis, and adaptive time integration, as well as core implicit linear and nonlinear solvers. TOPS is researching and developing and will deploy a toolkit of open source solvers for the nonlinear partial differential equations that arise in many application areas, including fusion, accelerator design, global climate change, and the collapse of supernovae. DOE SciDAC TSI: The Terascale Supernova Initiative is a multidisciplinary collaboration of one national lab and eight universities to develop models for core collapse supernovae and enabling technologies in radiation transport, radiation hydrodynamics, nuclear structure, linear systems and eigenvalue solution, and collaborative visualization. DOE Nano: Predicting the Electronic Properties of 3D, Million-Atom Semiconductor Nanostructure Architectures. Cyberinfrastructure Technology
Watch (CTW): CTW
serves two missions -- as a forum for opinion and analysis on issues related
to cyberinfrastructure and technology, and as a newsletter for cyber technology.
Per the first mission, the goal of CTW is to present thoughtful
opinion and analysis on engineering research, education, and practice;
science and technology policy; and the roles of engineering and technology
as they relate to cyberinfrastructure. |
|||
|
Older Projects |
- Past Projects - top Automatically
Tuned Linear Algebra Software (ATLAS): ATLAS is an approach for
the automatic generation and optimization of numerical software for processors
with deep memory hierarchies and pipelined functional units. The production
of such software for machines ranging from desktop workstations to embedded
processors can be a tedious and time consuming task. ATLAS has been designed
to automate much of this process. We concentrate our efforts on the widely
used linear algebra kernels called the Basic Linear Algebra Subroutines
(BLAS). BenchWeb: BenchWeb
is a starting point for finding information about computer system performance
benchmarks, benchmark results, and benchmark code. BLACS:
BLACS: a linear algebra oriented message passing interface for distributed
memory platforms. BLAS: machine constants, vector and
matrix * vector BLAS BLAS Technical Forum:
A forum to consider expanding the BLAS in a number of ways in light of
modern software, language, and hardware developments. CRPC: The Center
for Research on Parallel Computation, A National Science Foundation Science
& Technology Center. HPC Netlib: A
High Performance Branch of Netlib. HeNCE:
Heterogeneous Network Computing Environment. I2-DSI: The
Internet2 Distributed Storage Infrastructure Project The goal
of I2-DSI is to build storage resources into the network so as to support
a next generation services platform for universities in the next century.
I2-DSI is designing and deploying a reliable, scalable, high-performance
storage services infrastructure that will enable universities to more
fully exploit the power of Internet2's advanced network infrastructure,
vBNS and Abelene. IML++:
IML++ (Iterative Methods Library) is a C++ templated library of modern
iterative methods for solving both symmetric and nonsymmetric linear systems
of equations. Linpack
Benchmark in Java This applet runs the Linpack Benchmark
in Java. The Linpack Benchmark is a numerically intensive test that has
been used for years to measure the floating point performance of computers. Matrix Market:
The Matrix Market provides convenient access to a repository of matrix
test data for use in comparative studies of algorithms. MetaCenter Regional
Alliance at UTK: In 1996 UTK, under the direction of the Joint
Institute for Computational Science, was awarded a National Science Foundation
grant to establish a ``MetaCenter Regional Alliance (MRA) for Computational
Science Collaborations with Historically Black Colleges and Universities.''
MPI: message
passing interface. NHSE: An interface to a
distributed collection of HPCC repositories. Parkbench:
parallel benchmark working group. PDS: The Performance Database Server. PTLib: Parallel Tools
Library, click here to
mail the maintainers. PVM3:
parallel virtual machine version 3, click here to mail the authors. PVMPI:
A software system that allows independent MPI applications to inter-communicate
even if they are running under different MPI implementations using different
language bindings. RCDS (Resource
Cataloging and Distribution Service): an architecture for cataloging
the characteristics of Internet-accessible resources, for replicating
such resources to improve their accessibility, and for cataloging the
current locations of the resources so replicated. Scalable Networked
Information Processing Environment (SNIPE): SNIPE is a metacomputing
system that aims to provide a reliable, secure, fault-tolerant environment
for long-term distributed computing applications and data stores across
the global InterNet. This system combines global naming and replication
of both processing and data to support large scale information processing
applications leading to better availability and reliability than currently
available with typical cluster computing and/or distributed computer environments.
To facilitate this the system supports: distributed data collection, distributed
computation, distributed control and resource management, distributed
output and process migration. The underlying system supports multiple
communication paths, media and routing methods to aid performance and
robustness across both local and global networks. Sparse Matrix
Benchmark : SparseBench is a benchmark suite of iterative methods
on sparse data. Sparse matrices, such as derived from PDEs, form an important
problem area in numerical analysis. Unlike in the case of dense matrices,
handling them does not entail much reuse of data. Thus, algorithms for
sparse matrices will be more bound by memory-speed than by processor-speed.
Templates
Report: Templates for the Solution of Linear Systems, Building
Blocks for Iterative Method, click here
to mail the authors. TORC:
The TORC project is a collaborative effort between the University of Tennessee's
Innovative Computing Laboratory in the Computer Science Department and
the Mathematical Science Section of Oak Ridge National Laboratories. TORC's
goal is to provide a open, extensible platform for research. It is comprised
primarily of commodity hardware and software. TORC benefits from a variety
of interconnection media. TORC-1 is located at the University of Tennessee
and TORC-2 is located at Oak Ridge National Laboratory. Current work is
focusing on aggregations of the two groups into a common resource. |