NAG Ping - Issue 10 (December 2002)

This issue covers:

• Advancing Research in Canada - NAG Partners with HPCVL
• Visualization Comes to the Grid - NAG Joins g-Viz Research Project
• NAG’s SMP Library - Now Optimized for the IBM Power4
• NAG Expands Presence on West Coast for ISV Software Partners
• How to Get Your Financial Data Really Clean
• Product Updates - C Library - Fortran Library
• Technical Tips - Using the NAG SMP Library on Heavily Loaded Systems
• White Papers - Now Available on NAG’s Web Site
• In the Next NAG Ping
• Your privacy - our policy
• Subscribe/Unsubscribe - Manage Your Account
• About The NAG Ping
• Contact us / Copyright information

The High Performance Computing Virtual Laboratory (HPCVL), a consortium of four leading Canadian universities, and NAG recently announced a partnership to bring NAG’s Parallel, SMP, and other numerical libraries to Canadian research projects in bioinformatics, computer-aided molecular design, econometrics, computational fluid dynamics, drug discovery, photonics and other disciplines. The addition of NAG libraries to HPCVL, one of the premier supercomputing centers in North America, is expected to help HPCVL-affiliated researchers accelerate their development of applications for solving complex problems.

HPCVL (www.hpcvl.org) has quickly grown to become the largest High Performance Computing Center in Canada, and is listed in the top 500 supercomputers in the world. HPCVL was formed by four universities located in Eastern Ontario (Carleton University, Queen’s University, The Royal Military College of Canada, and the University of Ottawa). HPCVL is dedicated to providing researchers at member institutions and researchers from elsewhere in Canada with the high performance computing resources they need to conduct innovative research in a broad range of disciplines. To read more about this partnership, please see our recent Press Release.

NAG has joined an effort to bring 3D visualization to high performance grid computing. We are a participant in g-Viz (short for "grid Visualization"), a research project funded under the UK's e-Science Programme to investigate and develop visualization middleware for grid computing. Professor Ken Brodlie of the University of Leeds heads the two-year project, begun in August 2002. In addition to Leeds, other participants include the Universities of Oxford and Oxford Brookes, Rutherford Appleton Laboratory, IBM, Streamline Computing, and NAG. Among the major areas of investigation are the development of a "grid-enabled" form of NAG’s IRIS Explorer, a similar exercise for the pV3 parallel visualization toolkit, a study of data and geometry compression techniques, and exploration of the use of XML for describing visualization datatypes and pipeline connections between visualization and related modules.
For further information about this exciting project, contact Professor Brodlie (kwb@comp.leeds.ac.uk) or Jeremy Walton of NAG Ltd (Jeremy.Walton@nag.co.uk). Find out more about the UK's e-Science Programme here.

NAG has optimized its SMP (Symmetric Multi-Processor) Library for the IBM POWER4 architecture.
Here’s a taste of the performance on the IBM eServer pSeries p690 system with 1.1 GHz POWER4 processors:

• Common dense linear algebra factorization routines (eg LU, Cholesky and QR), and their associated solvers, sustain a performance of up to 2.8 GFLOPS on a single processor, and scale in performance up to 21 GFLOPS on 8 processors.
• The QR-algorithm symmetric eigenproblem routine (F08JEF/DSTEQR) achieves 2.4 times the performance on 1 processor, and 16.6 times the performance on 8 processors, of the algorithmically equivalent LAPACK routine DSTEQR.
• The QR-algorithm Singular Value Decomposition (SVD) routine (F08MEF/DBDSQR) achieves over 6 times the performance on 1 processor of the algorithmically equivalent LAPACK routine. While the standard LAPACK code does not scale in performance on multiple processors, the SMP routine scales by, for example, a factor of 3.7 on 4 processors.

For detailed information about the above results and to read more about the NAG SMP Library for IBM Power4, click here.

Over the past several years, NAG has developed a reputation among independent software vendors (ISVs) as the preferred source for math, statistical, and data mining components. To better serve these partners on the US West Coast, NAG has recently added Greg Smith to its staff as Business Development Manager. Prior to NAG, Greg was in a similar role at Citeline, Inc., a provider of specialized search engine software, clinical trial databases, and editorial services to biotechnology and pharmaceutical companies worldwide. Prior to Citline, Greg was Director of Marketing at eManage.com, a provider of Internet managed services, and served as Business Development Manager for Apple Computer's developer relations division. Greg also was founder & CEO of RoundBook Publishing Group, Inc. where he produced RoundBook’s Computer Education Series with Prentice-Hall. BYTE magazine called the series, "a remarkable programming resource" - http://www.byte.com/art/9502/sec6/art3.htm.

Prior to RoundBook, Greg was general manager of Sony's first software company in the U.S., and he built the #1 worldwide CD-ROM pre-mastering system and networking sales force as VP of Sales at Meridian Data, Inc. in the late 1980s. Greg began his professional career as a software engineer in Silicon Valley.

Greg holds an undergraduate degree in Computer Technology from Cayuga College and pursued graduate studies in microbiology and food science at Cornell University. Most recently, Greg completed UC Santa Cruz Extension's award-winning Bioinformatics Certificate Program.

Increasingly, sophisticated analytical and data mining methods are available to decision makers for financial data. In practice, that data can be full of errors and missing values. Often, the more sophisticated methods seem to be particularly sensitive to the presence of bad values in the data. Dealing with bad data before the modeling takes place can improve the quality of the results. Geoff Morgan of NAG Ltd. recently discussed this topic in the June/July 2002 issue of Financial Engineering News (www.fenews.com). To read the compete article, see "Cleaning Financial Data" in NAG’s white papers.

NAG C Library

The recently launched Mark 7 of the NAG C Library is now available. NAG is committed to providing its numerical components for a wide range of platforms and will be adding new implementations on a regular basis. Mark 7 now offers double the functionality of Mark 6 and remains the largest and most comprehensive collection of numerical algorithms for C and C++ programmers in the world. The NAG C Library, Mark 7, is now available for the following platforms:

• Linux
• Silicon Graph IRIX 6
• Sun SPARC Solaris
• HP9000/700
• IBM RISC
• Microsoft Windows (Microsoft C/C++ compiler)
• Fujitsu PRIMEPOWER

For more information about the availability of NAG's C Library on other platforms, or for information about the functionality of NAG's C Library, please see www.nag.com/numeric/cl/CLdescription.asp. Remember, you can download a free, fully functional trial.

NAG Fortran Library

The NAG Fortran Library provides over 1,280 of NAG’s globally renowned user callable routines for a broad range of mathematical and statistical areas. NAG Fortran algorithms are readily integrated into applications written in a comprehensive range of languages. The Fortran Library, Mark 20, is now available for the following platforms:

• Microsoft Windows (CVF compiler)
• Digital Alpha UNIX
• Fujitsu PRIMEPOWER, VPP5000
• HP9000/700
• Hitachi SR8000
• HP PA-RISC 2.0
• IBM RS/6000 Power 3, RISC
• Linux
• Silicon Graph IRIX 5, IRIX 6
• Sun 64-bit Solaris, SPARC Solaris

For more information about the availability of NAG's Fortran Library on other platforms, or for information about the functionality of NAG's Fortran Library, see www.nag.com/numeric/fl/FLdescription.asp. Remember, you can download a free, fully functional trial.

A heavily loaded SMP system, where many programs are executing simultaneously, may not seem like an obvious candidate for using the NAG SMP Library. However, the SMP Library can offer important advantages even in this scenario. The NAG SMP Library contains many algorithms, particularly in the area of dense eigen problems and singular value decomposition, which have been optimized to obtain much better reuse of data in the fast cache memory on each processor than the NAG Fortran Library and standard LAPACK implementations. This not only greatly increases the performance of these routines (in some cases by a factor of more than 6 on a single processor), but also reduces the demands placed on the SMP computer's memory system, which can result in a large improvement in overall system throughput.

Read more about the NAG SMP Library at www.nag.com/numeric/fl/FSdescription.asp or contact us by email at infodesk@nag.com with any specific questions.

At NAG we frequently write articles on technical computing for various publications. The topics range from avoiding the pitfalls of computer arithmetic to data cleaning methods to using quasi-random number generators for financial modeling. In a recent fit of organization (akin to rearranging your closet) we created a "White Papers" repository on our web site. Make a visit here if you are interested in browsing through our collection.

-- Origins of NAG
-- Faster Transforms - Getting more performance from FFTs
-- NAG’s Data Mining Components Reviewed

At NAG, we have a simple policy on the use of information you provide to us; it is to treat you how we would like to be treated. We will not sell or give your personal information to anyone and will only use it for the purposes we state and with the permission you grant us. NAG also is one of the first U.S. companies to voluntarily certify compliance with the safe harbor principles for privacy protection developed by the U.S. Department of Commerce in conjunction with the European Union.
To learn more, click here.

You have received NAG Ping because you have:

1. Registered on the NAG Web site
2. Hold a current license to NAG Software
3. Given NAG permission to send information to you from a prior research survey
4. Requested NAG add you to their mailing list

The NAG Ping is an occasional newsletter of technical tips, product information and discussions of technical computing topics. Its principal audience is in North America though it has readers throughout the world. It is the sister publication of the NAGNews, produced by our colleagues at NAG Ltd primarily for those in Europe and elsewhere outside North America.

The NAG Ping is named in memory of Mike Muuss, formerly of the Army Research Laboratory. Among the creative and useful works of his life was the "ping" program to test network connection. It was written by Mike in 1983 to help diagnose problems on his network. The name is tied to Mike's earlier work in sonar modeling and echolocation. To learn more, click here.

Contact us at at ping@nag.com or http://nag.com/contact_us.asp#northamerica.

Copyright 2001-2004 Numerical Algorithms Group, Inc.