NAG Logo
Numerical Algorithms Group

NAG Ping - Issue 11 (March 2003)

This issue covers:


Questions? Comments? Need more information about an article? Write to us at ping@nag.com.


Tips & Hints - Get a fast start with the NAG Libraries using example programs

Not sure how to get started with a NAG library routine? We've included some code to make a fast start. Each of our library products includes example programs and data for every routine. Along with NAG's detailed documentation, the example programs can be useful program templates for you to edit and help you get started.

A case in point: you want to use NAG's nag_opt_nlp (e04ucc), an optimization routine for a non-linear objective function with linear and non-linear constraints contained in NAG's C Library. To see the source code for the example program, go to http://www.nag.com/numeric/cl/manual/examples/source/e04ucce.c and to http://www.nag.com/numeric/cl/manual/examples/data/e04ucce.d to see sample data. Also included is a script file (Nagex.bat for Windows versions) that will compile and run any of the example programs to allow you to verify them on your system. Similar instructions apply to the Fortran 77 Library and the Fortran 90 Library. Note that the Fortran 77 Library example programs can be used in conjunction with the SMP Library.


AMD Opteron & Athlon 64 bit processors to get major speed boost from NAG

NAG and AMD are working together to produce a core numerical library optimized for existing and upcoming AMD processors including the AMD Opterontm and AMD Athlontm 64 processors.

The AMD Core Math Library (ACML) will initially incorporate BLAS, LAPACK and FFT routines, which can be used by a wide range of software developers to obtain excellent performance from their applications running on AMD platforms.

ACML is expected to be available to developers on the AMD website during the second quarter of 2003. NAG and AMD will continue to maintain and develop ACML so software developers will be able to rely on the proven accuracy, robustness and reliability of the library to give them the correct answers - quickly and efficiently.


Technical Computing on MacOS X - NAG announces a suite of products

Apple is becoming the platform of choice for many involved in scientific and technical computing. The combination of MacOS X, with its fast and stable Unix kernel, and Apple's long established ease of use has re-established the Mac as a platform for "serious" science. NAG began collaboration with Apple prior to the introduction of MacOS X that resulted in the introduction of the NAGWare f95 Compiler in August of 2002. This compiler was enthusiastically received by the community for applications such as modeling ocean currents and simulating the evolution of stars. Among many comments we get, one sums up the reaction of users - "It is by far the friendliest compiler I have used."

With the success of the compiler NAG has pushed ahead with the creation of numerical libraries for the MacOS platform. We are pleased to announce three new products: the NAG C Library, the NAG Fortran Library, and the NAG Fortran 90 Library. The NAG C Library is available for download and trial at http://www.nag.com/downloads/index_arch.asp#AMP. The other two are expected to be available within a few weeks.


Tips & Hints - The NAG SMP Library, OpenMP directives, and Pseudo-random number generation

The NAG SMP Library is the high performance version of NAG's traditional Fortran 77 Library. It offers an identical calling interface to the F77 Library but provides a high degree of scaling for 4 or more processor systems. It achieves this performance through 231 highly optimized routines using OpenMP directives. If you have a performance-sensitive application using the NAG Fortran 77 Library, you can generally boost its performance on a multi-processor system simply by re-linking (not re-compiling) it to the NAG SMP Library. As you use the SMP Library, here are a couple of tips to keep in mind.

Psuedo-Random Numbers

When a piece of code generates a large number of pseudo-random numbers, these are best generated in long arrays. This reduces the overhead associated with restarting on each call to the generator routines (G05 Chapter).

SMP/OpenMP Parallel Regions

If your code contains OpenMP directives to create parallel regions, please remember that this will switch off any parallelism in NAG routines called within the parallel region since multi-level SMP parallelism is not (yet) supported by any vendor.

For more information about the NAG SMP Library please click here: http://www.nag.com/numeric/fl/FSdescription.asp or email infodesk@nag.com with any specific inquiries.


NAG welcomes new ISV Software Partners SoftIntegration, Gylling Data Management

As the users of many enterprise and desktop applications have begun to demand sophisticated analytical capabilities, Independent Software Vendor firms (ISVs) have turned increasing to NAG for robust, fast, accurate, and thoroughly documented math, statistical, and data mining capabilities. These include such well-known firms as PeopleSoft, IBM Data Solutions, and Manugistics as well as many others. They trust NAG software to give their customers the accuracy, speed and performance NAG customers have enjoyed for over 30 years. Our Software Partners also rely on NAG in two other areas - available components that shave months and hundreds of labor hours from product development schedules and expert guidance and technical support on the use of these components.

In this issue, we would like to welcome two new NAG Software Partners - Gylling Data Management and SoftIntegration, Inc.

Gylling Data Management (GDM) recently introduced Agricultural Research Manager 7 software for establishing, managing, and summarizing of herbicide, insecticide, fungicide, seed/variety, fertilizer, general non-chemical, nematicide, and plant growth regulator research trials. GDM enhanced its product with the incorporation of sophisticated experimental design components from NAG. Learn more at http://www.gdmdata.com.

SoftIntegration makes Ch, an embeddable C/C++ interpreter for cross-platform scripting, shell programming, 2D/3D plotting and numerical computing. Ch allows users to use one language, anywhere and everywhere, for any programming tasks on platforms including Windows, Linux, MacOS X, and Unix. SoftIntegration plans to incorporate multiple NAG statistical functions into an upcoming release of their software. Learn more at http://www.softintegration.com.


Origins of NAG - Part 1

Every organization has a history that predicts its future. The NAG software you use today and the service you receive from us is shaped by our thirty plus year history and by the goals and practices of our founders, many of whom are with us today. NAG began as the Nottingham Algorithms Group in 1970, a collaborative project among the Universities of Birmingham, Leeds, Manchester, Nottingham and Oxford, and the Atlas Computer Laboratory. The aim was to develop a library of numerical and statistical subroutines for use on their ICL 1906A/S machines. Because of the different language emphasis in the centers and the difficulty of mixed-language programming, it was decided to develop the library in both Algol 60 and ANSI Fortran. The Mark 1 Library contained 98 user-callable routines and was released on October 1, 1971.

Universities with other types of computers became interested in the activity and various machine-range implementations were initiated beginning with Mark 2 of the Library. Thus began growth of an organization dedicated to the following goals:

  • Create a balanced, general-purpose library of algorithms meeting the numerical and statistical needs of computer users;
  • Thoroughly document the library, giving advice on problem identification, algorithm selection, and routine usage;
  • Provide a substantial test suite, including example test programs, for certification of the library;
  • Implement the Library as widely as user demand requires.

NAG committed itself to a long-term program of library contents development with a strong emphasis on documentation, testing and portability. This emphasis has been maintained throughout the decades and is reflected in the present-day range of NAG products. Algorithm contributions and implementations were coordinated from Nottingham until August 1973 when the project offices were moved to Oxford University and the name of the project was changed from "The Nottingham Algorithms Group" to "The Numerical Algorithms Group".


In the Next NAG Ping

-- NAG and Waterloo Maple Announce Next Phase in Partnership
-- OR/MS Today, Financial Engineering News reviews NAG Data Mining Components
-- Intel Developer Tools & NAG Libraries - New products available
-- Origins of NAG - Part 2


Your Privacy - Our Policy

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.


Subscribe/Unsubscribe/Manage Your Account

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
You are free to forward this newsletter to colleagues or to post it to your company's Intranet site.
If your colleagues would like to receive this newsletter, they can either register online or contact us at ping@nag.com.

To manage your account with NAG (name, e-mail address, interests, etc.)
go to http://www.nag.com/local/userreg/users and login with your user name and password.

If you no longer wish to receive The NAG Ping, please reply with the word 'unsubscribe' in the subject line.


About The NAG Ping

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 / Copyright information

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

Copyright 2001-2004 Numerical Algorithms Group, Inc.

© The Numerical Algorithms Group 2008
Privacy Policy | Trademarks

© Numerical Algorithms Group

Visit NAG on the web at:

www.nag.co.uk (Europe and ROW)
www.nag.com (North America)
www.nag-j.co.jp (Japan)

http://www.nag.com/nagnews/nagping/np011.asp