Issue 86, 21 January 2010

Featuring


January, resolutions & bold plans


To paraphrase Thomas Paine, “January is the time that tries men's souls.” It is, for those of us in the northern hemisphere, the season of little daylight, low temperatures, snow, ice and rain. It is a month masquerading as a personality test. The optimists among us use the occasion to resolve (once more) to make all of those improvements we failed to live up to last year. The pessimists, of course, know better. I count myself as a pragmatist, defined as an optimistic idealist tempered by reality. The notion of new years and resolutions in January is a bit of an arbitrary tradition in a world where time is sliced into calendar years, academic years and fiscal years, the gift of accountants and auditors.

I've been thinking about new years and resolutions a lot because NAG has a significant milestone coming up on May 13, 2010 when we will celebrate the 40th anniversary of the start of the project that became The Numerical Algorithms Group Ltd. It is a worthy time to assess our past and plan our future. Perhaps it is the same for you and your projects.

When you've been developing numerical software for mathematics and statistics for as long as we have it's easy to focus only on the worthy task of adding new functions and making the existing ones better and suitable for the next version of compilers and operating systems. But the work we've been doing on GPUs . many-core CPUs and improving the performance of large research codes on the HECToR service has convinced us that a new generation of numerical software is needed; software that can fulfil the promise of the coming many-core machines and software that can readily cope with the increasing number of data-intensive applications.

If you've read this far down you're likely to have more than a passing interest in NAG and in numerical software so I'll pose the question to you: What is the “next big thing” you would put on our to-do list. Write to me with your ideas at (rob@nag.com). At NAG we're ready, willing and able and I'm inspired by the thoughts of Daniel Burnham, the planner most associated with my adopted home of Chicago. His plan, published over 100 years ago, was central to Chicago's transformation from “the land of stinky onions” (the literal translation of the Illinois-Miami Indian name) to the modern city it is today. Burnham summed it up best when he said “Make no little plans: They have no magic to stir men's blood.”

Rob Meyer, CEO NAG


New routines available in the NAG Library for .NET


NAG is delighted to announce the availability of the NAG Library for .NET (Beta 2). The NAG Library for .NET is a subset of the NAG Library and includes a selection of routines, including the widely used and trusted optimization routines. Beta 2 sees the inclusion of more routines developed specifically for those developing using the Microsoft .NET framework. The new routines include quadrature, eigenvalues and eigenvectors, determinants, simultaneous linear equations, linear algebra support routines, linear equations (LAPACK), least-squares and eigenvalue problems (LAPACK), simple calculations on statistical data, random number generators, time series analysis and approximation of special functions. Beta 1 of the NAG Library for .NET was only available for 32-bit systems, however at beta 2 it will also be available in both 32-bit and for 64-bit.

If you develop in the .NET environment, think you might benefit from NAG's routines, and want to help advance this exciting prototype library contact us nagnews@nag.co.uk. For more information on the NAG Library for .NET http://www.[token_custom_nag_site]/numeric/DT/DTdescription.asp.

It's also worth remembering that you can also call NAG Library routines from the .NET environment as described in the following papers: Calling the NAG C Library (DLL) from C# http://www.[token_custom_nag_site]/numeric/CL/cl23associnfo/CSharp/csharpinfo and Calling the NAG Fortran Library (DLL) from C# http://www.[token_custom_nag_site]/numeric/csharpinfo.asp


NAG in the news: Scientific software for the 21st century


Anybody working in the software industry will be aware of the changes in hardware that are causing us to re-factor and, in many cases, rewrite our software. As the number of cores increases and their clock speeds decrease a great deal of attention has focussed on the technologies that we can use to address these developments. Should we stick with tried and well-understood technologies such as OpenMP or MPI, or switch to one of the many new languages that are springing up? Perhaps we should drop our procedural and object-oriented programmes and write everything in a functional language such as Haskell or F#?

Maybe we should think less about the technologies that we use and focus more on the algorithms that we employ.

To read the full article by Mike Dewar, Chief Technical Officer at NAG, that was recently published in Scientific Computing World visit http://www.[token_custom_nag_site]/IndustryArticles/scw_InsideView_MikeDewar.pdf


Advantage of using NAG in applied mathematics: A NAG user's perspective


“I gained my PhD in the Laboratory of Mechanics and Acoustics in Marseilles, France. The objective was to define a numerical algorithm able to recover some properties of an immersed body (density, elastic coefficients, size or shape). Experimentally diffraction of ultrasonic pulses was used: measurement was done all around the body and the experimental acquisition was already so efficient that a huge number of measurements could be done.

In order to reconstruct obstacles (the inverse problem), a classical method consists of comparing measurements with numerical simulations of the wavefield built with some a priori knowledge on the geometrical and material properties of the body. The estimation is then defined as the set of parameters for which the error between simulation and measurement is a minimum. Hence, these types of algorithm are optimization processes.

Inverse problems are known as ill-posed problems in the sense that the solution may not exist, or be multiple or unstable. In the view of these difficulties a reliable and fast numerical model was necessary.

All my colleagues advised me to use the NAG Library and Compiler (it's remarkable in the sense that unanimity is not so common in a research team). I will give here some steps where NAG has been useful during these three years.”

To read the full paper by Lo'c Le Marrec, Institut de Recherche Math'matique de Rennes, visit http://www.[token_custom_nag_site]/IndustryArticles/advantages_NAG_applied_mathematics.pdf


See NAG Portfolio Optimization and Option Pricing in action


Recently created and published on the NAG website is a recorded webinar highlighting how Portfolio Optimization can be achieved using the NAG Library within Excel. The ability to optimize portfolios using NAG's renowned algorithms is well established and its use in this field contributes to the popularity of this set of routines in the NAG Library. Option Pricing routines were added to the Library at Mark 22 and there is a second recording showing these routines in action.

To view the webinars visit http://www.[token_custom_nag_site]/numeric/nagexcelexamples/webinars.asp


Supercomputing: Tough choices for supercomputing's legacy apps


The future direction of supercomputing may seem dazzling, but it raises difficult questions about the survival of important software at the heart of today's facilities, says Andrew Jones.

The supercomputing community generally agrees that the future holds a number of software challenges. The first of these is the massively increasing degrees of concurrency required ' heading towards billion-way. Then there is the complex hierarchy of parallelism ' from vector-like parallelism required at the local level, through multithreading and onwards to multi-level, massive parallel processing across many nodes. On top of those challenges comes the impending storm of verification, validation and resilience.

Programming issues

Evolving our applications and middleware to address these issues is going to be a difficult, but necessary, job over the coming years, as petascale computers become increasingly common for scientific use and in corporate high-performance computing (HPC) facilities. The derived technologies place many teraflops in the hands of individual researchers, but they raise the same programming issues.

To read Andrew Jones', VP HPC Business at NAG, column visit http://resources.zdnet.co.uk/articles/comment/0,1000002985,39869521,00.htm


Out & About with NAG


  • NAG and NVIDIA hosted seminar: Banking on Monte Carlo and GPUs
    28th January 2010, Paris
    This event is almost full. If you would like to attend this event please complete the online registration form here
     
  • 5th Annual CARISMA Conference 2010: The Interface of Behavioural Finance and Quantitative Finance.
    2-3 February 2010, London
    The organizers of CARISMA are delighted to offer 20% discount of the registration fee to all NAG users. Visit http://www.optirisk-systems.com/events/carisma2010.asp for more details and to register.
     
  • Frankfurt MathFinance Conference
    15-16 March 2010, Frankfurt
    NAG is delighted to attending this event once again. The conference is intended for practitioners in the areas of trading, quantitative or derivative research and risk and asset management as well as for academics studying or researching in the field of financial mathematics or finance in general.
     
  • DevWeek 2010
    15-19 March 2010, London
    NAG is pleased to once again be present at the UK's biggest conference for Developers, DBAs and IT Architects.
     
  • GAMM 2010
    22-26 March 2010, Karlsruhe
    NAG is present at this event. The 81st Annual Meeting of the International Association of Applied Mathematics and Mechanics.
     
  • Young Statistician's Meeting (YSM)
    30-31 March 2010, University of Liverpool
    YSM is specifically designed for career young statisticians, by career young statisticians, providing an excellent opportunity to network amongst their peers, present their own research in a friendly atmosphere and discover a variety of other statistical topics currently being researched.
     
  • HECToR (High End Computing Terascale Resource) Training Courses
    Presented by the NAG HECToR Team
    A full list of forthcoming HECToR Training Courses can be viewed on the official HECToR website here.

For more information on any of the above events visit NAG's ‘Out & About’ webpage


New NAG product implementations


The NAG Toolbox for MATLAB, Mark 22 is now available for the following platforms:

  • Microsoft Windows 32-bit
  • Linux 64-bit interfacing to MATLAB 2009a and MATLAB R2009b

The NAG Fortran Library, Mark 22 is now available for the following platforms:

  • Apple Intel Mac64 OS X using the NAG Fortran Compiler, Release 5.1
  • x86-32bit Linux using the NAG Fortran Compiler, Release 5.1
  • x86-64 bit Linux using the pgf90 v.0-1 Compiler

The NAG Fortran Compiler, Release 5.2 is now available for the following platform:

  • Apple Intel Mac64 OS X

For full details of these and all other available implementations, visit the NAG site. Comprehensive technical details of each implementation are given in the relevant Installation and User Notes at http://www.[token_custom_nag_site]/doc/inun.asp


Managing your subscription


Please feel free to forward this newsletter to colleagues or to post it to your company's Intranet sites.

To subscribe, unsubscribe or learn more about managing your subscription, please see: http://www.[token_custom_nag_site]/NAGNews/index.asp.


NAGNews - Past Issues


We provide an online archive of past issues of NAGNews. For editions prior to these, please contact us.