NAGnews 127 | 4th December 2014

Posted on
19 Dec 2014

In this issue


New Fortran Builder 6.0 (Windows version of the NAG Compiler) Now Available


In a previous issue of NAGnews we announced availability of version 6.0 of the NAG Fortran Compiler. The Microsoft Windows version of the NAG Compiler is now also available. We call this version NAG Fortran Builder 6.0.

For readers unfamiliar with the NAG Fortran Builder it is an Integrated Development Environment (IDE) for the Microsoft Windows version of the NAG Fortran Compiler. The principle development aim of the Fortran Builder is to provide comprehensive features that will be invaluable to developers, including a GUI. The command line version of the compiler is still included in Fortran Builder for users that wish to continue with the previous interface. Fortran Builder has a wide range of key features that make it extremely easy to use. They include:

  • Produces either 32 or 64-bit code on 64-bit Windows systems (32-bit code only on 32-bit Windows systems)
  • GUI debugger
  • Integrated help system
  • "Pretty Printer" - also converts fixed to free format
  • Call graph generator
  • Interface block builder
  • Quad precision real
  • Informative runtime error messages
  • Set default Project Settings & reset option
  • Set default Project Settings & reset option
  • I/O library multi-threaded for performance
  • Improved Random Number Generator
  • Integrated NAG Library support
  • Additional integrated libraries: OpenGL, GTK+
  • Quick execution mode
  • Jump to location of runtime errors
  • Memory leak detection built into the compiler

Fortran Builder Error Detect image
NAG Fortran Builder 6.0 Error Detect Screenshot

More information on the Fortran Builder can be found on our website. Many readers will be entitled to use the NAG Fortran Builder as part of their NAG Software Licence - if you think this applies to you, email us and we'll do the checking. Trial licences are also available here.


Training Course (Oxford): Portable, Bug Free Fortran with the NAG Fortran Compiler


We are delighted to announce available places at "Portable, Bug-free Fortran with the NAG Fortran Compiler" training course at NAG Oxford on 12th December 2014.

Course Outline:

As all developers know their primary tool during the code development cycle is the compiler. A compiler that is strong on identifying bugs and portability issues clearly, accurately and early in that cycle is of huge benefit to the programmer. The NAG Fortran Compiler is widely recognized as one such compiler (for instance http://www.polyhedron.com/fortran-compiler-comparisons/win32-fortran-compiler-comparisons-diagnostic-capabilities). In this half-day course we will introduce its capabilities, showing how bugs can be easily found by use of the NAG Compiler, including classes of bugs that are often difficult to find, such as uninitialized variables.

The course will consist of a lecture explaining how to use the facilities available within the compiler. This will include real examples of bugs identified when developing well known Fortran codes.

It will then be followed by a practical session where attendees can use the NAG Compiler and see for themselves its capabilities.

Registration has now closed.


NAG Student Prize White Paper: "Continuous Wavelet Transform and Wavelet Coherence - Implementation and Application to the Diversification Analysis of Hedge Fund Returns"


NAG was delighted to award the 2014 "Direct Award" Student Prize to Warwick Business School student, Christoph Auth recently. Christoph received his award for his excellent paper "Continuous Wavelet Transform and Wavelet Coherence - Implementation and Application to the Diversification Analysis of Hedge Fund Returns".

Abstract

This paper investigates the diversification benefits of hedge funds as an alternative investment class. For this purpose, it carries out a time-frequency analysis. Wavelet squared coherence estimates based on the continuous wavelet transform measure the co-movements of monthly returns from six hedge-fund indices and the S&P500 Composite Index over the period from January 1994 through December 2012. Additionally, the paper compares the results to coherency spectra based on Fourier transforms. For the wavelet analysis, it employs the commonly used Morlet wavelet coherence estimates with the latter function. Results show an increased interrelation between hedge funds and stock markets during financial crises in general and the most recent subprime-mortgage crisis in particular. Therefore, diversification benefits attained through hedge-fund investments became less pronounced during the most recent years, especially for long- and medium-term investors. However, Equity Market Neutral and Global Macro funds still provide relatively good diversification gains.

Read Christoph's paper here.


Mark 24 New Functionality Spotlight: Matrix Square Roots and General Powers and Fréchet Derivative (Matrix Operations Chapter)


New routines have been added to compute matrix square roots. They are f01en* that computes a real matrix square root and f01ep* specifically for real upper quasi-triangular matrices, which have or may have some 2x2 blocks on the diagonal. The complex counterparts are f01fn* and f01fp*. Routines are also provided for general powers of a matrix, f01eq for real matrices and f01fq* for complex. They compute the pth power, for any real p. Please note these functions are currently available in the NAG C Library Mark 24, and will be made available in the NAG Fortran Library Mark 25 in 2015. (http://www.nag.co.uk/numeric/CL/nagdoc_cl24/html/F01/f01conts.html

The condition number of a matrix function is a measure of the sensitivity of the computed solution to small changes in the input data. The Library contains many routines for estimating matrix function condition numbers along with the function value. Another interesting quantity that can be computed is the Fréchet Derivative, which describes the first-order effect on the matrix function of perturbations to the input matrix - see http://www.nag.co.uk/numeric/CL/nagdoc_cl24/html/F01/f01kkc.html and http://www.nag.co.uk/numeric/CL/nagdoc_cl24/html/F01/f01kfc.html for more details.


New Thinking in Maths: TakeAIM Winners Announced


Organised by the Smith Institute for Industrial Mathematics and System Engineering, the goal of TakeAIM is to make visible the crucial role that mathematics plays and will increasingly play in all aspects of our lives. The competition highlights some of the areas where new thinking is coming over the horizon, with influence far outside the academic world. It enables young researchers to express in their own words the potential future impact of their work.

NAG is delighted to have supported TakeAIM since its inception. A key element of our support is to be a member of the judging panel. Each year the quality of entries is very high and 2014 was no exception. However a winner was decided upon and at the Alan Taylor day, the winner and five joint runners-up were awarded their prizes.

Congratulations to this years' winner, Luca Calatroni for his work "Image Ma-therapy: detect or defect?" Luca was awarded £1,000 of Apple vouchers and each runner-up received £500. Well done to all for some outstanding work.

Links to Luca's and the runners-up work can be found here.

Fortran Builder Error Detect image
Dr David Sayers, NAG Honorarium (right) with TakeAIM runner up Peter Baddoo of the University of Oxford.


Training Courses & Events


NAG will be at the following exhibitions and conferences over the next few months.


NAGnews - Past Issues


We provide an online archive of past issues of NAGnews. For editions prior to 2010, please contact us.