NAGnews 148

Posted on
18 May 2017

In this issue:


The NAG Library for Python updated with new Optimization and Nearest Correlation Matrix Techniques


The updated NAG Library for Python gives users of the Python programming language access to hundreds of fully tested, robust, and high performing mathematical and statistical routines in line with the latest releases of Python. The NAG Library for Python has recently been updated with some new features including a linear and nonlinear semidefinite programming (SDP) solver, and a new interior point method for large-scale nonlinear programming problems. To make these new optimization solvers accessible, NAG has introduced the Optimization Modelling Suite which allows users to build up their problem to be solved in stages, instead of calling one monolithic solver with many arguments; making it simpler to use and easier to avoid mistakes.

Other additions to the NAG Library for Python: Gaussian Quadrature, Least Squares and Eigenvalue Problems (LAPACK), Nearest Correlation Matrix functions, and OpenMP Utilities.

Enhancements for users also include updated Example Scripts which give detailed explanation of the routines to aid in their use. For more information see NAG Library for Python - product trials are available.


NAG and MathJax - Documentation Affected


The NAG Library documentation makes use of the MathML format for displaying mathematics within web pages. This is supported natively in the Firefox browser, but for users of other browsers we use the freely available MathJax javascript library. Unfortunately the MathJax consortium has had to shut down its server as detailed here. As outlined in that page, users may instead use a locally installed copy of MathJax, or may use a different freely available server.

The copies of the NAG Library documentation on our website have already been updated, for example Fortran Library Documentation and C Library Documentation.

We emailed NAG Library users directly about this issue and blogged about it here. If you have any questions please contact NAG Technical Support.


Working to Identify the Challenges of Code/Theory Translation


In January this year a Code/Theory Workshop was held at the University of Manchester. The event brought together programmers and architects from academia and industry to discuss the challenges of translating between 'theory' - scientific or logical concepts - and 'code' - software, ontologies or linked data. One of the participants was Jonathan Boyle, HPC Application Analyst at NAG.

The workshop started with each participant giving a lightning talk outlining their experiences in this area. This was followed by two discussion sessions in small groups: "Why is code/theory translation challenging?" and "How can we improve code/theory translation?".

In his presentation Jonathan described theory as collective knowledge, often contained within papers, that is converted into personal knowledge, particularly into understanding, and then into code design. This process can be hard work, and the way in which it happens is not always obvious, and not well understood; it occurs at both a conscious and unconscious level, involving creativity and reflection. Translating theory into design is hard, while actually writing the code is more straightforward. The challenge involved in translating theory to code also varies considerably, with different levels of complexity, and some implementations being much more straightforward than others. There are also social aspects to consider, as the process requires teamwork and different sets of skills, as well as potentially having a variety of goals and constraints. It is important to consider the context in which it is happening.

You can read the workshop round-up report here.


Webinar: Dissecting the Myths of Cloud, GPUs and HPC


Andrew Jones, Vice-President, HPC Business at NAG will deliver this impartial webinar dissecting the myths surrounding cloud, GPUs and HPC. A must for anyone involved in computing.

Are you trying to answer either of these two questions?

  1. Which processor architecture is right for your HPC?
  2. Should your HPC system be in the cloud or in house?

What you will get from the webinar:

  • A live webinar so you can have your questions answered, either during the session or post event.
  • An expert practitioner's personal view - independent, with deep technical knowledge and experience.
  • Real-world advice on how to select the best HPC system for your specific needs.
  • The truth behind the marketing statements made by vendors and solution providers.
  • An understanding of new ways to mitigate against risk in your decision processes.

Register here


Best of the Blog


To users of NAG Library Documentation - MathJax


Out & About with NAG


Come and see us at various conferences and events over the next few months.

Training Courses & Webinars

Introduction to High Performance Computing
17-18 May 2017

Dissecting the Myths of Cloud, GPUs and HPC
7 June 2017

Fortran Modernization Workshop, Universitat Politecnica de Catalunya
24-26 July 2017

NAG-TACC Institute Summer School
11-15 September 2017

Exhibitions, Conferences and Trade Shows

The Trading Show Chicago 2017
17-18 May 2017

PyCon 2017
17-25 May 2017

POP User Forum at HPC Summit Week 2017
19 May 2017

STAC Summit 2017
5 June 2017

NAFEMS World Congress 2017
11-14 June 2017

ISC High Performance 2017
18-22 June 2017