NAGnews 159
Published 24/01/2019 By NAG

We wish you a very happy and prosperous 2019.

In this issue:

 New Release of AAD Tool for Accelerators offers ease-of-use and performance improvements

We are excited to announce a host of improvements to our cross-platform, accelerator ready Adjoint Algorithmic Differentiation (AAD) Tool, dco/map. dco/map is used to create adjoints of performance-critical sections of code, be they C++/OpenMP or CUDA. It has found application notably in accelerated XVA platforms where it helps deliver first and second order sensitivities.

What's new in v1.6 dco/map

  • New bitwise-copyable reduction push array. For many workloads, reduction push is still the fastest array type, and the new class makes it easy to access this performance in existing C++ codes
  • Array management functions now allocate bitwise-copyable adjoint arrays directly, making it significantly easier to integrate the array classes into existing codes and class hierarchies. It is now easier to apply dco/map to an existing C++ code base
  • Performance improvements to atomic push arrays - these are now approximately 50% faster in double precision
  • Improved dco/map external adjoint object for easier interoperability with dco/c++
  • Enhanced MAP_PRINT functionality to give more information to the user and make it easier to process the data
  • Overhauled the training material: users should now find it easier to get up to speed with dco/map

For more information see https://www.nag.com/content/algorithmic-differentiation-software or contact us at info@nag.com with questions or just to learn more about the product.

NAG HPC Consulting Team enhance computational fluid dynamics application for optimized ship design

Cetena is a study centre, based in Italy, that carries out a range of research and consulting work for the maritime industries. One of the software tools Cetena utilise as part of their virtual prototyping work is a computational fluid dynamics (CFD) application, known as HELYX-EcoMarine from ENGYS.

NAG, as HPC experts, were asked to investigate the optimal system configuration for the CFD based application to achieve good performance on the chosen HPC cluster platform.

Project Scope

Ship builders and operators, wanting to enhance their fleets, require ever more efficient and safer hull designs to cope with the demands of rough seas. Often, specialist consulting businesses are needed to provide expertise in fluid and environmental modelling. These subject specialists, such as Cetena, use advanced software for virtual modelling to optimize ship design. To assist with their virtual modelling work Cetena turn to application vendors, such as ENGYS, to provide the best software tools. NAG fits into this ecosystem of subject experts by providing the knowledge and experience of running a variety of numerically based software applications on large HPC clusters.

Read about the improvements NAG made to the project here.

Webinar: Getting Started with the NAG Library for Python

Watch the new video 'Getting Started with the NAG Library for Python' for a short tour of the documentation of the new package, Installation Guide, and 'Quick Start' - solving an optimization problem.

NAG and Manchester University Knowledge Transfer Partnership to develop new numerical linear algebra for emerging computer architectures

NAG is delighted to be working in collaboration once again with the Numerical Linear Algebra Group at The University of Manchester. After the success of two previous Knowledge Transfer Partnerships (KTPs) with Manchester, NAG announces a third project focussing on the development of new numerical linear algebra routines for emerging computer architectures.

The project pulls together an impressive group: Professors Jack Dongarra, Nick Higham, and Françoise Tisseur, all who will be working with the KTP Associate, Mawussi Zounon, primarily based at NAG's Manchester office. Mawussi holds a PhD on the resilience of numerical solvers for linear systems and eigenvalue problems in the context of HPC applications and modern HPC architectures. Following his PhD, he fulfilled a postdoctoral research role at the University of Manchester working on the EU project "Parallel Numerical Linear Algebra for Extreme Scale Systems" (NLAFET). It was in this position that he worked closely with, former NAG Principal Technical Consultant, Professor Sven Hammarling, who changed Mawussi's perception of a career in industry based on Sven's experience at NAG. Learn more about the KTP here https://www.nag.com/content/ktp-partnership

NAG Prize for Outstanding Performance on the Numerical Analysis Undergraduate Study Pathway

On Friday 6th July 2018, Craig Lucas presented awards for the NAG Prize for Outstanding Performance on the Numerical Analysis Undergraduate Study Pathway at the University of Manchester.

Instead of the two prizes that we have presented annually in the past, NAG is now awarding three winners. The 2018 recipients were: Thomas Arthur, Guannan Chen and Jingbang Liu. Unfortunately, Craig was only able to present the prize to Thomas, as Guannan and Jingbang were unable to attend the prize giving.

We asked the students what their future plans were: Thomas Arthur said "I'm about to start an internship with Systra, a global transport and infrastructure consultancy. Not too sure what after, but hopefully something numerical analysis related."

Guannan Chen replied, "I am going to University of Oxford to study MSc Mathematical Modelling and Scientific Computing in October 2018."

And Jingbang Liu said "I am continuing my studies in Oxford. In the future, I would like to stay in academia."

Congratulations to Thomas, Guannan and Jingbang from all at NAG! We wish you every success in your future endeavours.

Out & About with NAG

Exhibitions, Conferences, Trade Shows and Webinars

Fortran Modernization Workshop
University of Manchester, 4-5 April 2019

This workshop is open to all and is free to attend.

PyCon 2019
Cleveland - OH, 1-9 May 2019

The Trading Show
Chicago, 8-9 May, 2019

Quant Minds International
Vienna, 13-17 May 2019