In this issue:
- Algorithmic Differentiation algorithms enter the NAG Library for the first time
- Re-engineered NAG Library for Python – sign up for NAG’s ‘tutorial’ webinar
- CVA at Scale with Adjoint Sensitivities
- Introducing Origami: A Grid/Cloud Task Execution Framework
- NAG Library for Java updated to Mark 26.2
- Webinar: Getting Started with the NAG Library for Python
- Tech Report: Using the NAG Library with kdb+ in a Pure q Environment
- Tech Report: The Role of Matrix Functions
- In other news: NAG’s HPC role in EU POP project set to continue
Algorithmic Differentiation algorithms enter the NAG Library for the first time
The NAG AD Library, launched at the WBS Quant Finance Conference recently, provides AD enabled numerical routines to make the AD process quicker, more efficient and productive, and eliminate the need to write your own code or rely on unsupported code.
The NAG AD Library has been designed so that it can be used with or without any Algorithmic Differentiation (AD) tool; however, the conversion from code containing calls to primal routines to an adjoint version computes seamlessly when combined with NAG's AD tool, dco/c++ – recently updated to version 3.3.0.
Learn more about how NAG continues to pioneer in AD Solutions here. It might be that you are entitled to use the NAG AD Library as part of your NAG software licence agreement. If you’re unsure about your or your organisation’s eligibility, do get in touch and we’ll check for you.
Re-engineered NAG Library for Python – sign up for NAG's 'tutorial' webinar
The latest NAG Library for Python release brings new algorithmic content in the area of mathematical optimization. All NAG Library content works effortlessly with NumPy data types.
The re-engineered NAG Library gives developers of Python many improvements to usability, enabling quicker application development including: self-contained routine documentation, thread-safety, curated examples demonstrating important algorithmics and Python-related functionality, and support for native Python callbacks and C language routine equivalents giving a unified route from prototyping in Python to deploying as compiled code.
Alongside the NAG Library for Python significant usability improvements, a number of new solvers have been added to the Library including two new NAG Optimization algorithms: Derivative-free Optimization for Data Fitting and Interior Point Method for Large Scale Linear Programming Problems further expanding the number of world-class maths and stats algorithms available in the NAG Library.
The 64-bit NAG Library for Python is available for Microsoft Windows, Linux and Mac OS, and is compatible with Python 2.7, 3.5 and 3.6.
Full product trials are available for the NAG Library for Python.
Sign up for NAG’s ‘tutorial’ Webinar: ‘Getting Started with the NAG Library for Python’ – 29 November 2018
CVA at Scale with Adjoint Sensitivities
A new Technical Poster 'CVA at Scale with Adjoint Sensitivities' was showcased earlier in the Autumn. The poster tackles the performance challenge in calculating XVA and presents a software-based solution using AD-enabled (adjoint and tangent) versions of NAG Library algorithms, dco/c++, and Origami (a light-weight task execution framework for grid and cloud). This solution offers users a way to manage the CVA at scale challenge in a cohesive and cost-effective way.
Introducing Origami: A Grid/Cloud Task Execution Framework
Origami is the framework that enables optimal task execution across your HPC infrastructure connecting workstations, datacentre and the cloud. Motivated by frustration with dated, inflexible products Senior Quant Dev professionals created Origami to deliver simple set-up, rapid deployment, ease of programming and monitoring of distributed quant solutions – typically quant/risk libraries and applications.
Xifintiq have partnered with NAG – the market leading numerical library and tools provider to investment banking – to launch Origami. NAG's experience in rigorous testing and correctness combined with its respected position in the quant finance community deliver an unrivalled level of service to Origami users. This service includes tailored implementation and installation ensuring users gain maximum benefit from the framework, user training, plus first-line technical support.
- Origami puts the power of on-premise, hybrid and cloud infrastructure resources at your fingertips
- Origami removes constraints that are inherent in other grid solutions and empowers you to deliver faster solutions
- Origami delivers optimized workload scheduling using Directed Acyclic Graphs (DAGs), and performance can scale with the size of the grid
- Origami offers game-changing robustness and flexibility
- Origami reduces cost by enabling simultaneous easy selection of heterogenous hardware (CPU/GPU)
- Origami lets you identify and correct compute bottlenecks
New NAG Library for Java updated to Mark 26.2 – featuring new Optimization routines
NAG is pleased to report the release of the new NAG Library for Java, Mark 26.2. Many of our customers in finance use this version of the Library. This release sees the availability for Java users of a number of new solvers, including new Optimization functionality: Derivative-free Optimization for Data Fitting and an Interior Point Method for Large Scale Linear Programming Problems.
Webinar: Getting Started with the NAG Library for Python | 29 November 2018
This webinar will give, a short tour of the documentation of the new package, installation guide, ‘Quick Start’ - solving an optimization problem, and give further examples from maths and stats. More information and registration.
Technical and Numerical Research
Using the NAG Library with kdb+ in a Pure q Environment
NAG and Kx share several customers, many of whom use the NAG Library with the q programming language. Within this report, we describe a procedure for using the NAG Library with kdb+ in a pure q environment that will enable developers to write less code, employ fewer development tools, and shorten overall development time. Read the report here.
In this report we discuss functions of matrices. What are they, and why might they be of interest to you? We use the NAG Library for Python to provide examples and have included some code snippets that can run if you have your own installation of the Library. Read the report here.
In other news: NAG's HPC Role in EU POP Project set to continue
NAG is delighted that the EU group called POP (Performance Optimisation and Productivity) that helps to improve the performance of parallel software is set to be recommissioned very soon. In brief, POP offers to analyse software and recommend improvements, free of charge, with a focus on HPC and parallelisation. NAG, along with the other POP partners have proven the benefits of application analysis to many EU organisations already – read about some of their work here. If you’re interested in the work that the POP project undertakes and would like to apply for their services do get in touch.