Discover the Technology Powering the Next Generation of Research at the University of Nottingham

The University of Nottingham, one of the UK’s leading research Universities and member of the Russell Group of Universities, is renowned for the strength and quality of its research and teaching capabilities. In order to further strengthen their unwavering commitment to research excellence, the University of Nottingham is making a significant investment in a state-of-the-art new High Performance Computing (HPC) system. This article provides more insight into the University's latest ￡3 million cluster named Ada, and how it will enhance research capabilities across a wide range of disciplines, fuelling innovation. Channel partner OCF was responsible for the design, development, and installation of this cutting-edge HPC cluster.

Chapter 1: Introducing Ada - Nottingham Universitys New HPC Cluster.

The new cluster is named after Ada Lovelace, a renowned and gifted English mathematician, recognised for creating the first computer programme, and foreseeing a digital future. Famous for her contributions to Charles Babbages proposed mechanical general-purpose computer, the Analytical Engine, Ada was the first to recognise the machines potential beyond pure calculation. The name is a fitting tribute to the female pioneer, who also had ancestral ties to Nottinghamshire.

‘Ada’ is set to replace the current Augusta cluster, which will reach its end of life in October 2023. The arrival of Ada brings forth exciting possibilities, facilitating broader participation and unlocking numerous opportunities.

These advancements will drive growth in strategic areas of the University, such as quantum technologies, nanoscience, artificial intelligence, imaging, and bioinformatics. This, in turn, will lead to increased research income and outputs.

Chapter 2: The Technology Behind Ada

As technology continues to evolve at an exponential pace, it is crucial to ensure the University’s HPC infrastructure can meet growing demand accordingly. ‘Ada’ introduces a significant upgrade in power and capability, providing the University with the necessary infrastructure to process and analyse ever-growing volumes of data. Its speed enables researchers to perform complex calculations, yielding high-quality and impactful outcomes in key research domains.

2.1 Hardware Components

Ada's strength lies in its powerful combination of hardware and software components. It leverages Lenovo ThinkSystem Servers and Storage, AMD EPYC Processors, and NVIDIA A100 GPUs to provide increased computational power and storage capacity. Ada empowers users in AI/ML with enhanced GPU capabilities, enabling faster and more accurate training of neural networks for quicker insights and better decision-making. The NVIDIA A100 GPUs, equipped with third-generation Tensor Cores, deliver greater speed and efficiency for AI workloads, allowing for simultaneous training of multiple models without performance loss. The combination of AMD EPYC processors and NVIDIA A100 GPUs efficiently manages and processes large amounts of data, crucial for AI/ML applications. Compared to Augusta, Ada's advanced hardware setup provides a more robust and efficient environment for developing and deploying AI/ML models. Additionally, Ada offers dedicated data visualisation nodes for pre and post processing, a significant improvement over Augusta's? remote graphics.

2.2 Software Components

The HPC utilises OCF’s Steel HPC Management suite alongside a host of professional and managed services to ensure the smooth running of the system over its lifetime.

While the majority of the upgrades in Ada compared to Augusta primarily focus on hardware replacements, the software stack is also a crucial aspect. The software plays a significant role in controlling key functions of the resulting system, ensuring that maximum benefits are achieved.

One of Ada's main goals is to minimise the disruption to existing users' work. To achieve this, Ada carries over the existing end-user software from Augusta, while also adding new application software that enhances users' productivity.

Ada includes:

• Compilers: Ensure effective compilation on the new system, maximizing new investments while minimising user demands.
• Storage management: Ada's high-speed storage is also software-managed and optimized, building on Augusta's success. Hardware and controlling software are provided as a compatible package.
• Management software: Performs functions like user authentication, tracking and billing, scheduling, and equipment status monitoring (including OCF Steel).
• User applications software and support: Optimised to minimise necessary changes for the new configuration.
• Interfaces: Offers a baseline interface, improved GUI, web portal, and application streaming capability.
• These enhancements aim to improve the overall performance, user experience, and compatibility of Ada.

2.3 Data Transfer Capabilities, Integration and Accessibility

Ada addresses the challenge of slow data transfer, a crucial aspect when dealing with large datasets. By integrating and simplifying access processes, the HPC cluster becomes more accessible, broadening its range of applications and users.

Ada has also implemented seamless links between the new HPC and research storage, effectively resolving issues related to data movement and access. One of the major challenges addressed by Ada is the ability to access the HPC from the research data storage (on-premise NetApp) without unnecessary data duplication. Additionally, Ada has successfully addressed GDPR implications and version control issues, which previously led to time-consuming and inefficient processes.

Chapter 3: Ada's Impact on Research

The cluster addresses a wide range of topics, with a current focus on applications such as quantum technologies, computational chemistry, artificial intelligence, computational fluid dynamics modelling, nanoscience, and medical imaging.

The service will also be available for experimentation beyond larger, funded projects, leading to more ambitious research and resulting funding, more high-quality outputs, increased inter-disciplinary collaboration, attracting researchers and PGRs.

Ada will help address research in the area below:

• Simulation and modelling (memory, visualisation, data storage, data movement, bandwidth)
• Visualisation (data movement, data sharing, data backup, data security, bandwidth)
• Algorithm or software development (GPGPU, cores, data movement, data backup)
• Image analysis (interactive, data backup, data movement, data sharing, GPGPU)
• Machine learning (memory, cores, data storage, data movement, data sharing, bandwidth, GPGPU)
• Medical data (visualisation, data movement, data sharing, bandwidth)
• Statistics (interactive, data movement, data backup, data storage, Linux)
• Bioinformatics (data storage, data movement, data backup, bandwidth)

The HPC cluster's ability to handle 'big data' applications also opens up new avenues for research.

Conclusion

Ada marks a significant milestone in the University of Nottingham's journey towards research innovation and excellence. By providing a more powerful, efficient, and accessible HPC service, it enables researchers to push the boundaries of innovation and discovery. This investment underscores the University's commitment to nurturing a conducive environment for research and making significant contributions to knowledge and society.