EPCC Tier 2 HPC Service
Lead Research Organisation:
University of Edinburgh
Department Name: Edinburgh Parallel Computing Centre
Abstract
EPCC is the supercomputer centre at the University of Edinburgh. Throughout our 26-year history, EPCC has significant experience of running supercomputer services for both academic and industrial researchers. We currently run the UK's national supercomputer, ARCHER, that is used for modelling new materials, designing more efficient engines and investigating climate change.
This proposal is to enhance an existing supercomputer with additional compute nodes, and to add a significant amount of next generation accelerators. The design of supercomputers is going through a period of change and accelerators are becoming increasingly important to provide faster supercomputers while using less electricity. This enhanced system will be useful for researchers across the UK to prepare their computer programmes to make effective use of future national supercomputer systems after ARCHER.
As computers become more powerful, they generate ever-increasing amounts of data. Managing and analysing this data is very important to help researchers make new scientific discoveries. This proposal also includes increasing the size of the large data storage facility at EPCC. This will allow researchers to store their data, share it with other people, move it between different supercomputers, and make new discoveries that will improve society and people's lives.
There will be a number of these medium-size supercomputers funded alongside the existing national supercomputer, ARCHER. These will have different features and so will be useful for different research problems. It is important that these supercomputers work together to provide maximum benefits for everyone. We will work with all the other successful proposals to provide a coherent service across all the different supercomputers. Most of the time on our supercomputer will be freely available for researchers from throughout the UK. We will work together with the other successful proposals to ensure that researchers can get access to the most appropriate system/ In addition, we have developed a we.-based portal that makes it easier for users to access the different systems. As part of our proposal, we will work with all the other successful proposals so that they can use our web-based portal completely free of cost.
This proposal is to enhance an existing supercomputer with additional compute nodes, and to add a significant amount of next generation accelerators. The design of supercomputers is going through a period of change and accelerators are becoming increasingly important to provide faster supercomputers while using less electricity. This enhanced system will be useful for researchers across the UK to prepare their computer programmes to make effective use of future national supercomputer systems after ARCHER.
As computers become more powerful, they generate ever-increasing amounts of data. Managing and analysing this data is very important to help researchers make new scientific discoveries. This proposal also includes increasing the size of the large data storage facility at EPCC. This will allow researchers to store their data, share it with other people, move it between different supercomputers, and make new discoveries that will improve society and people's lives.
There will be a number of these medium-size supercomputers funded alongside the existing national supercomputer, ARCHER. These will have different features and so will be useful for different research problems. It is important that these supercomputers work together to provide maximum benefits for everyone. We will work with all the other successful proposals to provide a coherent service across all the different supercomputers. Most of the time on our supercomputer will be freely available for researchers from throughout the UK. We will work together with the other successful proposals to ensure that researchers can get access to the most appropriate system/ In addition, we have developed a we.-based portal that makes it easier for users to access the different systems. As part of our proposal, we will work with all the other successful proposals so that they can use our web-based portal completely free of cost.
Planned Impact
The EPCC National Tier 2 Integration and HPC Accelerator Service will deliver strong scientific and industrial impact now and in the future in three specific areas:
1. Through support via SAFE for the creation of a national network of Tier 2 centres;
2. By preparing key codes from the UK scientific software base for use on Intel Xeon Phi accelerators on the road to the Exascale by the mid-2020s; and
3. Delivering an enhanced EPCC industry service to companies around the UK, Europe and beyond.
SAFE
EPCC has run national services for EPSRC since the Cray T3D in 1994. Our SAFE software has developed over the past 20 years from simple scripts to the comprehensive web-based system we use today. Although most UK users know SAFE from the point of view of account and resource management, there is also significant value in its helpdesk and reporting functionality.
By deploying SAFE across the Tier 2 network we hope to amplify the value of this funding, demonstrating the impact a joined-up Tier 2 service will have on EPSRC's scientific community and the wider UK scientific and industrial communities.
SCIENCE
Although Exascale computers are 6-8 years away, the technologies that are likely to be used in them are becoming clearer. There will be an increase in the use of accelerators, a greater focus on heterogeneity, more complex memory hierarchies, and new models of I/O. The HPC Accelerator Service that will be provided by this funding will allow EPCC to work with the EPSRC scientific community to prepare their scientific software codes and simulation techniques for the very high core counts and complex memory hierarchies expected to dominate in the future. By starting this work now, the impact of future national services, such as ARCHER 2, will be maximised.
Additionally, the funding will support the creation of a Tier 2 RDF service. Here we will work with the other Tier 2 centres and their user base to explore the value of different high level data services such that the future impact of investment in EPSRC's RDF can be maximised.
INDUSTRY
Since 1990 when EPCC was created, we have had a strong focus on working with industry. Not only is such work a key part of our business model, it is also a key part of our mission - "to accelerate the effective uptake of HPC and novel computing by industry, academia and commerce". Our value proposition delivers competitive advantage in the form of lower costs, improved products and services, and accelerated time to market for our industrial customers in segments such as engineering, energy, oil and gas, renewables, manufacturing and materials, life sciences, aerospace and automotive, and financial services.
Our successful and sustainable business model has been instrumental in allowing us to invest in HPC resources for industry. In 2012, EPCC, at its own expense, bought the INDY system that it has used, along with ARCHER, to deliver cycle-sales to industry over the past 4 years. In April 2016, this system was replaced by INDY 2, a state of the art SGI ICE XA system dedicated to industrial use. It is this system we intend to expand with this funding, further encouraging industrial collaboration and our long-term sustainability.
1. Through support via SAFE for the creation of a national network of Tier 2 centres;
2. By preparing key codes from the UK scientific software base for use on Intel Xeon Phi accelerators on the road to the Exascale by the mid-2020s; and
3. Delivering an enhanced EPCC industry service to companies around the UK, Europe and beyond.
SAFE
EPCC has run national services for EPSRC since the Cray T3D in 1994. Our SAFE software has developed over the past 20 years from simple scripts to the comprehensive web-based system we use today. Although most UK users know SAFE from the point of view of account and resource management, there is also significant value in its helpdesk and reporting functionality.
By deploying SAFE across the Tier 2 network we hope to amplify the value of this funding, demonstrating the impact a joined-up Tier 2 service will have on EPSRC's scientific community and the wider UK scientific and industrial communities.
SCIENCE
Although Exascale computers are 6-8 years away, the technologies that are likely to be used in them are becoming clearer. There will be an increase in the use of accelerators, a greater focus on heterogeneity, more complex memory hierarchies, and new models of I/O. The HPC Accelerator Service that will be provided by this funding will allow EPCC to work with the EPSRC scientific community to prepare their scientific software codes and simulation techniques for the very high core counts and complex memory hierarchies expected to dominate in the future. By starting this work now, the impact of future national services, such as ARCHER 2, will be maximised.
Additionally, the funding will support the creation of a Tier 2 RDF service. Here we will work with the other Tier 2 centres and their user base to explore the value of different high level data services such that the future impact of investment in EPSRC's RDF can be maximised.
INDUSTRY
Since 1990 when EPCC was created, we have had a strong focus on working with industry. Not only is such work a key part of our business model, it is also a key part of our mission - "to accelerate the effective uptake of HPC and novel computing by industry, academia and commerce". Our value proposition delivers competitive advantage in the form of lower costs, improved products and services, and accelerated time to market for our industrial customers in segments such as engineering, energy, oil and gas, renewables, manufacturing and materials, life sciences, aerospace and automotive, and financial services.
Our successful and sustainable business model has been instrumental in allowing us to invest in HPC resources for industry. In 2012, EPCC, at its own expense, bought the INDY system that it has used, along with ARCHER, to deliver cycle-sales to industry over the past 4 years. In April 2016, this system was replaced by INDY 2, a state of the art SGI ICE XA system dedicated to industrial use. It is this system we intend to expand with this funding, further encouraging industrial collaboration and our long-term sustainability.
Organisations
Publications
Townsend-Nicholson A
(2020)
Educating and engaging new communities of practice with high performance computing through the integration of teaching and research
in Interface Focus
Tian T
(2020)
Electronic Polarizability as the Fundamental Variable in the Dielectric Properties of Two-Dimensional Materials.
in Nano letters
Thouvenin P
(2023)
Parallel faceted imaging in radio interferometry via proximal splitting (Faceted HyperSARA) - II. Code and real data proof of concept
in Monthly Notices of the Royal Astronomical Society
Thouvenin P
(2023)
Parallel faceted imaging in radio interferometry via proximal splitting (Faceted HyperSARA): I. Algorithm and simulations
in Monthly Notices of the Royal Astronomical Society
Terris M
(2023)
Image reconstruction algorithms in radio interferometry: From handcrafted to learned regularization denoisers
in Monthly Notices of the Royal Astronomical Society
Terris M
(2020)
Building Firmly Nonexpansive Convolutional Neural Networks
Terris M
(2021)
Enhanced Convergent PNP Algorithms For Image Restoration
Tagliafierro B
(2022)
Numerical Assessment of a Tension-Leg Platform Wind Turbine in Intermediate Water Using the Smoothed Particle Hydrodynamics Method
in Energies
Taddei M
(2019)
Band gap modulation in zirconium-based metal-organic frameworks by defect engineering
in Journal of Materials Chemistry A
Sutcliffe E
(2023)
Transient magneto-optical spectrum of photoexcited electrons in the van der Waals ferromagnet Cr 2 Ge 2 Te 6
in Physical Review B
Su J
(2024)
Low-frequency still-air acoustic inertia of inclined circular aperture in an infinite flat plate of finite thickness
in Journal of Sound and Vibration
Strungaru Mara
(2022)
Ultrafast laser-driven topological spin textures on a 2D magnet
in NPJ COMPUTATIONAL MATERIALS
Strungaru M
(2022)
Ultrafast laser-driven topological spin textures on a 2D magnet
in npj Computational Materials
Strathearn A
(2018)
Efficient non-Markovian quantum dynamics using time-evolving matrix product operators.
in Nature communications
Stewart A
(2021)
Chemical functionality at the liquid surface of pure unsaturated fatty acids
in Environmental Science: Atmospheres
Stewart A
(2023)
Surface functionality of sub- to full-monolayer organic coverage of water aerosols determined by molecular dynamics simulations
in Environmental Science: Atmospheres
Stewart A
(2022)
The influence of saturation on the surface structure of mixed fatty acid-on-water aerosol: a molecular dynamics study
in Environmental Science: Atmospheres
Sterling AJ
(2020)
Rationalizing the diverse reactivity of [1.1.1]propellane through s-p-delocalization.
in Chemical science
Sterling A
(2022)
Beyond strain release: Delocalisation-enabled organic reactivity
Sterling A
(2023)
Beyond strain release: Delocalization-enabled organic reactivity
Sterling A
(2021)
Beyond strain release: Delocalisation-enabled organic reactivity
Spracklen P
(2023)
Automated streamliner portfolios for constraint satisfaction problems
in Artificial Intelligence
Smith D
(2022)
Noise Source Analysis in Counter-Rotating Open Rotors
in AIAA Journal
Smith D
(2020)
Acoustic Analysis of Counter-Rotating Open Rotors with a Locked Blade Row
in AIAA Journal
Silveri F
(2022)
Catalytic Reduction of Carbon Dioxide on the (001), (011), and (111) Surfaces of TiC and ZrC: A Computational Study
in The Journal of Physical Chemistry C
Shire T
(2020)
DEM simulations of polydisperse media: efficient contact detection applied to investigate the quasi-static limit
in Computational Particle Mechanics
Shi G
(2020)
A Fluid-Structure Interaction Study on a Bionic Fish Fin With Non-Uniform Stiffness Distribution
in Journal of Offshore Mechanics and Arctic Engineering
Sansom HC
(2021)
Highly Absorbing Lead-Free Semiconductor Cu2AgBiI6 for Photovoltaic Applications from the Quaternary CuI-AgI-BiI3 Phase Space.
in Journal of the American Chemical Society
Sahoo L
(2021)
Unravelling charge-transfer in Pd to pyrrolic-N bond for superior electrocatalytic performance
in Journal of Materials Chemistry A
Plekhanov E
(2021)
Calculating dynamical mean-field theory forces in ab initio ultrasoft pseudopotential formalism
in Physical Review B
Park KHK
(2022)
Collective Synthesis of Illudalane Sesquiterpenes via Cascade Inverse Electron Demand (4 + 2) Cycloadditions of Thiophene S,S-Dioxides.
in Journal of the American Chemical Society
Papamarkou T
(2023)
Approximate blocked Gibbs sampling for Bayesian neural networks
in Statistics and Computing
Pal P.
(2021)
The University of Edinburgh's Bengali-Hindi Submissions to the WMT21 News Translation Task
in WMT 2021 - 6th Conference on Machine Translation, Proceedings
Osiecki J
(2022)
Periodic corner holes on the Si(111)-7×7 surface can trap silver atoms
in Nature Communications
Nugent J
(2019)
A General Route to Bicyclo[1.1.1]pentanes through Photoredox Catalysis
in ACS Catalysis
Neale S
(2020)
Accurate computed spin-state energetics for Co( iii ) complexes: implications for modelling homogeneous catalysis
in Dalton Transactions
Description | It has been possible to operate the EPSRC Tier 2 HPC service for the benefit of both academic scientific research and industrial use. In particular, over the past 3 years we have had around 80 industry users annually on the system, either using it in collaborative projects with academic researchers or on a pay-per-use basis. |
Exploitation Route | If appropriate support for users is provided, coupled to a properly skilled business development team, any HPC centre should be able to develop local support for businesses on their HPC service. |
Sectors | Aerospace, Defence and Marine,Agriculture, Food and Drink,Chemicals,Construction,Creative Economy,Digital/Communication/Information Technologies (including Software),Energy,Financial Services, and Management Consultancy,Healthcare,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology |
Description | EPCC's Tier 2 HPC Service - called Cirrus - has provided access to a number of companies over the past five years to help them improve their products and services. Many of these companies have come from the manufacturing and engineering sectors. Cirrus also formed part of the EC-funded Fortissimo Cloud of HPC Resources which provided pay-on-demand access to its resources for companies across Europe. A small company, the Fortissimo Marketplace Ltd has been established with the PI as Managing Director which markets Cirrus and other Fortissimo Cloud systems to companies across Europe. |
First Year Of Impact | 2017 |
Sector | Aerospace, Defence and Marine,Agriculture, Food and Drink,Construction,Digital/Communication/Information Technologies (including Software),Energy,Environment,Financial Services, and Management Consultancy,Healthcare,Manufacturing, including Industrial Biotechology,Pharmaceuticals and Medical Biotechnology,Transport |
Impact Types | Economic |
Description | IAA Grant to Accelerate the impact of Cirrus uptake across Scottish Manufacturing SMEs via outreach in collaboration with CEED |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | Regional |
Primary Audience | Industry/Business |
Results and Impact | We received an Impact Accelerator Activity award to allow us to promote the Cirrus Tier 2 service to Scottish Manufacturing SMEs. The work was undertaken in collaboration with CEED - an organisation representing Scottish Manufacturing and Engineering companies. |
Year(s) Of Engagement Activity | 2017 |