Computing Services

The department behind IT services at the University of Bath

Topic: High Performance Computing (HPC)

EPSRC launches six new Tier2 HPC Centres

  , , , , , , , , , ,

📥  Computing Services, High Performance Computing (HPC), Research

The Engineering and Physical Sciences Research Council (EPSRC) has just launched six new High Performance Computational (HPC) Centres, worth a combined £20 million, at the Thinktank Science museum in Birmingham. These new Tier2 regional centres are aimed at supporting both academics and industry providing access to a diverse range of powerful supercomputers for scientific research and engineering.

These Tier2 systems will sit between the National Tier1 (e.g. ARCHER) and local campus Tier3 systems (e.g. Balena), addressing the gap in capability between these two levels. This new layer of Tier2 HPC will enable new discoveries and drive innovation, it will be open to any UK ESPRC researcher and provide easy local access, and be integrated with the HPC ecosystem across the UK, both vertically into Tier1 and Tier3 systems and horizontally to other Tier2 centres.  The Tier2 Centres will provide access to new and advanced technologies such as Intel’s Knights Landing (KNL) Xeon Phi, NVIDIAs Pascal, POWER8, 64-bit ARM and data burst buffers. To complement this minefield of diverse technologies, all six Centres will provide support though Research Software Engineers (RSEs) to assist with the training and skills development, porting and optimisation of applications and codes.

 

Great Western 4 (GW4) HPC Centre for Advanced Architectures: Isambard
Isambard, after Victorian engineer Isambard Kingdom Brunel, will be the first of its kind 64-bit ARM-based supercomputer providing multiple advanced architectures within the same system to enable evaluation and comparison across a diverse range of hardware platforms in a production environment. This is a joint project between the GW4 Alliance (universities of Bath, Bristol, Cardiff and Exeter) together with Cray Inc. and the Met Office. The service will provide over 10,000 high performance ARMv8 cores, as well as NVIDIA P100 and Intel Xeon Phi (KNL) accelerators cards.
http://gw4.ac.uk/isambard/

A National Facility for Peta-scale Data Intensive Computation and Analytics: Peta-5
Peta-5, a multi-disciplinary facility providing large-scale data simulation and high performance data analytics designed to enable advances in material science, computational chemistry, computational engineering and health informatics. This project, led by University of Cambridge, will provide peta-flop performance across Intel’s KNL, GPU, x86 computational architectures and peta-scale storage on spinning disk and tape storage.

Hub in Materials and Molecule Modelling: Thomas
Thomas, after the polymath Thomas Young, a Materials and Molecular Modelling (MMM) Hub will have applications in energy, healthcare and the environment. The project is led by UCL, with partners in the Materials Chemistry Consortium (MCC) and UKCP, and will provide a large x86-64 based system.
https://www.ucl.ac.uk/news/news-articles/1116/021116-ucl-hpc-hub-materials-science

Joint Academic Data science Endeavour: JADE
JADE, led by the University of Oxford, will provide the largest GPU-based system in the UK. Working with NVIDIA, will provide around 3.7 Petaflops of performance using NVIDIA’s latest DGX-1 platform, which makes use of the new NVlink technology.  This centre will be optimised for Deep Learning which will greatly benefit research applications involved in machine learning, image/video/audio analysis and molecular dynamics.
http://www.ox.ac.uk/news/2017-03-30-£3m-awarded-oxford-led-consortium-national-computing-facility-support-machine

HPC Midlands Plus
HPC Midlands Plus will primarily focus on data-intensive applications in fields ranging from engineering, manufacturing, healthcare and energy. Led by Loughborough University, the Centre will provide a large x86 based system, to accompany this will be a modest sized component of POWER8 systems each with 1TB memory.
http://www.lboro.ac.uk/media-centre/press-releases/2017/february/32m-funding-for-midlands-based-high-performance-computing-centre.html

Edinburgh Parallel Computing Centre (EPCC): Cirrus
Cirrus, will be hosted by the Edinburgh Parallel Computing Centre (EPCC). It will offer over 10,000 x86 cores to both science and industry. This will nicely complement EPSRC’s National HPC service, ARCHER. As part of this project, EPCC will host a mini-RDF (Research Data Facility) to provide common object-based data store to the Tier2 ecosystem and other supercomputers.
https://www.epcc.ed.ac.uk/facilities/cirrus

 

Recent press releases about the launch:

  1. https://www.epsrc.ac.uk/newsevents/news/sixhpccentresofficiallylaunch
  2. https://www.top500.org/news/uk-antes-up-20-million-for-six-new-supercomputer-centers
  3. http://www.bath.ac.uk/news/2017/03/30/gw4-world-first-supercomputer-launched-at-national-exhibition

 

 

Save the date - 6th Annual Symposium on HPC on 12th June 2017

  , , , , ,

📥  High Performance Computing (HPC), Research

We are pleased to announce that the 6th Annual Symposium on High Performance Computing will be healed on Monday 12th June 2017.

This symposium organized by the University of Bath we will bring together staff, researchers and students working with HPC in the different areas of science and engineering along with other invited specialists of the field.

Save the date in your agenda and don’t miss the opportunity for networking and get in touch with the exciting developments on HPC.

Best regards,
The Organising Committee

 

Balena maintenance - 2nd February 2017

  , , , ,

📥  High Performance Computing (HPC)

During Inter-Semester break, on Thursday 2 February between 09:00 and 13:00, we will be placing the cluster into maintenance mode whilst we perform failover tests between the pair of master nodes and BeeGFS node couplets.

These tests will ensure that pairs of master nodes and BeeGFS node couplets are in good working order should an unexpected system issue occur that triggers a system failover.

While the system failovers are being tested, all users will be able to access data from /home and /beegfs, but you may notice a momentary freeze while the storage areas are transferred between the failover pairs. All compute nodes will be placed into a scheduler reservation to prevent any workloads from running while these tests are carried out.

Sorry for the short notice of this announcement, I hope this will not cause too much disruption for anyone.

Team HPC

 

GW4 joins industry partners to develop ‘first of its kind’ supercomputer

  , , , , , , , ,

📥  High Performance Computing (HPC), Research

GW4 Alliance, together with Cray Inc. and the Met Office, has been awarded £3m by EPSRC to deliver a new Tier 2 high performance computing (HPC) service for UK-based scientists. This unique new service, named ‘Isambard’ after the renowned Victorian engineer Isambard Kingdom Brunel, will provide multiple advanced architectures within the same system in order to enable evaluation and comparison across a diverse range of hardware platforms. It will also provide a service to the community that will enable algorithm development and the porting of scientific codes. Isambard will include 10,000+ ARMv8 64-bit cores, in addition to a smattering of x86 CPUs, Intel Knights Landing Xeon Phi processors, and NVIDIA P100 GPUs.
For more information, please see the below press releases:

Bath's press release:
http://www.bath.ac.uk/research/news/2017/01/17/gw4-joins-industry-partners-to-develop-%E2%80%98first-of-its-kind%E2%80%99-supercomputer/

GW4 press release:
http://gw4.ac.uk/news/gw4-joins-industry-partners-to-develop-first-of-its-kind-supercomputer/

 

Mathematicians use Balena to model global air pollution

  , , , , , , ,

📥  High Performance Computing (HPC), Research

Dr Gavin Shaddick, of the University of Bath’s Department of Mathematical Sciences and Deputy Director of the Bath Institute for Mathematical Innovation, has been leading an international research team, which includes the WHO, and have used the University of Bath's High Performance Computing (HPC) system, Balena, to investigate global air pollution.

The results show that nine out of ten people (92 per cent) on Earth live in places where air pollution is higher than acceptable limits – even when they are outside.

For more information visit:

http://www.bath.ac.uk/news/2016/09/27/air-pollution/

http://www.wired.co.uk/article/uk-air-pollution-levels-london

http://www.bath.ac.uk/research/news/2016/09/21/air-pollution/

 

 

Balena Maintenance - 8th to 12th August 2016

  , ,

📥  High Performance Computing (HPC)

The maintenance work will begin on Monday 8th August and is expected to take up to a week to complete. During this maintenance window there will be no access to the Balena system and all queued jobs will need to be cleared from the scheduler.

The majority of this work will be performed by ClusterVision. We are anticipating needing a full week to give ClusterVision and ourselves enough time to complete these maintenance tasks. We shall open up access once all disruptive tasks have been completed.

Below is a list of some of the maintenance work which will be taking place:

  • Upgrading the SLURM scheduler, security patching and enabling new features
  • Testing SLURM's node power management
  • Enabling global file locking on the BeeGFS scratch partition
  • ClusterVision will also be configuring new system monitoring tools

Regards,
Team HPC

 

HPC for researchers by Tobias Brewer

  , , , , , , ,

📥  High Performance Computing (HPC), Research

The annual symposium at Bath brings together people from across the University to discuss how they use High Performance Computing (HPC) in their research. This year, the event included talks from several departments as well as keynote speakers from Intel and Cardiff University. People with projects from many diverse research areas attended and presented their work on topics such as data sharing in biology, computational screening of inorganic materials and topological optimization. All of these projects benefit hugely from HPC.

My PhD research centres around the efficient simulation of rare events and determining the probability of their occurrence. How likely is it that a tsunami will strike this coastal city? What was the probability of this genetic mutation? Could this village be destroyed by a lava flow? To answer questions like these requires computer codes that concurrently simulate a very large number of sophisticated systems. The C++ program's that perform these simulations take a ludicrously long time to run. The University's HPC computing system, Balena, allows these computations to be sped up dramatically, often by a factor of more than 20. This makes it simpler to consider more sophisticated systems and to perform analysis of much rarer events. A simulation that would have otherwise taken a month to run now takes less than two days.

Furthermore, with Balena facilitating the execution of several codes simultaneously, researchers can obtain results from many different simulations at the same time. Balena has a huge amount of software and hardware available for researchers to employ: various compilers, a SLURM scheduler and MPI libraries for parallelizing code. Of particular use to my research are the Allinea profiling tools which make it easy to identify bottlenecks and bugs and to pinpoint memory leaks. The system also enables the use of accelerator cards, such as the Intel Xeon Phi's, which can reduce computation time even further. I'd encourage any researcher at Bath to utilise Balena for its wide range of computational tools and for the considerable amount of time that it can save you in your research.

 

 

 

 

5th Annual HPC Symposium was a great success

  , , , , , , ,

📥  High Performance Computing (HPC), Research

Attendees of the 5th HPC Symposium

We would like to thank everyone who joined us at the 5th Annual Bath HPC Symposium yesterday. Following the trend from previous years, the event was a great success for the University.

This year represented the fifth symposium and the event was attended by over 50 participants from within the university, collaborators, as well as external partners including Intel. The schedule for the day included contributions from across the University: physics, maths, chemistry, biology; mech-eng, elec-eng and chem-eng; management and architecture. These contributions showcased the sheer variety of innovative HPC work being done across the University, covering including molecular dynamics and quantum-chemical simulations through to high level finite-element methods to solve engineering problems. The Symposium provided a great opportunity for networking and for learning about new developments in hardware, software and applications.

We would like to say a big thankyou to the contributors, attendees, keynote speakers, session chairs and ground crew, who together made this meeting a day to remember.

Also, we would like to congratulate our two prize winners, Tobias Brewer and Will Saunders, once again for their outstanding poster presentation and contributed talk.

Best wishes,
Steven, Jonathan and Roshan

 

5th Annual HPC Symposium - Keynote Speakers

  , , , , , , , , , , , , , , , ,

📥  High Performance Computing (HPC), Research

We are pleased to announce two keynote talks for the upcoming 5th Annual Bath HPC Symposium on Thursday 9th June.

The first keynote will be given by Dr Tom Connor from Cardiff University, titled "The Cloud Infrastructure for Microbial Bioinformatics (CLIMB); overcoming barriers to software use and data sharing in biology".

Genome sequencing has made it possible to examine fundamental biological questions over vast scales; from bacteria to man. Since the first bacterial genome was published 20 years ago, research focused around bacterial pathogens has been at the very vanguard of the revolution that has transformed biology into a data rich, data intensive field of science. Thomas will discuss how cloud approaches are being used in the MRC’s £8.5m CLIMB project to overcome key challenges that exist for most biologists today.

The second keynote will be delivered by Andy Mallinson from Intel. Andy will give us an overview of the Xeon Phi (MIC) architecture, some tips and tricks for developing, optimising and porting HPC codes to use them, a glimpse of Intel's future roadmap and the benefit these development will have for next-generation HPC systems.

We would like to take this opportunity to remind those who have not yet registered to do so, and to encourage participants to submit an abstract for a contributed talk or poster on their research.

Registration is currently open at https://bathhpc-2016.eventbrite.co.uk (deadline: Friday 27th May), and information about abstract submission can be found on the symposium website at http://go.bath.ac.uk/hpc-symposium (deadline: today, Friday 13th May).

We will be in touch again with further updates soon.

Best wishes,
Jonathan and Steven

 

 

"HPC and Accelerator Cards" by Martin Schreiber, 29th April 2016

  , , , , , , , ,

📥  High Performance Computing (HPC), Research

Next Friday (29th of April) Martin Schreiber (Exeter, see https://emps.exeter.ac.uk/mathematics/staff/ms698 ) will give a guest lecture for our Scientific Computing course in the maths department, and I thought this might be of interest for some of you.

The title of Martin's talk is:

"High Performance Computing and Accelerator Cards"

The talk takes place from 11:15h - 12:05h in the Wolfson Lecture Theatre (4West 1.7).

See you next week,
Eike