N8 HPC Helps to Improve Animal Simulations

 Professor Bill Sellers from the School of Earth and Environmental Sciences, University of Manchester has used N8 HPC to improve the accuracy of computer-generated animal simulations.

The research, recently published by the Royal Society Open Science Journal, shows how simple changes to ‘machine learning’ algorithms can produce better looking, more accurate computer-generated animal simulations.  It will also help researchers investigate the ‘curious way’ that all primates walk and how this might be linked to stability whilst moving through the trees.

Professor Bill Sellers says: “Starting from an animal’s skeleton, computers using machine learning can now reconstruct how the animal could have moved. However, they don’t always do a good job”.

“But with some simple changes to the machine learning goals we can now create much more accurate simulations. We’ve now used this process to generate chimpanzee locomotion to explore why they walk the way they do”.

A full body CT scan of an adult male common chimpanzee was used to create a chimpanzee model.  This scan was then used to generate a skeletal model and a skin outline which was then used to define joint positions, muscle paths and limb contact points for the simulation.

Prof Sellers explains: “As technology has advanced and with musculoskeletal models becoming increasingly sophisticated, previous simulation models are becoming extremely unrealistic in relation to gait patterns so we have to adapt the way we think and research.”

The Future of N8 HPC

2018 will see a period of change within N8 HPC.  The current machine, Polaris is five years old, increasingly expensive to maintain and is comprised of dated hardware –  it is approaching end of life and is due to be switched off on the 17th July 2018.

The 8 member universities of N8 HPC have been in consultation for the past year as to how to proceed after this point.  Three universities (Lancaster, Manchester and York) are planning to jointly fund and procure a new cluster.  This will be hosted at Durham, who will receive 5% of available resources in return.  There will also be an investment in cloud-based resources by Manchester and Leeds.  (Sharing of resources will allow users from Lancaster and York access to cloud cycles and users from Leeds access to the new cluster.)

Users at Liverpool, Newcastle and Sheffield are advised to contact their local N8 HPC TMG representative to discuss resource availability.

All 8 universities will be contributing to a N8 HPC Centre of Excellence which will accelerate progress in areas of research that are of strategic importance to the N8 partners, including some that have not made heavy use of technology in the past.  Led by academics, each area will identify software tools, methodology, expertise etc that could benefit each research area.  A research roadmap will be developed for each selected theme, identifying key opportunities for us to accelerate progress, proposing a joint programme of work, and building a community of practice.

Further announcements will be made on the N8 HPC Users mailing list, website and Twitter feed, regarding removing your data from Polaris, queue shutdown, etc., which will be before the middle of July.

If you have any questions about your institution’s involvement in N8 HPC going forward, please contact your Steering Group member.

HPC Calls Out Now

There are several EPSRC HPC (high performance computing) calls out at the moment which researchers should be aware of.  All calls have a closing date of the 22nd of March.

ARCHER

EPSRC offers access to ARCHER through calls for proposals to the Resource Allocation Panel (RAP). Users can request significant amounts (>1,000kAUs or >66,667 ARCHER core hours) of computing resource over a maximum 1 year period.

A non-exclusive list of eligible projects includes:

  • Short computational projects that do not warrant a full grant application
  • UK led collaborative projects with international and/or industry partners
  • Joint applications from students (as Co-I) with proven HPC experience (e.g. a successfully completed instant access project) and their PIs
  • Projects that link consecutive standard grant applications or that aid the preparation of a grant or fellowship application
  • Extended feasibility studies and trialling application developments at scale

ARCHER application website

Closing date – 22nd March for Technical Assessment

ARCHER Top-up

The aim of this call is to provide top-up resource on ARCHER for all existing EPSRC grant-holders for a maximum of two years.

ARCHER top-up application website

Closing date – 22nd March for Technical Assessment

TIER 2

EPSRC is offering open access to five Tier-2 High Performance Computing facilities through this call for proposals. The five facilities users can access through this call are: Cirrus, GW4, CSD3, HPC Midlands +, and JADE. Further details on each of these centres can be found in the call document on the EPSRC website. For details on how to access The Materials and Molecular Modelling Hub (MMM Hub), please see their website.

The aim of this call is to provide access to national Tier-2 HPC facilities for adventurous high-risk, high-reward projects that will benefit from the diversity of computing architectures available at Tier-2.

A non-exclusive list of eligible projects includes:

  • Short computational projects that do not warrant a full grant application;
  • UK led collaborative projects with international and/or industry partners;
  • Joint applications from students (as Co-Is) with proven HPC experience and their PIs;
  • Projects that link consecutive standard grant applications or that aid the preparation of a grant or fellowship application;
  • Extended feasibility studies and trialling application developments at scale;
  • High-risk, high-reward projects that would benefit from using novel architectures.

Tier 2 application website

Closing date – 22nd March for Technical Assessment

 

Tackling pollution control in India

Exposure to ambient air pollution causes over 4 million premature deaths each year, with 25% of this burden in India. Half of the world’s 20 most polluted cities are in India and poor air quality there is predicted to worsen in the future. Despite this importance there has been comparatively little air quality research focused on India and our knowledge of the sources and processes causing air pollution here is limited. It is critical to understand the contribution of different emission sources (e.g., traffic, industry) to ambient air pollution to design effective policies to reduce this substantial disease burden.

A study by Luke Conibear from Prof Dominick Spracklen’s group at the University of Leeds is the first to combine high resolution computer simulations with extensive observations to estimate the contribution of different emission sources to ambient air pollution and the related disease burden from exposure across India. The researchers used N8 HPC to run a numerical weather prediction model to simulate air quality over India. In their paper, recently published in Nature Communications, they demonstrate that combustion of solid fuels for household cooking and heating is the dominant source of ambient particulate matter pollution across India. Reducing emissions from this source provides much larger benefits to health compared to control of traffic or industry. Luke’s study provides the evidence needed to design effective pollution control efforts in India.

Modelling the Palaeoclimate with N8 HPC

When the supercontinent, Pangea, broke up during the Triassic and Jurassic periods, there were dramatic changes in Earth’s climate.  Prof David Schultz and his students used N8 HPC to run Palaeoclimate model simulations which revealed the spatial changes in climate between the Triassic and Jurassic.  Their results illustrate that the subtropics became slightly cooler and wetter despite the warming trend for the Earth’s average temperature.  To find out more about the use of “Build Your Own Earth”, please see the N8 HPC case study or the recently published paper in Geology Today.

Login Form