High-Performance Computing 2019-20

researcher at the Fondazione Bruno Kessler. Te HPC infrastructure allows the team

to integrate information on global climatic variables ‘as they arrive’, according to Poletti and Guzzetta, ‘refining the estimates of risk in almost real-time, from static seasonal maps describing the general risk of a given area at a given time of year, to dynamically changing maps describing the specific risk of the area under current climatic conditions.’

The time is now

Te next step for the Vestec project is to specify detailed execution plans for these use cases to prove the efficiency of its architectural design and also specify its KPIs. However, Brown said there is scope for ‘more than the focus of a single project’ with the team organising workshops to bring the HPC community together to further understand how to use supercomputers for urgent decision making. He concluded: ‘Tere are numerous

people doing work in and around this area, and [we need to share] our knowledge and experience to better understand how we can drive our own activities, and that of the HPC community forward in this area.’ n

Say Cheese

A new European-funded Center of Excellence for Exascale in Solid Earth (Cheese) is preparing ten open-source flagship codes to work efficiently on pre-exascale and upcoming exascale computers. Tese codes on seismic, tsunami, volcanic

and magnetohydrodynamics modelling will then be used in 12 pilot demonstrators. Tese are intended as proofs of concept to address a range of exascale computational challenges, and will eventually become services on urgent computing, hazard assessment, and early warning forecast. Tere are challenges from both a data and

complexity standpoint, as Arnau Folch, Cheese co-ordinator at the Barcelona Supercomputing Centre, explained: ‘As occurs in other scientific disciplines, community codes in Solid Earth are complex applications built over [many] years by different academic and research groups, oſten prioritising scientific results over soſtware engineering and, in some cases, eventually yielding to legacy codes.’ Te research teams have ‘quickly shiſted

from the ‘do it yourself’ times to a much transversal scenario, where multiple actors must co-operate,’ according to Folch, who

said: ‘Cheese includes a co-design component, but the alignment of science and industry, oſten with different timings and roadmaps, is a major challenge we face in the path towards exascale.’ ‘A second challenging aspect is the real-time

integration of large volumes of edge-generated data and model execution and post-process, which requires complex workflows,’ Folch added. Te project’s pilot demonstrators are being

designed in collaboration with the Italian and Icelandic civil protections, once physics-based ground shaking (impact) maps are generated shortly aſter an earthquake or real-time tsunami inundation forecasts. However, the project’s potential extends

far beyond these 12 use cases, once the right protocols are in place. Folch said: ‘Tese examples can also pave the road for urgent computing in other domains, from fire propagation to atmospheric or marine toxic dispersals. However, an important gap needs to be filled in terms of rules, protocols, triggering mechanisms and (limited) access to results before urgent computing becomes operational.’

HPC Yearbook 19/20

SoftIron creates the world’s finest unified storage solutions for the HPC data center

Our HyperDrive® platform is custom-built

to optimize Ceph for high-performance, scale-out, wire-speed storage.

Read more on page 10 about how Ceph is powering the most exciting research projects on Earth. Then, visit our website to discover how HyperDrive and our Ceph Storage Management tool will save time and money in your HPC data center.

For more information, visit

+1 855 450 CEPH (2374)

@SoftIron @SoftIron @SoftIronNews

See HyperDrive in action at SC19! Meet us and learn more at booth #1594 7

Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32