Cardiff University expands HPC cluster with Lenovo

Expansion of university’s Hawk high-efficiency computing cluster promises to double the efficiency of some functions

Cliff Saran


Published: 24 Jan 2024 13: 07

Cardiff University has deployed 90 Lenovo ThinkSystem servers, with Logicalis expertise, to red meat up high-efficiency computing (HPC) be taught. The university’s Evolved Examine Computing @Cardiff (ARCCA) carrier labored with Logicalis and the Lenovo ThinkSystem platform to enhance the capabilities of its Hawk HPC cluster by up to 100%. The HPC will be at possibility of red meat up be taught into the origins of the universe and big title formation.

Lenovo has offered 75 Lenovo ThinkSystem SR645 servers before the fourth two-year gravitational wave detection slump of the Laser Interferometer Gravitational-Wave Observatory (LIGO) Scientific Collaboration. The servers provide a minimal efficiency enhance of 46% in contrast with a earlier HPC environment using the same hardware. Per Lenovo, for some functions, the unusual servers double the compute efficiency on hand to researchers. This uplift in efficiency enables researchers to task detection events and fragment recordsdata sooner.

The hardware contains Lenovo ThinkSystem SR645 servers powered by AMD Epyc processors, which connect to the Hawk cluster using Nvidia ConnectX-6 connectivity by Ubiquity, an birth offer cluster management draw offered by Logicalis.

Stephen Fairhurst, head of the gravity exploration institute at Cardiff University, acknowledged: “We’re getting even bigger per-core efficiency than we expected. Crucially, the unusual cluster became up and working in time for the most contemporary LIGO staring at slump. We are able to now task detection events sooner and can fragment the tips with astronomers sooner, serving to them discover sources of the signals more without tell.”

A 2d be taught community at Cardiff University learning huge title formation, led by Royal Society University be taught fellow Ana Duarte Cabral, will use 15 Lenovo ThinkSystem SR630 servers with two Lenovo ThinkSystem SR650 servers for storage. Funded by the Royal Society, the community’s be taught specializes within the creation of stars in spiral galaxies esteem our hang Milky System, requiring computationally annoying simulations of full galaxies.

“We’re getting even bigger per-core efficiency than we expected. We are able to now task detection events sooner and can fragment the tips with astronomers sooner, serving to them discover sources of the signals more without tell”
Stephen Fairhurst, Cardiff University

Commenting on the unusual HPC setup, Cabral acknowledged: “The unusual nodes will allow us to cascade down from the dimensions of entire galaxies to the dimensions of specific huge title-forming areas. With a dedicated cluster, we are able to seriously magnify the amount of recordsdata we are able to catch.”

The technical implementation affords simultaneous multi-threading, the attach Ubiquity is at possibility of mix the Lenovo hardware into the Hawk cluster. This has allowed ARCCA to automate management tasks, allowing be taught instrument engineers to focal level their efforts on science. Per Lenovo, the configuration balances impress to efficiency when dealing with high-throughput computing and parallelised workloads.

Discussing the advantages of Ubiquity, Martyn Visitor, director of Evolved Examine Computing at Cardiff University, acknowledged: “We desired to mix the unusual pc handy resource with the present Hawk cluster. When Logicalis showed us the capabilities of Ubiquity, we had been happy our purpose became achievable. Our ability to slither progressed computing instruments into a professionally managed HPC environment has allowed our be taught teams to safe R&D funding in Wales.”

Learn more on Clustering for high availability and HPC

Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button