Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 17: Data, AI, Computing 2 (analysis tools)

T 17.8: Vortrag

Montag, 4. März 2024, 17:45–18:00, Geb. 30.34: LTI

Workflow Optimization for HEP Jobs on Opportunistic Resources with XRootD — •Robin Hofsaess, Manuel Giffels, Günter Quast, Maximilian Horzela, and Matthias Schnepf — Karlsruher Institut für Technologie

The WorldWide LHC Computing Grid (WLCG) is tailored to support High Energy Physics (HEP) workflows, offering an optimal environment for job execution. However, alongside the mostly homogeneous WLCG infrastructure, we increasingly rely on diverse 'opportunistic resources' such as High-Performance Computing (HPC) centers and cloud service providers. These resources are more heterogeneous and not inherently suited for HEP workflows that for example might require a high throughput of data. Since the future German HEP computing strategy foresees an increased share on such resources, we are currently working on concepts that not only aim to maximize computational throughput, but also align with the broader goals of sustainable resource utilization in an era of exponential data growth.

In this talk, I will present a first Proof of Concept implementation using an XRootD cache to increase the efficiency of HEP jobs on the HoreKa HPC cluster in Karlsruhe.

Keywords: XRootD; HPC; Workflow Optimization