Transparent extension of INFN-T1 with heterogeneous computing architectures
<p><span>The INFN-CNAF Tier-1 is engaged for years in a continuous effort to integrate</span><br><span>its computing centre with more typologies of computing resources. In particular, the challenge</span><br><span>of providing opportunistic access to nonstandard CPU architectures, such as PowerPC, ARM</span><br><span>or hardware accelerators (GPUs) has been actively exploited.</span> <span>In this work, we describe a</span><br><span>solution to transparently integrate access to ppc64 and GPUs.</span> <span>This solution has been tested</span><br><span>to transparently extend the INFN-T1 Grid computing centre with Power9 based machines and</span><br><span>V100 GPUs from the Marconi 100 HPC cluster managed by CINECA, and eventually used for</span><br><span>the next CINECA machine, Leonardo. We also discuss further possible improvements and how</span><br><span>this will meet requirements and future plans for the new tecnopolo centre, where the CNAF</span><br><span>Tier-1 will be hosted soon.</span></p>
Publication Details
- Journal
- INFM-OAR (INFN Catania)
- Published
- 2024-09-02
- DOI
- https://doi.org/10.15161/oar.it/51zr5-w9317
- Primary Topic
- Parallel Computing and Optimization Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00