{"language":[{"iso":"eng"}],"date_updated":"2024-04-09T07:12:06Z","title":"Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL","publication":"arXiv:2403.18374","abstract":[{"lang":"eng","text":"Most FPGA boards in the HPC domain are well-suited for parallel scaling because of the direct integration of versatile and high-throughput network ports. However, the utilization of their network capabilities is often challenging and error-prone because the whole network stack and communication patterns have to be implemented and managed on the FPGAs. Also, this approach conceptually involves a trade-off between the performance potential of improved communication and the impact of resource consumption for communication infrastructure, since the utilized resources on the FPGAs could otherwise be used for computations. In this work, we investigate this trade-off, firstly, by using synthetic benchmarks to evaluate the different configuration options of the communication framework ACCL and their impact on communication latency and throughput. Finally, we use our findings to implement a shallow water simulation whose scalability heavily depends on low-latency communication. With a suitable configuration of ACCL, good scaling behavior can be shown to all 48 FPGAs installed in the system. Overall, the results show that the availability of inter-FPGA communication frameworks as well as the configurability of framework and network stack are crucial to achieve the best application performance with low latency communication."}],"_id":"53364","status":"public","date_created":"2024-04-09T07:04:46Z","external_id":{"arxiv":["2403.18374"]},"author":[{"first_name":"Marius","id":"40778","full_name":"Meyer, Marius","last_name":"Meyer"},{"last_name":"Kenter","full_name":"Kenter, Tobias","first_name":"Tobias","id":"3145"},{"first_name":"Lucian","full_name":"Petrica, Lucian","last_name":"Petrica"},{"last_name":"O'Brien","full_name":"O'Brien, Kenneth","first_name":"Kenneth"},{"first_name":"Michaela","last_name":"Blott","full_name":"Blott, Michaela"},{"last_name":"Plessl","full_name":"Plessl, Christian","first_name":"Christian","id":"16153","orcid":"0000-0001-5728-9982"}],"year":"2024","user_id":"40778","citation":{"mla":"Meyer, Marius, et al. “Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL.” <i>ArXiv:2403.18374</i>, 2024.","chicago":"Meyer, Marius, Tobias Kenter, Lucian Petrica, Kenneth O’Brien, Michaela Blott, and Christian Plessl. “Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL.” <i>ArXiv:2403.18374</i>, 2024.","apa":"Meyer, M., Kenter, T., Petrica, L., O’Brien, K., Blott, M., &#38; Plessl, C. (2024). Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL. In <i>arXiv:2403.18374</i>.","short":"M. Meyer, T. Kenter, L. Petrica, K. O’Brien, M. Blott, C. Plessl, ArXiv:2403.18374 (2024).","ama":"Meyer M, Kenter T, Petrica L, O’Brien K, Blott M, Plessl C. Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL. <i>arXiv:240318374</i>. Published online 2024.","bibtex":"@article{Meyer_Kenter_Petrica_O’Brien_Blott_Plessl_2024, title={Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL}, journal={arXiv:2403.18374}, author={Meyer, Marius and Kenter, Tobias and Petrica, Lucian and O’Brien, Kenneth and Blott, Michaela and Plessl, Christian}, year={2024} }","ieee":"M. Meyer, T. Kenter, L. Petrica, K. O’Brien, M. Blott, and C. Plessl, “Optimizing Communication for Latency Sensitive HPC Applications on up to 48 FPGAs Using ACCL,” <i>arXiv:2403.18374</i>. 2024."},"project":[{"_id":"52","name":"PC2: Computing Resources Provided by the Paderborn Center for Parallel Computing"}],"type":"preprint","department":[{"_id":"27"}]}