Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service
Computing demands for large scientific experiments, such as the CMS experiment at the CERN LHC, will increase dramatically in the next decades. To complement the future performance increases of software running on central processing units (CPUs), explorations of coprocessor usage in data processing...
Main Authors: | , , , , , , , , , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Springer International Publishing
2024
|
Online Access: | https://hdl.handle.net/1721.1/156703 |
_version_ | 1810989992592277504 |
---|---|
author | Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Hussain, P. S. Jeitler, M. Krammer, N. Li, A. Liko, D. Mikulec, I. Schieck, J. Schöfbeck, R. Schwarz, D. |
author_facet | Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Hussain, P. S. Jeitler, M. Krammer, N. Li, A. Liko, D. Mikulec, I. Schieck, J. Schöfbeck, R. Schwarz, D. |
author_sort | Hayrapetyan, A. |
collection | MIT |
description | Computing demands for large scientific experiments, such as the CMS experiment at the CERN LHC, will increase dramatically in the next decades. To complement the future performance increases of software running on central processing units (CPUs), explorations of coprocessor usage in data processing hold great potential and interest. Coprocessors are a class of computer processors that supplement CPUs, often improving the execution of certain functions due to architectural design choices. We explore the approach of Services for Optimized Network Inference on Coprocessors (SONIC) and study the deployment of this as-a-service approach in large-scale data processing. In the studies, we take a data processing workflow of the CMS experiment and run the main workflow on CPUs, while offloading several machine learning (ML) inference tasks onto either remote or local coprocessors, specifically graphics processing units (GPUs). With experiments performed at Google Cloud, the Purdue Tier-2 computing center, and combinations of the two, we demonstrate the acceleration of these ML algorithms individually on coprocessors and the corresponding throughput improvement for the entire workflow. This approach can be easily generalized to different types of coprocessors and deployed on local CPUs without decreasing the throughput performance. We emphasize that the SONIC approach enables high coprocessor usage and enables the portability to run workflows on different types of coprocessors. |
first_indexed | 2024-09-23T12:30:43Z |
format | Article |
id | mit-1721.1/156703 |
institution | Massachusetts Institute of Technology |
language | English |
last_indexed | 2024-09-23T12:30:43Z |
publishDate | 2024 |
publisher | Springer International Publishing |
record_format | dspace |
spelling | mit-1721.1/1567032024-09-12T03:32:23Z Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Hussain, P. S. Jeitler, M. Krammer, N. Li, A. Liko, D. Mikulec, I. Schieck, J. Schöfbeck, R. Schwarz, D. Computing demands for large scientific experiments, such as the CMS experiment at the CERN LHC, will increase dramatically in the next decades. To complement the future performance increases of software running on central processing units (CPUs), explorations of coprocessor usage in data processing hold great potential and interest. Coprocessors are a class of computer processors that supplement CPUs, often improving the execution of certain functions due to architectural design choices. We explore the approach of Services for Optimized Network Inference on Coprocessors (SONIC) and study the deployment of this as-a-service approach in large-scale data processing. In the studies, we take a data processing workflow of the CMS experiment and run the main workflow on CPUs, while offloading several machine learning (ML) inference tasks onto either remote or local coprocessors, specifically graphics processing units (GPUs). With experiments performed at Google Cloud, the Purdue Tier-2 computing center, and combinations of the two, we demonstrate the acceleration of these ML algorithms individually on coprocessors and the corresponding throughput improvement for the entire workflow. This approach can be easily generalized to different types of coprocessors and deployed on local CPUs without decreasing the throughput performance. We emphasize that the SONIC approach enables high coprocessor usage and enables the portability to run workflows on different types of coprocessors. 2024-09-11T18:34:29Z 2024-09-11T18:34:29Z 2024-09-04 2024-09-08T03:08:37Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/156703 The CMS Collaboration., Hayrapetyan, A., Tumasyan, A. et al. Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service. Comput Softw Big Sci 8, 17 (2024). PUBLISHER_CC en https://doi.org/10.1007/s41781-024-00124-1 Computing and Software for Big Science Creative Commons Attribution https://creativecommons.org/licenses/by/4.0/ The Author(s) application/pdf Springer International Publishing Springer International Publishing |
spellingShingle | Hayrapetyan, A. Tumasyan, A. Adam, W. Andrejkovic, J. W. Bergauer, T. Chatterjee, S. Damanakis, K. Dragicevic, M. Hussain, P. S. Jeitler, M. Krammer, N. Li, A. Liko, D. Mikulec, I. Schieck, J. Schöfbeck, R. Schwarz, D. Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title | Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title_full | Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title_fullStr | Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title_full_unstemmed | Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title_short | Portable Acceleration of CMS Computing Workflows with Coprocessors as a Service |
title_sort | portable acceleration of cms computing workflows with coprocessors as a service |
url | https://hdl.handle.net/1721.1/156703 |
work_keys_str_mv | AT hayrapetyana portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT tumasyana portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT adamw portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT andrejkovicjw portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT bergauert portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT chatterjees portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT damanakisk portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT dragicevicm portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT hussainps portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT jeitlerm portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT krammern portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT lia portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT likod portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT mikuleci portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT schieckj portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT schofbeckr portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice AT schwarzd portableaccelerationofcmscomputingworkflowswithcoprocessorsasaservice |