[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-117973-en":3,"doc-seo-117973-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":4,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":27,"seo_description":14,"update_tm":28,"read_time":29},117973,3848291630094,"Emma Wilson","https://eur-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45",8,"Research & Report","Leveraging Intra-Function Parallelism in Serverless Machine Learning","Running stateful machine learning on serverless architectures introduces overheads because serverless functions are not directly network-addressable and must rely on remote storage for shared state. This work applies intra-function parallelism to exploit multicore resources and port two stateful algorithms—k-means clustering and logistic regression—to serverless. Experiments show that intra-function parallelism improves serverless learning performance, with up to 68% gains for k-means. The results also indicate that fewer workers with multiple vCPUs outperform many single-vCPU workers due to reduced synchronization costs.","This is a repository copy of Leveraging Intra-Function Parallelism in Serverless Machine Learning.  \nWhite Rose Research Online URL for this paper:  \n[https://eprints.whiterose.ac.uk/208734/](https://eprints.whiterose.ac.uk/208734/)  \n[Version: Accepted Version](Version: Accepted Version)  \nProceedings Paper:  \nPredoaia, [Ionut orcid.org/0000-0002-2009-4054 and Garc](Ionut orcid.org/0000-0002-2009-4054 and Garc) ía-López, Pedro (2023)  \nLeveraging Intra-Function Parallelism in Serverless Machine Learning. In: WoSC '23: Proceedings of the 9th International Workshop on Serverless Computing. 9th International Workshop on Serverless Computing, WoSC '23, 11-15 Dec 2023 WoSC '23 . ACM , ITA , 36–41.  \n[https://doi.org/10.1145/3631295.3631399](https://doi.org/10.1145/3631295.3631399)  \nReuse  \nItems deposited in White Rose Research Online are protected by copyright, with all rights reserved unless indicated otherwise. They may be downloaded and/or printed for private study, or other acts as permitted by national copyright laws. The publisher or other rights holders may allow further reproduction and re-use of the full text version. This is indicated by the licence information on the White Rose Research Online record for the item.  \nTakedown  \nIf you consider content in White Rose Research Online to be in breach of UK law, please notify us by  \nemailing [eprints@whiterose.ac.uk](eprints@whiterose.ac.uk) including the URL of the record and the reason for the withdrawal request.  \n[eprints@whiterose.ac.uk](eprints@whiterose.ac.uk)[ ](eprints@whiterose.ac.uk)[https://eprints.whiterose.ac.uk/](https://eprints.whiterose.ac.uk/)  \nLeveraging Intra-Function Parallelism in Serverless Machine Learning  \nIonut Predoaia  \nUniversity of York York, United Kingdom [ionut.predoaia@york.ac.uk](ionut.predoaia@york.ac.uk)  \nPedro García-López  \nUniversitat Rovira i Virgili Tarragona, Spain [pedro.garcia@urv.cat](pedro.garcia@urv.cat)  \nAbstract  \nRunning stateful machine learning algorithms with serverless architectures inherently induces overheads, as serverless functions are not directly network-addressable, hence one must rely on a remote storage service for storing the shared state. To hide the access latency to the remote storage, one can employ intra-function parallelism to take advantage of the multicore computing resources of the serverless functions. In this work, we port to serverless two stateful machine learning algorithms, k-means clustering and logistic regression, and then adopt intra-function parallelism to parallelize the execution of the serverless functions. Several experiments have demonstrated that intra-function parallelism delivers performance improvements in serverless machine learning. Improved performances of up to 68% have been achieved when running k-means on serverless functions that employ intra-function parallelism. We demonstrate with k-means and logistic regression that from a performance perspective it is preferable to execute a smaller number of multiple-vCPUs workers than a larger number of singlevCPU workers, due to decreased synchronization overheads.  \nCCS Concepts: · Computing methodologies → Machine learning; Parallel computing methodologies; · Computer systems organization → Cloud computing.  \nKeywords: Serverless, Machine Learning, Intra-Function Parallelism, Multicore Functions, Stateful, Lithops  \nACM Reference Format:  \nIonut Predoaia and Pedro García-López. 2023. Leveraging IntraFunction Parallelism in Serverless Machine Learning. In 9th International Workshop on Serverless Computing (WoSC’23), December 11􀀕15, 2023, Bologna, Italy. ACM, New York, NY, USA, 6 pages.  \n[https://doi.org/10.1145/3631295.3631399](https://doi.org/10.1145/3631295.3631399)  \nPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation","cbCailb9DkArWzPI","https://ap.wps.com/l/cbCailb9DkArWzPI","pdf",418093,1,7,"English","en",105,"# Abstract\n# Introduction\n## Shared-state and overheads in serverless ML\n## Communication and synchronization costs (k-means, logistic regression)\n# Methods\n## Porting stateful algorithms to serverless\n## Applying intra-function parallelism\n# Experimental Evaluation\n## Performance improvements and up to 68% gains\n## Worker scaling: multiple vCPUs vs single vCPU\n# Conclusion\n## Key findings and practical guidance","[{\"question\":\"Why do stateful machine learning algorithms incur overhead on serverless architectures?\",\"answer\":\"Serverless functions are not directly network-addressable, so shared iterative state must be stored and accessed via remote storage over the network. This creates communication overheads and, for synchronization-based algorithms, additional synchronization overhead.\"},{\"question\":\"How does intra-function parallelism help serverless machine learning?\",\"answer\":\"Intra-function parallelism leverages multicore resources within a single serverless function to reduce the impact of remote storage access latency and improve overall execution performance.\"},{\"question\":\"What performance trend is observed for k-means with intra-function parallelism?\",\"answer\":\"Experiments demonstrate performance improvements for serverless ML, with improved results of up to 68% when running k-means functions that employ intra-function parallelism.\"}]","Leveraging Intra-Function Parallelism in Serverless Machine Learning | PDF",1785680599,18,{"code":4,"msg":31,"data":32},"ok",{"site_id":24,"language":23,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"leveraging-intra-function-parallelism-in-serverless-machine-learning","",{"@graph":36,"@context":85},[37,54,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":53},"https://docshare.wps.com/document/leveraging-intra-function-parallelism-in-serverless-machine-learning/117973/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":23,"description":14,"dateModified":62,"datePublished":62,"encodingFormat":61,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":41,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-02",true,{"@type":65,"interactionType":66,"userInteractionCount":4},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"Why do stateful machine learning algorithms incur overhead on serverless architectures?","Question",{"text":75,"@type":76},"Serverless functions are not directly network-addressable, so shared iterative state must be stored and accessed via remote storage over the network. This creates communication overheads and, for synchronization-based algorithms, additional synchronization overhead.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does intra-function parallelism help serverless machine learning?",{"text":80,"@type":76},"Intra-function parallelism leverages multicore resources within a single serverless function to reduce the impact of remote storage access latency and improve overall execution performance.",{"name":82,"@type":73,"acceptedAnswer":83},"What performance trend is observed for k-means with intra-function parallelism?",{"text":84,"@type":76},"Experiments demonstrate performance improvements for serverless ML, with improved results of up to 68% when running k-means functions that employ intra-function parallelism.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,119,122,127,130,134],{"id":20,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":21,"doc_module":4,"doc_module_name":46,"category_name":116,"show_sort_weight":117,"slug":118},"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":120,"slug":121},30,"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]