[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-121438-en":3,"doc-seo-121438-105":29,"detail-sidebar-cat-0-en-105":89},{"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":20,"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":11},121438,962075114101,"Seraphina","https://ap-avatar.wpscdn.com/avatar/e000253a75eb197efd?x-image-process=image/resize,m_fixed,w_180,h_180&k=1780044092746381165",8,"Research & Report","Benchmarking Energy and Performance of Parallel Machine Learning Models Using Hardware and Software Power Meters - SAML-25 Workshop on Statistical and Machine Learning","Growing reliance on machine learning in healthcare, transportation, and finance increases the need for high-performance computing, where runtime speed is often prioritized while energy efficiency becomes an essential sustainability and cost factor. This study compares energy consumption and performance for four algorithms—K-means clustering, Ant Colony Optimization, Logistic Regression, and Random Search—in serial and parallel forms. Experiments run on an HPC testbed using both hardware and software power meters to measure energy usage. Parallel versions reduce execution time and also lower overall energy consumption compared with serial implementations.","Technological University Dublin  \nARROW@TU Dublin  \n\n| SAML-25 Workshop on Statistical and Machine Learning | Research Institutes/Centres/Groups |\n| --- | --- |\n| 2025-06-05\u003Cbr>Benchmarking Energy and Performance of Parallel Machine Learning Models Using Hardware and Software Power Meters\u003Cbr>Urooj Asgher\u003Cbr>Technological University Dublin, [b00168170@mytudublin.ie](b00168170@mytudublin.ie)\u003Cbr>Tania Malik\u003Cbr>Technological University Dublin, [tania.malik@tudublin.ie](tania.malik@tudublin.ie)\u003Cbr>Follow this and additional works at: [https://arrow.tudublin.ie/saml](https://arrow.tudublin.ie/saml)\u003Cbr> Part of the Statistics and Probability Commons |  |\n\nRecommended Citation  \nAsgher, Urooj and Malik, Tania, \"Benchmarking Energy and Performance of Parallel Machine Learning Models Using Hardware and Software Power Meters\" (2025) . SAML-25 Workshop on Statistical and Machine Learning. 6.  \n[https://arrow.tudublin.ie/saml/6](https://arrow.tudublin.ie/saml/6)  \nThis Conference Paper is brought to you by the EUt+ Academic Press a free to read and publish press of the European University of Technology.  \nThis work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.  \nBenchmarking Energy and Performance of Parallel Machine Learning Models Using Hardware and Software Power Meters  \nUrooj Asgher  \n[B00168170@mytudublin.ie](B00168170@mytudublin.ie)[ ](B00168170@mytudublin.ie)Technological University of Dublin Dublin, Ireland  \nDr. Tania Malik  \n[tania.malik@tudublin.ie](tania.malik@tudublin.ie)[ ](tania.malik@tudublin.ie)Technological University of Dublin Dublin, Ireland  \nABSTRACT  \nThe growing reliance on machine learning algorithms across domains such as healthcare, transportation, and finance has led to their increased deployment on high-performance computing platforms. While performance optimization remains a central concern, energy efficiency is emerging as a critical design consideration, particularly in light of global sustainability goals. This study presents a comparative analysis of the energy consumption and performance of serial and parallel implementations of four machine learning algorithms, K-means clustering, Ant Colony Optimization, Logistic Regression, and Random Search. Experiments were conducted on an HPC testbed using both hardware-based and software-based power meters to measure energy consumption. The results demonstrate that parallel implementations not only achieve substantial reductions in execution time but also lead to lower overall energy consumption compared to their serial counterparts.  \nKEYWORDS  \nHigh-Performance Computing, HPC, Energy Efficiency, Machine Learning, Parallel Algorithms, Energy Measurement  \n1 INTRODUCTION AND MOTIVATION  \nMachine learning (ML) algorithms are increasingly utilized in fields such as healthcare, finance, logistics, and infrastructure. With the growing complexity of these algorithms and the size of datasets, high-performance computing (HPC) systems are often employed to accelerate training and inference through parallel processing. Although these platforms significantly improve runtime efficiency, they raise concerns about energy consumption, which impacts both environmental sustainability and operational costs. Despite widespread efforts to optimize ML algorithms for accuracy and speed, their energy efficiency, particularly in comparing serial and parallel implementations, remains underexplored [5] .  \nIn this study, we evaluated the energy consumption and performance of four widely used ML algorithms, K-means clustering [1], Ant Colony Optimization (ACO) [3], Random Search (RS) [6], and Logistic Regression (LR) [4], in both serial and parallel versions. These algorithms were chosen because of their diverse computational patterns and widespread use in optimization and classification tasks. For instance, K-means is a centroid-based clustering method, ACO mimics swarm intelligence for solving combinatorial problems, RS offers a sim","cbCaicLhF3SfgxGa","https://ap.wps.com/l/cbCaicLhF3SfgxGa","pdf",473286,1,3,"English","en",105,"# Abstract\n# 1 Introduction and Motivation\n## High-performance computing and energy concerns\n## Algorithms selected for comparison\n## Hardware vs software energy measurement\n# 2 Energy Measurement Analysis and Experimental Insights of ML Applications\n## Dynamic energy consumption and baseline energy\n## Total energy and dynamic energy calculation","[{\"question\":\"Which machine learning algorithms are benchmarked in the study?\",\"answer\":\"The paper benchmarks K-means clustering, Ant Colony Optimization, Random Search, and Logistic Regression in both serial and parallel implementations.\"},{\"question\":\"How does the study measure energy consumption?\",\"answer\":\"It uses hardware-based power meters for real-time systemwide measurements and software-based tools that estimate CPU/DRAM energy using processor counters such as Intel RAPL.\"},{\"question\":\"What overall effect do parallel implementations have compared with serial ones?\",\"answer\":\"Parallel implementations achieve substantial reductions in execution time and lead to lower overall energy consumption than their serial counterparts.\"}]","Benchmarking Energy and Performance of Parallel Machine Learning Models Using Hardware and Software Power Meters - SAML-25 Workshop on Statistical and Machine Learning | PDF",1785735666,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":84,"head_meta":86,"extra_data":88,"updated_unix":28},"benchmarking-energy-and-performance-of-parallel-machine-learning-models-using-hardware-and-software-power-meters-saml-25-workshop-on-statistical-and-machine-learning","",{"@graph":35,"@context":83},[36,52,66],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,49],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":21},"https://docshare.wps.com/document/research-report/",{"item":50,"name":13,"@type":42,"position":51},"https://docshare.wps.com/document/benchmarking-energy-and-performance-of-parallel-machine-learning-models-using-hardware-and-software-power-meters-saml-25-workshop-on-statistical-and-machine-learning/121438/",4,{"url":50,"name":13,"@type":53,"author":54,"headline":13,"publisher":56,"fileFormat":59,"inLanguage":23,"description":14,"dateModified":60,"datePublished":60,"encodingFormat":59,"isAccessibleForFree":61,"interactionStatistic":62},"DigitalDocument",{"name":9,"@type":55},"Person",{"url":40,"name":57,"@type":58},"DocShare","Organization","application/pdf","2026-08-03",true,{"@type":63,"interactionType":64,"userInteractionCount":20},"InteractionCounter",{"@type":65},"ViewAction",{"@type":67,"mainEntity":68},"FAQPage",[69,75,79],{"name":70,"@type":71,"acceptedAnswer":72},"Which machine learning algorithms are benchmarked in the study?","Question",{"text":73,"@type":74},"The paper benchmarks K-means clustering, Ant Colony Optimization, Random Search, and Logistic Regression in both serial and parallel implementations.","Answer",{"name":76,"@type":71,"acceptedAnswer":77},"How does the study measure energy consumption?",{"text":78,"@type":74},"It uses hardware-based power meters for real-time systemwide measurements and software-based tools that estimate CPU/DRAM energy using processor counters such as Intel RAPL.",{"name":80,"@type":71,"acceptedAnswer":81},"What overall effect do parallel implementations have compared with serial ones?",{"text":82,"@type":74},"Parallel implementations achieve substantial reductions in execution time and lead to lower overall energy consumption than their serial counterparts.","https://schema.org",{"og:url":50,"og:type":85,"og:title":13,"og:site_name":57,"og:description":14},"article",{"robots":87,"canonical":50},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":90},[91,95,99,103,108,113,118,121,126,129,133],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":92,"show_sort_weight":93,"slug":94},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":96,"show_sort_weight":97,"slug":98},"Literature",80,"literature",{"id":51,"doc_module":4,"doc_module_name":45,"category_name":100,"show_sort_weight":101,"slug":102},"Exam",70,"exam",{"id":104,"doc_module":4,"doc_module_name":45,"category_name":105,"show_sort_weight":106,"slug":107},5,"Comic",60,"comic",{"id":109,"doc_module":4,"doc_module_name":45,"category_name":110,"show_sort_weight":111,"slug":112},6,"Technology",50,"technology",{"id":114,"doc_module":4,"doc_module_name":45,"category_name":115,"show_sort_weight":116,"slug":117},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":119,"slug":120},30,"research-report",{"id":122,"doc_module":4,"doc_module_name":45,"category_name":123,"show_sort_weight":124,"slug":125},9,"Religion & Spirituality",20,"religion-spirituality",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":127,"show_sort_weight":124,"slug":128},"World Cup","world-cup",{"id":130,"doc_module":4,"doc_module_name":45,"category_name":131,"show_sort_weight":130,"slug":132},10,"Lifestyle","lifestyle",{"id":134,"doc_module":4,"doc_module_name":45,"category_name":135,"show_sort_weight":104,"slug":136},19,"General","general"]