{"id":{"repo_id":"mtu","oai_identifier":"oai:digitalcommons.mtu.edu:etdr-2077"},"canonical_url":"https://search.ndltd.org/etd/mtu/oai:digitalcommons.mtu.edu:etdr-2077","repository":{"repo_id":"mtu","name":"Michigan Technological University","base_url":"https://digitalcommons.mtu.edu/do/oai/"},"display":{"title":"Contextual Bandit Modeling for Dynamic Runtime Control in Computer Systems","abstract":"Modern operating systems and microarchitectures provide a myriad of mechanisms for monitoring and affecting system operation and resource utilization at runtime. Dynamic runtime control of these mechanisms can tailor system operation to the characteristics and behavior of the current workload, resulting in improved performance. However, developing effective models for system control can be challenging. Existing methods often require extensive manual effort, computation time, and domain knowledge to identify relevant low-level performance metrics, relate low-level performance metrics and high-level control decisions to workload performance, and to evaluate the resulting control models. This dissertation develops a general framework, based on the contextual bandit, for describing and learning effective models for runtime system control. Random profiling is used to characterize the relationship between workload behavior, system configuration, and performance. The framework is evaluated in the context of two applications of progressive complexity; first, the selection of paging modes (Shadow Paging, Hardware-Assisted Page) in the Xen virtual machine memory manager; second, the utilization of hardware memory prefetching for multi-core, multi-tenant workloads with cross-core contention for shared memory resources, such as the last-level cache and memory bandwidth. The resulting models for both applications are competitive in comparison to existing runtime control approaches. For paging mode selection, the resulting model provides equivalent performance to the state of the art while substantially reducing the computation requirements of profiling. For hardware memory prefetcher utilization, the resulting models are the first to provide dynamic control for hardware prefetchers using workload statistics. Finally, a correlation-based feature selection method is evaluated for identifying relevant low-level performance metrics related to hardware memory prefetching.","abstract_html":"Modern operating systems and microarchitectures provide a myriad of mechanisms for monitoring and affecting system operation and resource utilization at runtime. Dynamic runtime control of these mechanisms can tailor system operation to the characteristics and behavior of the current workload, resulting in improved performance. However, developing effective models for system control can be challenging. Existing methods often require extensive manual effort, computation time, and domain knowledge to identify relevant low-level performance metrics, relate low-level performance metrics and high-level control decisions to workload performance, and to evaluate the resulting control models. This dissertation develops a general framework, based on the contextual bandit, for describing and learning effective models for runtime system control. Random profiling is used to characterize the relationship between workload behavior, system configuration, and performance. The framework is evaluated in the context of two applications of progressive complexity; first, the selection of paging modes (Shadow Paging, Hardware-Assisted Page) in the Xen virtual machine memory manager; second, the utilization of hardware memory prefetching for multi-core, multi-tenant workloads with cross-core contention for shared memory resources, such as the last-level cache and memory bandwidth. The resulting models for both applications are competitive in comparison to existing runtime control approaches. For paging mode selection, the resulting model provides equivalent performance to the state of the art while substantially reducing the computation requirements of profiling. For hardware memory prefetcher utilization, the resulting models are the first to provide dynamic control for hardware prefetchers using workload statistics. Finally, a correlation-based feature selection method is evaluated for identifying relevant low-level performance metrics related to hardware memory prefetching.","abstract_has_math":false,"creators":["Hiebel, Jason"],"institution":null,"degree_name":"Doctor of Philosophy in Computer Science (PhD)","degree_level":"Open Access Dissertation","degree_discipline":"Department of Computer Science","degree_department":null,"school":null,"contributors":["Laura E. Brown","Zhenlin Wang"],"advisors":[],"committee_chairs":[],"committee_members":["Nilufer Onder","Allan A. Struthers"],"year":null,"date_issued":"","date_published":null,"updated_at":"2026-11-19T08:17:59Z","subjects":["Contextual Bandits","Virtual Memory","Hardware Memory Prefetching","Hardware Performance Monitoring","Artificial Intelligence and Robotics","Systems Architecture"],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["https://digitalcommons.mtu.edu/etdr/942"],"render_values":[{"text":"https://digitalcommons.mtu.edu/etdr/942","href":"https://digitalcommons.mtu.edu/etdr/942","code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/10.37099/mtu.dc.etdr/942","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/10.37099/mtu.dc.etdr/942","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://digitalcommons.mtu.edu/do/oai/?verb=GetRecord&metadataPrefix=oai_etdms&identifier=oai%3Adigitalcommons.mtu.edu%3Aetdr-2077","prefix":"oai_etdms"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Laura E. Brown","Zhenlin Wang"]},{"key":"dc:contributor.committeemember","label":"Committee Member","values":["Nilufer Onder","Allan A. Struthers"]},{"key":"dc:creator","label":"Author","values":["Hiebel, Jason"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.created","label":"Dc Date Created","values":["2019-01-01T08:00:00Z"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Department of Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Open Access Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Doctor of Philosophy in Computer Science (PhD)"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Contextual Bandits","Virtual Memory","Hardware Memory Prefetching","Hardware Performance Monitoring","Artificial Intelligence and Robotics","Systems Architecture"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["10.37099/mtu.dc.etdr/942","https://digitalcommons.mtu.edu/etdr/942"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Modern operating systems and microarchitectures provide a myriad of mechanisms for monitoring and affecting system operation and resource utilization at runtime. Dynamic runtime control of these mechanisms can tailor system operation to the characteristics and behavior of the current workload, resulting in improved performance. However, developing effective models for system control can be challenging. Existing methods often require extensive manual effort, computation time, and domain knowledge to identify relevant low-level performance metrics, relate low-level performance metrics and high-level control decisions to workload performance, and to evaluate the resulting control models. This dissertation develops a general framework, based on the contextual bandit, for describing and learning effective models for runtime system control. Random profiling is used to characterize the relationship between workload behavior, system configuration, and performance. The framework is evaluated in the context of two applications of progressive complexity; first, the selection of paging modes (Shadow Paging, Hardware-Assisted Page) in the Xen virtual machine memory manager; second, the utilization of hardware memory prefetching for multi-core, multi-tenant workloads with cross-core contention for shared memory resources, such as the last-level cache and memory bandwidth. The resulting models for both applications are competitive in comparison to existing runtime control approaches. For paging mode selection, the resulting model provides equivalent performance to the state of the art while substantially reducing the computation requirements of profiling. For hardware memory prefetcher utilization, the resulting models are the first to provide dynamic control for hardware prefetchers using workload statistics. Finally, a correlation-based feature selection method is evaluated for identifying relevant low-level performance metrics related to hardware memory prefetching."]},{"key":"dc:description.note","label":"Dc Description Note","values":["<p>Fixed the suite of margin issues introduced in Appendix A. These documents were added last and I failed to perform my due diligence to ensure they followed the formatting specifications.</p>"]},{"key":"dc:title","label":"Title","values":["Contextual Bandit Modeling for Dynamic Runtime Control in Computer Systems"]}]}],"canonical_facts":{"dc:contributor":["Laura E. Brown","Zhenlin Wang"],"dc:contributor.committeemember":["Nilufer Onder","Allan A. Struthers"],"dc:creator":["Hiebel, Jason"],"dc:date.created":["2019-01-01T08:00:00Z"],"dc:description.abstract":["Modern operating systems and microarchitectures provide a myriad of mechanisms for monitoring and affecting system operation and resource utilization at runtime. Dynamic runtime control of these mechanisms can tailor system operation to the characteristics and behavior of the current workload, resulting in improved performance. However, developing effective models for system control can be challenging. Existing methods often require extensive manual effort, computation time, and domain knowledge to identify relevant low-level performance metrics, relate low-level performance metrics and high-level control decisions to workload performance, and to evaluate the resulting control models. This dissertation develops a general framework, based on the contextual bandit, for describing and learning effective models for runtime system control. Random profiling is used to characterize the relationship between workload behavior, system configuration, and performance. The framework is evaluated in the context of two applications of progressive complexity; first, the selection of paging modes (Shadow Paging, Hardware-Assisted Page) in the Xen virtual machine memory manager; second, the utilization of hardware memory prefetching for multi-core, multi-tenant workloads with cross-core contention for shared memory resources, such as the last-level cache and memory bandwidth. The resulting models for both applications are competitive in comparison to existing runtime control approaches. For paging mode selection, the resulting model provides equivalent performance to the state of the art while substantially reducing the computation requirements of profiling. For hardware memory prefetcher utilization, the resulting models are the first to provide dynamic control for hardware prefetchers using workload statistics. Finally, a correlation-based feature selection method is evaluated for identifying relevant low-level performance metrics related to hardware memory prefetching."],"dc:description.note":["<p>Fixed the suite of margin issues introduced in Appendix A. These documents were added last and I failed to perform my due diligence to ensure they followed the formatting specifications.</p>"],"dc:identifier":["10.37099/mtu.dc.etdr/942","https://digitalcommons.mtu.edu/etdr/942"],"dc:subject":["Contextual Bandits","Virtual Memory","Hardware Memory Prefetching","Hardware Performance Monitoring","Artificial Intelligence and Robotics","Systems Architecture"],"dc:title":["Contextual Bandit Modeling for Dynamic Runtime Control in Computer Systems"],"thesis:degree_discipline":["Department of Computer Science"],"thesis:degree_level":["Open Access Dissertation"],"thesis:degree_name":["Doctor of Philosophy in Computer Science (PhD)"]},"updated_at":"2026-11-19T08:17:59Z"}