{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/29453"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/29453","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"A multi-level scalable startup for parallel applications","abstract":"High performance parallel machines with hundreds of thousands of processors and petascale performance are already in use, and even larger Exa flops scale computing systems which may have hundreds of millions of cores are planned. To run parallel applications on machines of such massive scale, one of the biggest challenges is the parallel startup process. This task involves two components: (1) parallel launching of appropriate processes on the given set of processors and (2) setting up communication channels to enable the processes to communicate with each other after process launching has completed. Most current startup mechanisms focus on either using special purpose daemons which waste system resources or using a startup manager which becomes a scalability bottleneck. In this thesis, we investigate the design and scalability of a SMP-aware, multi-level startup scheme with batching of remote shell sessions, which provides a complete solution to startup of a parallel application and facilitates its management during execution. It still supports existing Charm++ runtime capabilities including process health monitoring, facilitation of recovery from failures and scalable interaction with the application. We demonstrate the performance and scalability of this scheme by applying it to startup Charm++ applications. In particular, starting up a Charm++ program on 16,384 cores of Ranger (at TACC) with Ethernet as the underlying communication layer takes only 25 seconds and attains a speedup of over 400% compared to MPICH2-1.3 startup (using Hydra as process manager) and over 800% compared to Open MPI 1.3.1 startup on Ranger.","abstract_html":"High performance parallel machines with hundreds of thousands of processors and petascale performance are already in use, and even larger Exa flops scale computing systems which may have hundreds of millions of cores are planned. To run parallel applications on machines of such massive scale, one of the biggest challenges is the parallel startup process. This task involves two components: (1) parallel launching of appropriate processes on the given set of processors and (2) setting up communication channels to enable the processes to communicate with each other after process launching has completed. Most current startup mechanisms focus on either using special purpose daemons which waste system resources or using a startup manager which becomes a scalability bottleneck. In this thesis, we investigate the design and scalability of a SMP-aware, multi-level startup scheme with batching of remote shell sessions, which provides a complete solution to startup of a parallel application and facilitates its management during execution. It still supports existing Charm++ runtime capabilities including process health monitoring, facilitation of recovery from failures and scalable interaction with the application. We demonstrate the performance and scalability of this scheme by applying it to startup Charm++ applications. In particular, starting up a Charm++ program on 16,384 cores of Ranger (at TACC) with Ethernet as the underlying communication layer takes only 25 seconds and attains a speedup of over 400% compared to MPICH2-1.3 startup (using Hydra as process manager) and over 800% compared to Open MPI 1.3.1 startup on Ranger.","abstract_has_math":false,"creators":["Gupta, Abhishek"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Computer Science","degree_department":null,"school":null,"contributors":["Kale, Laxmikant V."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2012,"date_issued":"2012-02-01T00:46:59Z","date_published":"2012-02-01T00:46:59Z","updated_at":"2026-07-22T22:25:27Z","subjects":["parallel startup","multi-level startup","Hierarchical","tree startup","runtime","Charm++"],"languages":["en"],"rights":["Copyright 2011 Abhishek Gupta"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/29453","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Kale, Laxmikant V."]},{"key":"dc:creator","label":"Author","values":["Gupta, Abhishek"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2012-02-01T00:46:59Z","2014-02-01T11:00:27Z","2011-12"]},{"key":"dc:type","label":"Dc Type","values":["Dissertation / Thesis","text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Thesis"]},{"key":"thesis:degree_name","label":"Degree Name","values":["M.S."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["parallel startup","multi-level startup","Hierarchical","tree startup","runtime","Charm++"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2011 Abhishek Gupta"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/29453"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["High performance parallel machines with hundreds of thousands of processors and petascale performance are already in use, and even larger Exa flops scale computing systems which may have hundreds of millions of cores are planned. To run parallel applications on machines of such massive scale, one of the biggest challenges is the parallel startup process. This task involves two components: (1) parallel launching of appropriate processes on the given set of processors and (2) setting up communication channels to enable the processes to communicate with each other after process launching has completed. Most current startup mechanisms focus on either using special purpose daemons which waste system resources or using a startup manager which becomes a scalability bottleneck. In this thesis, we investigate the design and scalability of a SMP-aware, multi-level startup scheme with batching of remote shell sessions, which provides a complete solution to startup of a parallel application and facilitates its management during execution. It still supports existing Charm++ runtime capabilities including process health monitoring, facilitation of recovery from failures and scalable interaction with the application. We demonstrate the performance and scalability of this scheme by applying it to startup Charm++ applications. In particular, starting up a Charm++ program on 16,384 cores of Ranger (at TACC) with Ethernet as the underlying communication layer takes only 25 seconds and attains a speedup of over 400% compared to MPICH2-1.3 startup (using Hydra as process manager) and over 800% compared to Open MPI 1.3.1 startup on Ranger.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2011-12-02T20:42:28Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5)","Made available in DSpace on 2012-02-01T00:46:59Z (GMT). No. of bitstreams: 2 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5) license.txt: 4063 bytes, checksum: 140b56a563fb6eea0eb3029aa1327849 (MD5)","Item marked as restricted to the 'UIUC Users [automated]' Group (id=2) by William Ingram (wingram2@illinois.edu) on 2012-02-01T00:50:33Z Item is restricted until 2014-02-01T00:50:07Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2014-02-01T11:00:27Z Item was in collections: Dissertations and Theses - Computer Science (ID: 587) Graduate Theses and Dissertations at Illinois (ID: 204) No. of bitstreams: 3 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5) license.txt: 4063 bytes, checksum: 140b56a563fb6eea0eb3029aa1327849 (MD5) Gupta_Abhishek.pdf.txt: 67190 bytes, checksum: 172f003fdf1a1a64c1c64e8881b159e9 (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2014-02-01T11:00:27Z"]},{"key":"dc:title","label":"Title","values":["A multi-level scalable startup for parallel applications"]}]}],"canonical_facts":{"dc:contributor":["Kale, Laxmikant V."],"dc:creator":["Gupta, Abhishek"],"dc:date":["2012-02-01T00:46:59Z","2014-02-01T11:00:27Z","2011-12"],"dc:description":["High performance parallel machines with hundreds of thousands of processors and petascale performance are already in use, and even larger Exa flops scale computing systems which may have hundreds of millions of cores are planned. To run parallel applications on machines of such massive scale, one of the biggest challenges is the parallel startup process. This task involves two components: (1) parallel launching of appropriate processes on the given set of processors and (2) setting up communication channels to enable the processes to communicate with each other after process launching has completed. Most current startup mechanisms focus on either using special purpose daemons which waste system resources or using a startup manager which becomes a scalability bottleneck. In this thesis, we investigate the design and scalability of a SMP-aware, multi-level startup scheme with batching of remote shell sessions, which provides a complete solution to startup of a parallel application and facilitates its management during execution. It still supports existing Charm++ runtime capabilities including process health monitoring, facilitation of recovery from failures and scalable interaction with the application. We demonstrate the performance and scalability of this scheme by applying it to startup Charm++ applications. In particular, starting up a Charm++ program on 16,384 cores of Ranger (at TACC) with Ethernet as the underlying communication layer takes only 25 seconds and attains a speedup of over 400% compared to MPICH2-1.3 startup (using Hydra as process manager) and over 800% compared to Open MPI 1.3.1 startup on Ranger.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2011-12-02T20:42:28Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5)","Made available in DSpace on 2012-02-01T00:46:59Z (GMT). No. of bitstreams: 2 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5) license.txt: 4063 bytes, checksum: 140b56a563fb6eea0eb3029aa1327849 (MD5)","Item marked as restricted to the 'UIUC Users [automated]' Group (id=2) by William Ingram (wingram2@illinois.edu) on 2012-02-01T00:50:33Z Item is restricted until 2014-02-01T00:50:07Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2014-02-01T11:00:27Z Item was in collections: Dissertations and Theses - Computer Science (ID: 587) Graduate Theses and Dissertations at Illinois (ID: 204) No. of bitstreams: 3 Gupta_Abhishek.pdf: 350833 bytes, checksum: 49ab95fea7a0bf4cf69ba6862f79ce94 (MD5) license.txt: 4063 bytes, checksum: 140b56a563fb6eea0eb3029aa1327849 (MD5) Gupta_Abhishek.pdf.txt: 67190 bytes, checksum: 172f003fdf1a1a64c1c64e8881b159e9 (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2014-02-01T11:00:27Z"],"dc:identifier":["http://hdl.handle.net/2142/29453"],"dc:language":["en"],"dc:rights":["Copyright 2011 Abhishek Gupta"],"dc:subject":["parallel startup","multi-level startup","Hierarchical","tree startup","runtime","Charm++"],"dc:title":["A multi-level scalable startup for parallel applications"],"dc:type":["Dissertation / Thesis","text"],"thesis:degree_discipline":["Computer Science"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:27Z"}