{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/31986"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/31986","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Adaptive strategy for one-sided communication in MPICH2","abstract":"The one-sided communication model (or remote memory access) supported by MPI-2 is more convenient to use than the regular two-sided send/receive communication model, because it allows the sender to specify all data transfer parameters and avoids the receiver to be explicitly involved in data receiving. One-sided communication model also has potential to provide higher performance. The MPI-2 standard provides flexibility about when RMA operations can be issued and completed, which makes the MPI implementation possible to be optimized internally. The current MPICH2 implementation uses a lazy approach to issue one-sided operations by queuing the operations and issuing them during the later synchronization phase. This has certain benefits with respect to short operations in terms of reduced network operations. For large data transfers, issuing operations in an eager fashion could be more beneficial as well as provide more scope for overlapping communication and computation. In this thesis we describe the design and implementation of an adaptive approach for all three synchronization mechanisms defined in MPI: fence, post-start-complete-wait, and lock-unlock. We evaluate our implementation with both micro benchmarks and the Graph500 benchmark to demonstrate the performance impact of our approach. The performance results show that our hybrid approach performs as good as the lazy approach for small data transfers and achieves performance similar to the eager approach for large data transfers, and its overlapping percentage is good.","abstract_html":"The one-sided communication model (or remote memory access) supported by MPI-2 is more convenient to use than the regular two-sided send/receive communication model, because it allows the sender to specify all data transfer parameters and avoids the receiver to be explicitly involved in data receiving. One-sided communication model also has potential to provide higher performance. The MPI-2 standard provides flexibility about when RMA operations can be issued and completed, which makes the MPI implementation possible to be optimized internally. The current MPICH2 implementation uses a lazy approach to issue one-sided operations by queuing the operations and issuing them during the later synchronization phase. This has certain benefits with respect to short operations in terms of reduced network operations. For large data transfers, issuing operations in an eager fashion could be more beneficial as well as provide more scope for overlapping communication and computation. In this thesis we describe the design and implementation of an adaptive approach for all three synchronization mechanisms defined in MPI: fence, post-start-complete-wait, and lock-unlock. We evaluate our implementation with both micro benchmarks and the Graph500 benchmark to demonstrate the performance impact of our approach. The performance results show that our hybrid approach performs as good as the lazy approach for small data transfers and achieves performance similar to the eager approach for large data transfers, and its overlapping percentage is good.","abstract_has_math":false,"creators":["Zhao, Xin"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"M.S.","degree_level":"Thesis","degree_discipline":"Computer Science Computer Science","degree_department":null,"school":null,"contributors":["Gropp, William D."],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2012,"date_issued":"2012-06-27T21:23:13Z","date_published":"2012-06-27T21:23:13Z","updated_at":"2026-07-22T22:25:30Z","subjects":["One-sided Communication","MPI implementation","Synchronization","Adaptive Strategy","Message Passing Interface (MPI)"],"languages":["en"],"rights":["Copyright 2012 Xin Zhao"],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/2142/31986","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Gropp, William D."]},{"key":"dc:creator","label":"Author","values":["Zhao, Xin"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2012-06-27T21:23:13Z","2014-06-28T10:00:24Z","2012-05"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Computer Science 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":["One-sided Communication","MPI implementation","Synchronization","Adaptive Strategy","Message Passing Interface (MPI)"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["en"]},{"key":"dc:rights","label":"Dc Rights","values":["Copyright 2012 Xin Zhao"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/31986"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["The one-sided communication model (or remote memory access) supported by MPI-2 is more convenient to use than the regular two-sided send/receive communication model, because it allows the sender to specify all data transfer parameters and avoids the receiver to be explicitly involved in data receiving. One-sided communication model also has potential to provide higher performance. The MPI-2 standard provides flexibility about when RMA operations can be issued and completed, which makes the MPI implementation possible to be optimized internally. The current MPICH2 implementation uses a lazy approach to issue one-sided operations by queuing the operations and issuing them during the later synchronization phase. This has certain benefits with respect to short operations in terms of reduced network operations. For large data transfers, issuing operations in an eager fashion could be more beneficial as well as provide more scope for overlapping communication and computation. In this thesis we describe the design and implementation of an adaptive approach for all three synchronization mechanisms defined in MPI: fence, post-start-complete-wait, and lock-unlock. We evaluate our implementation with both micro benchmarks and the Graph500 benchmark to demonstrate the performance impact of our approach. The performance results show that our hybrid approach performs as good as the lazy approach for small data transfers and achieves performance similar to the eager approach for large data transfers, and its overlapping percentage is good.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2012-04-25T17:03:55Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5)","Made available in DSpace on 2012-06-27T21:23:13Z (GMT). No. of bitstreams: 2 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5) license.txt: 4058 bytes, checksum: b29106ea1f21866a07c044441a656b8d (MD5)","Item marked as restricted to the 'UIUC Users [automated]' Group (id=2) by William Ingram (wingram2@illinois.edu) on 2012-06-27T21:24:45Z Item is restricted until 2014-06-27T21:24:27Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2014-06-28T10:00:24Z Item was in collections: Graduate Theses and Dissertations at Illinois (ID: 204) Dissertations and Theses - Computer Science (ID: 587) No. of bitstreams: 2 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5) license.txt: 4058 bytes, checksum: b29106ea1f21866a07c044441a656b8d (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2014-06-28T10:00:24Z"]},{"key":"dc:title","label":"Title","values":["Adaptive strategy for one-sided communication in MPICH2"]}]}],"canonical_facts":{"dc:contributor":["Gropp, William D."],"dc:creator":["Zhao, Xin"],"dc:date":["2012-06-27T21:23:13Z","2014-06-28T10:00:24Z","2012-05"],"dc:description":["The one-sided communication model (or remote memory access) supported by MPI-2 is more convenient to use than the regular two-sided send/receive communication model, because it allows the sender to specify all data transfer parameters and avoids the receiver to be explicitly involved in data receiving. One-sided communication model also has potential to provide higher performance. The MPI-2 standard provides flexibility about when RMA operations can be issued and completed, which makes the MPI implementation possible to be optimized internally. The current MPICH2 implementation uses a lazy approach to issue one-sided operations by queuing the operations and issuing them during the later synchronization phase. This has certain benefits with respect to short operations in terms of reduced network operations. For large data transfers, issuing operations in an eager fashion could be more beneficial as well as provide more scope for overlapping communication and computation. In this thesis we describe the design and implementation of an adaptive approach for all three synchronization mechanisms defined in MPI: fence, post-start-complete-wait, and lock-unlock. We evaluate our implementation with both micro benchmarks and the Graph500 benchmark to demonstrate the performance impact of our approach. The performance results show that our hybrid approach performs as good as the lazy approach for small data transfers and achieves performance similar to the eager approach for large data transfers, and its overlapping percentage is good.","Item withdrawn by Mark Zulauf (zulauf@illinois.edu) on 2012-04-25T17:03:55Z Item was in collections: University of Illinois Theses & Dissertations (ID: 1) No. of bitstreams: 1 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5)","Made available in DSpace on 2012-06-27T21:23:13Z (GMT). No. of bitstreams: 2 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5) license.txt: 4058 bytes, checksum: b29106ea1f21866a07c044441a656b8d (MD5)","Item marked as restricted to the 'UIUC Users [automated]' Group (id=2) by William Ingram (wingram2@illinois.edu) on 2012-06-27T21:24:45Z Item is restricted until 2014-06-27T21:24:27Z","Item reinstated by Sarah Shreeves (sshreeve@illinois.edu) on 2014-06-28T10:00:24Z Item was in collections: Graduate Theses and Dissertations at Illinois (ID: 204) Dissertations and Theses - Computer Science (ID: 587) No. of bitstreams: 2 Zhao_Xin.pdf: 749248 bytes, checksum: 600e46d662705cf8067ae831450c8e77 (MD5) license.txt: 4058 bytes, checksum: b29106ea1f21866a07c044441a656b8d (MD5)","Item released from any restrictions by Sarah Shreeves (sshreeve@illinois.edu) on 2014-06-28T10:00:24Z"],"dc:identifier":["http://hdl.handle.net/2142/31986"],"dc:language":["en"],"dc:rights":["Copyright 2012 Xin Zhao"],"dc:subject":["One-sided Communication","MPI implementation","Synchronization","Adaptive Strategy","Message Passing Interface (MPI)"],"dc:title":["Adaptive strategy for one-sided communication in MPICH2"],"dc:type":["text"],"thesis:degree_discipline":["Computer Science Computer Science"],"thesis:degree_level":["Thesis"],"thesis:degree_name":["M.S."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:25:30Z"}