File Download
Links for fulltext
(May Require Subscription)
- Publisher Website: 10.1109/ICPADS.2011.144
- Scopus: eid_2-s2.0-84863080401
- WOS: WOS:000299395900038
- Find via
Supplementary
- Citations:
- Appears in Collections:
Conference Paper: TrC-MC: decentralized software transactional memory for multi-multicore computers
Title | TrC-MC: decentralized software transactional memory for multi-multicore computers |
---|---|
Authors | |
Keywords | Algorithm Cache contention Multicore Software transactional memory |
Issue Date | 2011 |
Publisher | IEEE. |
Citation | The 17th IEEE International Conference on Parallel and Distributed Systems (ICPADS 2011), Tainan, Taiwan, 7-9 December 2011. In Proceedings of the 17th IEEE ICPADS, 2011, p. 292-299 How to Cite? |
Abstract | To achieve single-lock atomicity in software transactional memory systems, the commit procedure often goes through a common clock variable. When there are frequent transactional commits, clock sharing becomes inefficient. Tremendous cache contention takes place between the processors and the computing throughput no longer scales with processor count. Therefore, traditional transactional memories are unable to accelerate applications with frequent commits regardless of thread count. While systems with decentralized data structures have better performance on these applications, we argue they are incomplete as they create much more aborts than traditional transactional systems. In this paper we apply two design changes, namely zone partitioning and timestamp extension, to optimize an existing decentralized algorithm. We prove the correctness and evaluate some benchmark programs with frequent transactional commits. We find it as much as several times faster than the state-of-the-art software transactional memory system. We have also reduced the abort rate of the system to an acceptable level. |
Persistent Identifier | http://hdl.handle.net/10722/144627 |
ISSN | 2023 SCImago Journal Rankings: 0.397 |
ISI Accession Number ID |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Chan, K | en_US |
dc.contributor.author | Wang, CL | en_US |
dc.date.accessioned | 2012-02-03T06:16:58Z | - |
dc.date.available | 2012-02-03T06:16:58Z | - |
dc.date.issued | 2011 | en_US |
dc.identifier.citation | The 17th IEEE International Conference on Parallel and Distributed Systems (ICPADS 2011), Tainan, Taiwan, 7-9 December 2011. In Proceedings of the 17th IEEE ICPADS, 2011, p. 292-299 | en_US |
dc.identifier.issn | 1521-9097 | - |
dc.identifier.uri | http://hdl.handle.net/10722/144627 | - |
dc.description.abstract | To achieve single-lock atomicity in software transactional memory systems, the commit procedure often goes through a common clock variable. When there are frequent transactional commits, clock sharing becomes inefficient. Tremendous cache contention takes place between the processors and the computing throughput no longer scales with processor count. Therefore, traditional transactional memories are unable to accelerate applications with frequent commits regardless of thread count. While systems with decentralized data structures have better performance on these applications, we argue they are incomplete as they create much more aborts than traditional transactional systems. In this paper we apply two design changes, namely zone partitioning and timestamp extension, to optimize an existing decentralized algorithm. We prove the correctness and evaluate some benchmark programs with frequent transactional commits. We find it as much as several times faster than the state-of-the-art software transactional memory system. We have also reduced the abort rate of the system to an acceptable level. | - |
dc.language | eng | en_US |
dc.publisher | IEEE. | en_US |
dc.relation.ispartof | Proceedings of the International Conference on Parallel and Distributed Systems | en_US |
dc.subject | Algorithm | - |
dc.subject | Cache contention | - |
dc.subject | Multicore | - |
dc.subject | Software transactional memory | - |
dc.title | TrC-MC: decentralized software transactional memory for multi-multicore computers | en_US |
dc.type | Conference_Paper | en_US |
dc.identifier.email | Chan, K: kchan@cs.hku.hk | en_US |
dc.identifier.email | Wang, CL: clwang@cs.hku.hk | - |
dc.identifier.authority | Wang, CL=rp00183 | en_US |
dc.description.nature | link_to_subscribed_fulltext | - |
dc.identifier.doi | 10.1109/ICPADS.2011.144 | - |
dc.identifier.scopus | eid_2-s2.0-84863080401 | - |
dc.identifier.hkuros | 198270 | en_US |
dc.identifier.spage | 292 | - |
dc.identifier.epage | 299 | - |
dc.identifier.isi | WOS:000299395900038 | - |
dc.description.other | The 17th IEEE International Conference on Parallel and Distributed Systems (ICPADS 2011), Tainan, Taiwan, 7-9 December 2011. In Proceedings of the 17th IEEE ICPADS, 2011, p. 292-299 | - |
dc.identifier.issnl | 1521-9097 | - |