The Community for Technology Leaders
2007 IEEE 13th International Symposium on High Performance Computer Architecture (2007)
Scottsdale, AZ, USA
Feb. 10, 2007 to Feb. 14, 2007
ISBN: 1-4244-0804-0
pp: 97-108
Hassan Chafi , Computer Systems Laboratory, Stanford University. hchafi@stanford.edu
Jared Casper , Computer Systems Laboratory, Stanford University. jaredc@stanford.edu
Brian D. Carlstrom , Computer Systems Laboratory, Stanford University. bdc@stanford.edu
Austen McDonald , Computer Systems Laboratory, Stanford University. austenmc@stanford.edu
Chi Cao Minh , Computer Systems Laboratory, Stanford University. caominh@stanford.edu
Woongki Baek , Computer Systems Laboratory, Stanford University. wbaek@stanford.edu
Christos Kozyrakis , Computer Systems Laboratory, Stanford University. kozyrakis@stanford.edu
Kunle Olukotun , Computer Systems Laboratory, Stanford University. kunle@stanford.edu
ABSTRACT
Transactional Memory (TM) provides mechanisms that promise to simplify parallel programming by eliminating the need for locks and their associated problems (dead-lock, livelock, priority inversion, convoying). For TM to be adopted in the long term, not only does it need to deliver on these promises, but it needs to scale to a high number of processors. To date, proposals for scalable TM have relegated livelock issues to user-level contention managers. This paper presents the first scalable TM implementation for directory-based distributed shared memory systems that is livelock free without the need for user-level intervention. The design is a scalable implementation of optimistic concurrency control that supports parallel commits with a two-phase commit protocol, uses write-back caches, and filters coherence messages. The scalable design is based on Transactional Coherence and Consistency (TCC), which supports continuous transactions and fault isolation. A performance evaluation of the design using both scientific and enterprise benchmarks demonstrates that the directory-based TCC design scales efficiently for NUMA systems up to 64 processors.
INDEX TERMS
null
CITATION

A. McDonald et al., "A Scalable, Non-blocking Approach to Transactional Memory," 2007 IEEE 13th International Symposium on High Performance Computer Architecture(HPCA), Scottsdale, AZ, USA, 2007, pp. 97-108.
doi:10.1109/HPCA.2007.346189
93 ms
(Ver 3.3 (11022016))