Skip to main content

Timestamp snooping: An approach for extending SMPs

Publication ,  Journal Article
Martin, MMK; Sorin, DJ; Ailamaki, A; Alameldeen, AR; Dickson, RM; Mauer, CJ; Moore, KE; Plakal, M; Hill, MD; Wood, DA
Published in: SIGPLAN Notices (ACM Special Interest Group on Programming Languages)
January 1, 2000

Symmetric multiprocessor (SMP) servers provide superior performance for the commercial workloads that dominate the Internet. Our simulation results show that over one-third of cache misses by these applications result in cache-to-cache transfers, where the data is found in another processor's cache rather than in memory. SMPs are optimized for this case by using snooping protocols that broadcast address transactions to all processors. Conversely, directory-based shared-memory systems must indirectly locate the owner and sharers through a directory, resulting in larger average miss latencies. This paper proposes timestamp snooping, a technique that allows SMPs to i) utilize high-speed switched interconnection networks and ii) exploit physical locality by delivering address transactions to processors and memories without regard to order. Traditional snooping requires physical ordering of transactions. Timestamp snooping works by processing address transactions in a logical order. Logical time is maintained by adding a few bits per address transaction and having network switches perform a handshake to ensure on-time delivery. Processors and memories then reorder transactions based on their timestamps to establish a total order. We evaluate timestamp snooping with commercial workloads on a 16-processor SPARC system using the Simics full-system simulator. We simulate both an indirect (butterfly) and a direct (torus) network design. For OLTP, DSS, web serving, web searching, and one scientific application, timestamp snooping with the butterfly network runs 6-28% faster than directories, at a cost of 13-43% more link traffic. Similarly, with the torus network, timestamp snooping runs 6-29% faster for 17-37% more link traffic. Thus, timestamp snooping is worth considering when buying more interconnect bandwidth is easier than reducing interconnect latency. © 2000 ACM.

Duke Scholars

Altmetric Attention Stats
Dimensions Citation Stats

Published In

SIGPLAN Notices (ACM Special Interest Group on Programming Languages)

DOI

ISSN

0362-1340

Publication Date

January 1, 2000

Volume

35

Issue

11

Start / End Page

25 / 36

Related Subject Headings

  • Software Engineering
 

Citation

APA
Chicago
ICMJE
MLA
NLM
Martin, M. M. K., Sorin, D. J., Ailamaki, A., Alameldeen, A. R., Dickson, R. M., Mauer, C. J., … Wood, D. A. (2000). Timestamp snooping: An approach for extending SMPs. SIGPLAN Notices (ACM Special Interest Group on Programming Languages), 35(11), 25–36. https://doi.org/10.1145/356989.356992
Martin, M. M. K., D. J. Sorin, A. Ailamaki, A. R. Alameldeen, R. M. Dickson, C. J. Mauer, K. E. Moore, M. Plakal, M. D. Hill, and D. A. Wood. “Timestamp snooping: An approach for extending SMPs.” SIGPLAN Notices (ACM Special Interest Group on Programming Languages) 35, no. 11 (January 1, 2000): 25–36. https://doi.org/10.1145/356989.356992.
Martin MMK, Sorin DJ, Ailamaki A, Alameldeen AR, Dickson RM, Mauer CJ, et al. Timestamp snooping: An approach for extending SMPs. SIGPLAN Notices (ACM Special Interest Group on Programming Languages). 2000 Jan 1;35(11):25–36.
Martin, M. M. K., et al. “Timestamp snooping: An approach for extending SMPs.” SIGPLAN Notices (ACM Special Interest Group on Programming Languages), vol. 35, no. 11, Jan. 2000, pp. 25–36. Scopus, doi:10.1145/356989.356992.
Martin MMK, Sorin DJ, Ailamaki A, Alameldeen AR, Dickson RM, Mauer CJ, Moore KE, Plakal M, Hill MD, Wood DA. Timestamp snooping: An approach for extending SMPs. SIGPLAN Notices (ACM Special Interest Group on Programming Languages). 2000 Jan 1;35(11):25–36.

Published In

SIGPLAN Notices (ACM Special Interest Group on Programming Languages)

DOI

ISSN

0362-1340

Publication Date

January 1, 2000

Volume

35

Issue

11

Start / End Page

25 / 36

Related Subject Headings

  • Software Engineering