How to emulate shared memory

Ranade, Abhiram

doi:10.1016/0022-0000(91)90005-p

Cited by 170 publications

(30 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The rationale offered is that these tasks can be assigned to the physical processors, with each processor apportioning its time among the tasks assigned to it. This technique of multithreading is a convenient way of masking latency [16,19]. Since each physical processor simulates several virtual processors, computation does not have to be suspended during the processing of a remote request by one of the virtual processors.…”

Section: Discouraged Loopholes and Rewarded Techniquesmentioning

confidence: 99%

LogP

et al. 1996

View full text Add to dashboard Cite

Section: Discouraged Loopholes and Rewarded Techniquesmentioning

confidence: 99%

LogP

et al. 1996

View full text Add to dashboard Cite

“…Fujimoto's Time Warp machine [17] applies optimistic PDES to hardware to implement a sequential program execution on parallel hardware using speculation. Ranade [36] proposed hardware to mimic a PRAM model using techniques derived from conservative PDES.…”

Section: Related Workmentioning

confidence: 99%

“…Logical time is updated using a token-passing method that inspired how the timestamp snooping network manages the guarantee time (GT). Ranade's network ensures that memory references arrive at the correct PRAM cycle [36]. Multicast snooping's network [9] ensures that address multicasts arrive in the logical order determined by the network roots.…”

Section: Related Workmentioning

confidence: 99%

Timestamp snooping

Martin

Wood

Sorin

et al. 2000

Proceedings of the Ninth International Conference on Architectural Support for Programming Languages and Operating Systems

View full text Add to dashboard Cite

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.

show abstract

“…Several such ingenious results are known for specific networks like the mesh, butterfly, hypercube etc. For instance, given any routing problem with N packets on art Nnode butterfly, there is a randomized online routing algorithm that with high probability, routes in O(log IV) time using 0(1)-sized queues [24].…”

Section: Introductionmentioning

confidence: 99%

A constant-factor approximation algorithm for packet routing, and balancing local vs. global criteria

Srinivasan

Teo

1997

Proceedings of the Twenty-Ninth Annual ACM Symposium on Theory of Computing - STOC '97

View full text Add to dashboard Cite

We present the first constant-factor approximation algorithm for a fundamental problem: the store-and-fonvarzipacket muting prvbiem on arbitrary networks, Furthermore, the queue sizes required at the edges are bounded by an absolute constant. Thus, this algorithm balances a global criterion (routing time) with a local criterion (maximum queue size), and showshow to get simultaneousgood bounds for both (though, for this particular problem, approximating the routing time well, even witbout considering the queue-sizes, was open). We then consider a class of such local vs. global problems in the context of covering integer programs, and show how to improve the local criterion by more than a constant (roughly logarithmic) factor, by losing a constant factor in the global criterion; this is in comparison with the current-best algorithms.

show abstract

How to emulate shared memory

Cited by 170 publications

References 4 publications

LogP

LogP

Timestamp snooping

A constant-factor approximation algorithm for packet routing, and balancing local vs. global criteria

Contact Info

Product

Resources

About