2017 IEEE International Parallel and Distributed Processing Symposium (IPDPS) 2017
DOI: 10.1109/ipdps.2017.41
|View full text |Cite
|
Sign up to set email alerts
|

FFQ: A Fast Single-Producer/Multiple-Consumer Concurrent FIFO Queue

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
4
1

Citation Types

0
7
0

Year Published

2019
2019
2023
2023

Publication Types

Select...
3
2

Relationship

0
5

Authors

Journals

citations
Cited by 11 publications
(8 citation statements)
references
References 18 publications
0
7
0
Order By: Relevance
“…Unfortunately, in parallelizing high-performance applications by leveraging pipeline parallelism, we found that the performance of state-of-the-art queue implementations deteriorates dramatically, and the performance of the parallelized applications is limited by the queues. This performance deterioration has been noticed in literature [5], [7], [11], but, unfortunately, has not been well studied, let alone solved. When looking into existing lock-free queues, we found that the primary reason is as follows.…”
Section: Introductionmentioning
confidence: 93%
See 3 more Smart Citations
“…Unfortunately, in parallelizing high-performance applications by leveraging pipeline parallelism, we found that the performance of state-of-the-art queue implementations deteriorates dramatically, and the performance of the parallelized applications is limited by the queues. This performance deterioration has been noticed in literature [5], [7], [11], but, unfortunately, has not been well studied, let alone solved. When looking into existing lock-free queues, we found that the primary reason is as follows.…”
Section: Introductionmentioning
confidence: 93%
“…To provide a high-speed core-to-core communication mechanism, researchers typically utilize single-producersingle-consumer FIFO queues (e.g. [4]- [7], [10]- [13], [32], [33]). The major benefit of these implementations is that it is not necessary for the queues to handle the aforementioned ABA and dynamic memory allocation problems.…”
Section: Related Workmentioning
confidence: 99%
See 2 more Smart Citations
“…Prior research works improve the performance of wait-free algorithms mainly at the algorithmic level [9], [13]. For example, in 2012, Kogan and Petrank [6] and subsequent works [8], [9], [14] proposed a fast-path-slow-path methodology for creating practical wait-free queues. The fast-path, usually a lock-free algorithm, provides good performance; the slow-path ensures wait-freedom.…”
Section: Introductionmentioning
confidence: 99%