Looking into the Peak Memory Consumption of Epoch-Based Reclamation in Scalable in-Memory Database Systems

Mitake, Hitoshi; Yamada, Hiroshi; Nakajima, Tatsuo

doi:10.1007/978-3-030-27618-8_1

Cited by 2 publications

(2 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our work sheds new light on why techniques like Conditional Access that immediately free individual objects perform well in practice. Mitake et al [29] studied the impact of peak memory usage on in-memory database transaction latencies that used epoch based reclamation. They suggested freeing batches using a separate background thread to increase the frequency with which the main worker threads could participate in advancing the epoch.…”

Section: Related Workmentioning

confidence: 99%

Are Your Epochs Too Epic? Batch Free Can Be Harmful

Kim,

Brown,

Singh

2024

Proceedings of the 29th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

Epoch based memory reclamation (EBR) is one of the most popular techniques for reclaiming memory in lock-free and optimistic locking data structures, due to its ease of use and good performance in practice. However, EBR is known to be sensitive to thread delays, which can result in performance degradation. Moreover, the exact mechanism for this performance degradation is not well understood.This paper illustrates this performance degradation in a popular data structure benchmark, and does a deep dive to uncover its root cause-a subtle interaction between EBR and state of the art memory allocators. In essence, modern allocators attempt to reduce the overhead of freeing by maintaining bounded thread caches of objects for local reuse, actually freeing them (a very high latency operation) only when thread caches become too large. EBR immediately bypasses these mechanisms whenever a particularly large batch of objects is freed, substantially increasing overheads and latencies. Beyond EBR, many memory reclamation algorithms, and data structures, that reclaim objects in large batches suffer similar deleterious interactions with popular allocators.We propose a simple algorithmic fix for such algorithms to amortize the freeing of large object batches over time, and apply this technique to ten existing memory reclamation algorithms, observing performance improvements for nine out of ten, and over 50% improvement for six out of ten in experiments on a high performance lock-free ABtree. We also present an extremely simple token passing variant of EBR and show that, with our fix, it performs 1.5-2.6× faster than the fastest known memory reclamation algorithm, and 1.2-1.5× faster than not reclaiming at all, on a 192 thread four socket Intel system.

show abstract

Section: Related Workmentioning

confidence: 99%

Are Your Epochs Too Epic? Batch Free Can Be Harmful

Kim,

Brown,

Singh

2024

Proceedings of the 29th ACM SIGPLAN Annual Symposium on Principles and Practice of Parallel Programming

View full text Add to dashboard Cite

show abstract

“…The second shortcoming that must be overcome to develop a scalable and stable platform supporting practical services is managing the substantial number of visual perspectives: the platform must account for various system techniques such as those in Refs. [45,46] that make the platform scalable and stable. In a large scaled system, it is hard to support a large number of simultaneous visual perspectives at the same time.…”

Section: Three Shortcomings and Their Possible Solutions Of The Current Collectiveeyes Prototype Platformmentioning

confidence: 99%

Collectively Sharing People’s Visual and Auditory Capabilities: Exploring Opportunities and Pitfalls

Kimura

Nakajima

2020

SN COMPUT. SCI.

Self Cite

View full text Add to dashboard Cite

The sharing economy, virtualizing the peer-to-peer-based sharing of physical or logical resources such as cars, houses or people's spare time with digital technologies, has recently attracted a great deal of attention. Investigating the feasibility of sharing other types of physical resources, particularly human bodies, offers promising opportunities with which to expand the current scope of the sharing economy. This study explores the opportunities and pitfalls of collectively sharing parts of the body, particularly human eyes and ears, to build and explore novel services based on wearable cameras and microphones. The paper proposes a novel concept named CollectiveEyes to realize a digital platform for collectively sharing human seeing and hearing, then presents a prototype platform of CollectiveEyes to demonstrate key features of CollectiveEyes. The essential contribution of the paper is that it reveals potential opportunities and pitfalls of collectively sharing physical capabilities of the human eyes and ears for building a novel digital platform and diverse services on the platform. We mainly found the following two promising research directions extracted from the current study: (1) CollectiveEyes enhances human thinking abilities through multiple visual perspectives; (2) nonhuman's visual perspectives offer potential opportunities to access more diverse perspectives for augmenting human thinking abilities.

show abstract

Looking into the Peak Memory Consumption of Epoch-Based Reclamation in Scalable in-Memory Database Systems

Cited by 2 publications

References 16 publications

Are Your Epochs Too Epic? Batch Free Can Be Harmful

Are Your Epochs Too Epic? Batch Free Can Be Harmful

Collectively Sharing People’s Visual and Auditory Capabilities: Exploring Opportunities and Pitfalls

Contact Info

Product

Resources

About