Atomic snapshots of shared memory

Afek, Yehuda; Dolev, Danny; Attiya, Hagit; Gafni, Eli; Merritt, Michael; Shavit, Nir

doi:10.1145/93385.93394

Cited by 118 publications

(193 citation statements)

References 19 publications

Supporting

Mentioning

193

Contrasting

Order By: Relevance

“…• If it is a major, p i is required to invoke XCONS [1].xcons propose(v) in order the majors that participate agree (among themselves) on a single value (line 02), and that value is made public by writing it into PROP [1] (line 03).…”

Section: Xcons [0]mentioning

confidence: 99%

“…This occurs when the majors crash after having written PROP [1] and before giving a value to the atomic register WINNER. In that case, a correct minor can be blocked when it executes wait(WINNER = ⊥) (line 10).…”

Section: Xcons [0]mentioning

confidence: 99%

“…If the decided value is the value present in PROP [1], then WINNER has been set to 1 by a major process before any process decides. In that case, a major process has written the result of XCONS [1].xcons propose() in PROP [1] before setting WINNER, and this value has been proposed by one of the major processes.…”

Section: Proof Of the Constructionmentioning

confidence: 99%

See 2 more Smart Citations

The x-Wait-Freedom Progress Condition

Imbs

Raynal

2010

Euro-Par 2010 - Parallel Processing

View full text Add to dashboard Cite

The liveness of concurrent objects despite asynchrony and failures is a fundamental problem. To that end several progress conditions have been proposed. Wait-freedom is the strongest of these conditions: it states that any object operation must terminate if the invoking process does not crash. Obstruction-freedom is a weaker progress condition as it requires progress only when a process executes in isolation for a long enough period.This paper explores progress conditions in n-process asynchronous read/write systems enriched with base objects with consensus number x, 1 < x ≤ n (i.e., objects that wait-free solve consensus in a set of x processes). It is easy to solve consensus in such a system if progress is required only when one of the x processes allowed to access the underlying consensus object invokes this object and does not crash. This paper proposes and investigates a stronger progress condition that we call x-wait-freedom (n-wait-freedom is wait-freedom). While it does not need more assumptions than the previous one in order to ensure progress, that condition identifies additional scenarios in which progress is required despite the fact that none of the x processes allowed to access the underlying consensus object participates. The paper then presents and proves correct a consensus algorithm that satisfies this progress condition.

show abstract

Section: Xcons [0]mentioning

confidence: 99%

Section: Xcons [0]mentioning

confidence: 99%

See 1 more Smart Citation

The x-Wait-Freedom Progress Condition

Imbs

Raynal

2010

Euro-Par 2010 - Parallel Processing

View full text Add to dashboard Cite

show abstract

“…In the contemporary chips of these architectures, a read operation can atomically read 128 bytes. In general, such a multi-word read operation can be implemented as an atomic snapshot using only single-word read and single-word write primitives 9 [1]. FindHead nds the item start latest with the highest timestamp in start and searches for the head from locator start latest .ptr by following the next pointers until it nds a locator H whose next pointer is ⊥ (lines 3F-6F).…”

Section: Algorithmmentioning

confidence: 99%

NB-FEB: A Universal Scalable Easy-to-Use Synchronization Primitive for Manycore Architectures

Tsigas

Anshus

2009

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper addresses the problem of universal synchronization primitives that can support scalable thread synchronization for largescale manycore architectures. The universal synchronization primitives that have been deployed widely in conventional architectures, are the compare-and-swap (CAS) and load-linked/store-conditional (LL/SC) primitives. However, such synchronization primitives are expected to reach their scalability limits in the evolution to manycore architectures with thousands of cores. We introduce a non-blocking full/empty bit primitive, or NB-FEB for short, as a promising synchronization primitive for parallel programming on manycore architectures. We show that the NB-FEB primitive is universal, scalable, feasible and easy to use. NB-FEB, together with registers, can solve the consensus problem for an arbitrary number of processes (universality). NB-FEB is combinable, namely its memory requests to the same memory location can be combined into only one memory request, which consequently makes NB-FEB scalable (scalability). Since NB-FEB is a variant of the original full/empty bit that always returns a value instead of waiting for a conditional ag, it is as feasible as the original full/empty bit, which has been implemented in many computer systems (feasibility). We construct, on top of NB-FEB, a non-blocking software transactional memory system called NBFEB-STM, which can be used as an abstraction to handle concurrent threads easily. NBFEB-STM is space e cient: the space complexity of each object updated by N concurrent threads/transactions is Θ(N ), which is optimal.

show abstract

“…The first known wait-free implementations of snapshot from atomic registers [1,2,6] required Θ(n 2 ) steps to carry out a snapshot with n processes; subsequent work [7,8] reduced this cost to O(n), which was shown to be optimal in the worst case for non-blocking deterministic algorithms by Jayanti et al [9].…”

Section: Introductionmentioning

confidence: 99%

Atomic Snapshots in O(log3 n) Steps Using Randomized Helping

Aspnes

Censor-Hillel

2013

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract.A randomized construction of unbounded snapshots objects from atomic registers is given. The cost of each snapshot operation is O(log 3 n) atomic register steps with high probability, where n is the number of processes, even against an adaptive adversary. This is an exponential improvement on the linear cost of the previous best known unrestricted snapshot construction [7,8] and on the linear lower bound for deterministic constructions [9], and does not require limiting the number of updates as in previous sublinear constructions [4]. One of the main ingredients in the construction is a novel randomized helping technique that allows out-of-date processes to obtain up-to-date information without running into covering lower bounds.

show abstract

Atomic snapshots of shared memory

Cited by 118 publications

References 19 publications

The x-Wait-Freedom Progress Condition

The x-Wait-Freedom Progress Condition

NB-FEB: A Universal Scalable Easy-to-Use Synchronization Primitive for Manycore Architectures

Atomic Snapshots in O(log3 n) Steps Using Randomized Helping

Contact Info

Product

Resources

About