Lock-Free and Practical Doubly Linked List-Based Deques Using Single-Word Compare-and-Swap

Abstract

Abstract. We present an efficient and practical lock-free implementation of a concurrent deque that supports parallelism for disjoint accesses and uses atomic primitives which are available in modern computer systems. Previously known lock-free algorithms of deques are either based on non-available atomic synchronization primitives, only implement a subset of the functionality, or are not designed for disjoint accesses. Our algorithm is based on a general lock-free doubly linked list, and only requires single-word compare-and-swap atomic primitives. It also allows pointers with full precision, and thus supports dynamic deque sizes. We have performed an empirical study using full implementations of the most efficient known algorithms of lock-free deques. For systems with low concurrency, the algorithm by Michael shows the best performance. However, as our algorithm is designed for disjoint accesses, it performs significantly better on systems with high concurrency and non-uniform memory architecture. In addition, the proposed solution also implements a general doubly linked list, the first lock-free implementation that only needs the single-word compare-and-swap atomic primitive.

Efficient & Lock-Free Modified Skip List in Concurrent Environment

“…Our algorithm based on [11] . In this paper we present a lock-free algorithm of a concurrent modified skip list that is designed for efficient use in both preemptive as well as in fully concurrent environments.…”

“…Insert_node ( d) inserts adds d to the set and returns true iff d was not already in the set; del_node (v) removes v from the set and returns true iff v was in the set, the below Figure 4 & 5 shows the field of a node. Using the strategy of [11]. To insert or delete a node from the modified skip list we have to change the respective set of prev and next pointers.…”

“…These pointers have to be changed consistently, but it is not necessary to change them at once. According to Sundell & Tsigas [11] we can consider the doubly linked list as being a singly linked list with auxiliary information in the left pointers, with the right pointers being updated before the left pointers. Thus, the right pointers always form a consistent singly linked list and thus define the nodes positional relations in the logical abstraction of the doubly linked list, but the left pointers only give hints as to where to find the previous node.…”

Lock-free deques and doubly linked lists

“…Sundell and Tsigas [13,14] presented a practical lock-free deque implementation which can also support general doubly linked list operations. 2 The basic idea is to utilize the built-in redundancy of the doubly linked list and allow temporary inconsistency of the data structure while operations are performed on it.…”

“…However, the doubly linked list algorithm in [15] does not support traversals over deleted nodes 3 and do not guarantee that the underlying data structure is consistent 4 in both directions when it is idle. Recently Heller et al [18] showed that it can be beneficial for linked list structures to allow concurrent threads to traverse over locked nodes (i.e., logically deleted nodes), a technique also exploited in a lock-free manner in [14,19].…”