Concurrent Deterministic Skiplist and Other Data Structures

TL;DR

This paper proposes hierarchical usage of concurrent data structures in programs to improve memory latencies by reducing memory accesses from remote NUMA nodes.

Abstract

Skiplists are used in a variety of applications for storing data subject to order criteria. In this article we discuss the design, analysis and performance of a concurrent deterministic skiplist on many-core NUMA nodes. We also evaluate the performance of concurrent lock-free unbounded queue implementation and two concurrent multi-reader,multi-writer(MWMR) hash table implementations and compare them with those from Intel's Thread Building Blocks(TBB) library. We introduce strategies for memory management that reduce page faults and cache misses for the memory access patterns in these data structures. This paper proposes hierarchical usage of concurrent data structures in programs to improve memory latencies by reducing memory accesses from remote NUMA nodes.

Figures (9)

• Figure 1: Deterministic Skiplist with 3 levels
• Figure 2: Lock-free Queue : Memory Allocation
• Figure 3: Lock-free Queues : Performance for 100m and 1b operations
• Figure 4: Skiplist Performance 10m operations
• Figure 5: Skiplist Performance for 100m operations (Inserts,Finds,Erases)