Towards Lockfree Persistent Homology

10Citations
Citations of this article
6Readers
Mendeley users who have this article in their library.

Abstract

Persistent homology, which describes the shape of data by quantifying the sizes of its topological features, is one of the most ubiquitous algorithms in topological data analysis. All existing algorithms that compute persistence in parallel rely on the algebraic structure of the problem to subdivide the computation, either by partitioning the range, or the domain of the underlying scalar measurement. Instead, we exploit the inherent parallelism of the reduction algorithm and rely on hardware synchronization primitives, namely compare-and-swap operations, to develop a lockfree shared-memory algorithm that avoids having to decide how to partition the underlying data set. We demonstrate the algorithm's performance and scaling using a set of computational experiments.

Author supplied keywords

Cite

CITATION STYLE

APA

Morozov, D., & Nigmetov, A. (2020). Towards Lockfree Persistent Homology. In Annual ACM Symposium on Parallelism in Algorithms and Architectures (pp. 555–557). Association for Computing Machinery. https://doi.org/10.1145/3350755.3400244

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free