Search

We consider the following generalization of binary search in sorted arrays to tree domains. In each step of the search, an algorithm is querying a vertex $q$, and as a reply, it receives an answer, which either states that $q$ is the desired target, or it gives the neighbor of $q$ that is closer to the target than $q$. A further generalization assumes that a vertex-weight function $\omega$ gives the query costs, i.e., the cost of querying $q$ is $\omega(q)$. The goal is to find an adaptive search strategy requiring the minimum cost in the worst case. This problem is NP-complete for general weight functions and one of the challenging open questions is whether there exists a polynomial-time constant factor approximation algorithm for an arbitrary tree? In this work, we prove that there exist a constant-factor approximation algorithm for trees with a monotonic cost function, i.e., when the tree has a vertex $v$ such that the weights of the subsequent vertices on the path from $v$ to any leaf give a monotonic (non-increasing or non-decreasing) sequence $S$. This gives a constant factor approximation algorithm for trees with cost functions such that each such sequence $S$ has a fixed number of monotonic segments. Finally, we combine several earlier results to show that the problem is NP-complete when the number of monotonic segments in $S$ is at least $4$.

High performance large scale graph analytics is essential to timely analyze relationships in big data sets. Conventional processor architectures suffer from inefficient resource usage and bad scaling on graph workloads. To enable efficient and scalable graph analysis, Intel developed the Programmable Integrated Unified Memory Architecture (PIUMA). PIUMA consists of many multi-threaded cores, fine-grained memory and network accesses, a globally shared address space and powerful offload engines. This paper presents the PIUMA architecture, and provides initial performance estimations, projecting that a PIUMA node will outperform a conventional compute node by one to two orders of magnitude. Furthermore, PIUMA continues to scale across multiple nodes, which is a challenge in conventional multinode setups.