arXiv Science⌕ Search
Home/Archive/arXiv

arXiv · 2610.02385

Fine-Grained Analysis of SIMD-Based Hash Table Implementations

Cyril Nicaud·Pablo Rotondo
View original source ↗Read source document ↗Export metadata ↓

Abstract

In recent years, several variants of classical hash table schemes have been developed by engineers in order to take advantage of the processor's internal parallelism using SIMD instructions, which make it possible to operate on multiple bytes simultaneously. At a small additional memory cost, this enables a significant speedup, making it a data structure that is increasingly popular in practice, when very high performance is required. In this article, we provide a detailed theoretical analysis of the dynamics of such hash tables. From a methodological standpoint, we use and adapt a technique developed by Wormald in the 1990s to study dynamic graphs. This approach, which can be adapted to many variants, enables us to accurately estimate the quantities of interest by capturing the dynamics of the data structure through systems of differential equations. Although complex, we provide an explicit description of the solutions of these systems, which can furthermore be efficiently approximated numerically. Our main results are stated with high probability, which is significantly more precise than average-case analyses, and they match experimental results remarkably well, even for hash tables of moderate size.

Explore related subjects

cs.DS

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Cyril Nicaud, Pablo Rotondo. 2026-10-01. Fine-Grained Analysis of SIMD-Based Hash Table Implementations. https://arxiv.org/abs/2610.02385

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Parameterized Complexity of Temporal Connected Components

We study the parameterized complexity of maximum temporal connected components (tccs) in temporal graphs, that is, graphs whose edges are available only at specific points in time. In a tcc, every pair of vertices must be able to reach one another via time-respecting paths. We consider both maximum open tccs (openTCC), which allow temporal paths through vertices outside the component, and closed tccs (closedTCC), which require at least one temporal path entirely within the component for every pair of vertices. We perform a comprehensive study of the openTCC and closedTCC problems with respect to both structural parameters (treewidth, pathwidth, vertex cover number) and a temporal parameter (temporal path number). We show that the exact complexity, i.e., paraNP-hardness vs XP-tractability, depends on both whether we seek an open or closed tcc and on whether the temporal graph is directed or not. Vertex cover number suffices for XP algorithms for both openTCC and closedTCC on undirected temporal graphs only, while temporal path number suffices only for openTCC in both directed and undirected temporal graphs. Our results are tight: every XP algorithm is complemented by a matching W[1]-hardness result, and for every other case we prove NP-hardness for small constant values of the parameters even on planar graphs. Finally, we prove that both problems become fixed-parameter tractable on both directed and undirected graphs when parameterized by treewidth and temporal path number together.

cs.DS↗
arXiv2026

Temporal Graph Reconfiguration for Always-Connected Graphs

Network redesign problems ask for modifications to the edges of a given graph to satisfy certain properties. In temporal graphs, where edges are only active at certain times, we are sometimes only allowed to modify when the edges are going to be active. In practice, we might not even be able to perform all of the necessary modifications at once; changes must be applied step-by-step while the network is still in operation, meaning that the network must continue to satisfy some properties. To initiate a study in this area, we introduce the class of temporal graph reconfiguration problems. As a starting point, we consider the Layered Connectivity Reconfiguration (LCR) problem: Given two always-connected temporal graphs G1 and G2, determine if it is possible to transform G1 into G2 by changing the time at which a single temporal edge is active in each step, such that every intermediate temporal graph is always-connected. We provide a dynamic programming algorithm for the LCR problem. We also show that finding the shortest reconfiguration sequence between two temporal graphs cannot be approximated in polynomial time within a factor of c log M for some c > 0, unless P = NP. Additionally, we show that the LCR problem is equivalent to the Spanning Tree Sequence Reconfiguration (STSR) problem introduced by Hanaka et al. Therefore, our results also answer two questions raised by the authors: (i) find a simpler algorithm for the STSR problem, (ii) show that the STSR problem is inapproximable up to a factor of c log M.

cs.DS↗
arXiv2026

Model-agnostic distribution recovery in high dimension via Fourier measurements

We study the problem of reconstructing an unknown probability distribution to high accuracy, given noisy information about its low-degree Fourier coefficients. Prior results for this basic problem have typically imposed strong modeling assumptions on the distribution, often requiring that it is a linear combination of spatially separated point sources. In this work, we analyze a very general version of the high-dimensional recovery problem, by considering completely general distributions over the $d$-dimensional torus $[0,1)^d$; we do not assume any spatial separation between point sources, or even that the distribution is a finite linear combination of point sources. The question naturally arises: what can be said about distribution recovery in such a general setting? - As a warm-up, we first give results for reconstructing distributions with respect to the Wasserstein distance. We establish essentially matching upper and lower bounds on the cutoff frequency $T$ and the magnitude $κ$ of the noise for which accurate reconstruction is possible. Roughly speaking, our results here show that for $d$-dimensional distributions, estimates of $\approx \exp(d)$ many Fourier coefficients are necessary and sufficient for accurate reconstruction under the Wasserstein distance. - As our main result, we define a new notion of "heavy hitter" reconstruction for distributions, which essentially amounts to achieving high-accuracy reconstruction of all "sufficiently dense" regions of the distribution. We give essentially matching upper and lower bounds on the cutoff frequency $T$ and the magnitude $κ$ of the noise for which accurate reconstruction is possible under this notion. Our results show that (in sharp contrast with Wasserstein reconstruction) accurate estimates of only $\approx \exp(\sqrt{d})$ many Fourier coefficients are necessary and sufficient for heavy hitter reconstruction.

cs.DS↗