Towards a Complexity Classification of LCL Problems in Massively Parallel Computation

TL;DR

This paper establishes a complexity classification for locally checkable labeling problems on trees within the low-space Massively Parallel Computation model, demonstrating exponential speed-ups over traditional models and identifying fundamental limits.

Contribution

It introduces a general method for solving LCL problems faster in MPC than in LOCAL, achieving $O( ext{log} n)$ and $O( ext{log} ext{log} n)$ time algorithms, and provides new techniques like a tree rooting algorithm and pointer-chain analysis.

Findings

All solvable LCL problems on trees can be solved in $O( ext{log} n)$ time in MPC.

LCL problems with $n^{o(1)}$ complexity in LOCAL can be solved in $O( ext{log} ext{log} n)$ in MPC.

Conditional lower bounds show no faster algorithms are likely for certain problems like 3-coloring trees.

Abstract

In this work, we develop the low-space Massively Parallel Computation (MPC) complexity landscape for a family of fundamental graph problems on trees. We present a general method that solves most locally checkable labeling (LCL) problems exponentially faster in the low-space MPC model than in the LOCAL message passing model. In particular, we show that all solvable LCL problems on trees can be solved in $O(\log n)$ time (high-complexity regime) and that all LCL problems on trees with deterministic complexity $n^{o(1)}$ in the LOCAL model can be solved in $O(\log \log n)$ time (mid-complexity regime). We observe that obtaining a greater speed-up than from $n^{o(1)}$ to $\Theta(\log \log n)$ is conditionally impossible, since the problem of 3-coloring trees, which is a LCL problem with LOCAL time complexity $n^{o(1)}$, has a conditional MPC lower bound of $\Omega(\log \log n)$ [Linial, FOCS'87; Ghaffari, Kuhn and Uitto, FOCS'19]. We emphasize that we solve LCL problems on constant-degree trees, and that our algorithms are deterministic, component-stable, and work in the low-space MPC model, where local memory is $O(n^\delta)$ for $\delta \in (0,1)$ and global memory is $O(m)$. For the high-complexity regime, there are two key ingredients. One is a novel $O(\log n)$-time tree rooting algorithm, which may be of independent interest. The other is a novel pointer-chain technique and analysis that allows us to solve any solvable LCL problem on trees in $O(\log n)$ time. For the mid-complexity regime, we adapt the approach by Chang and Pettie [FOCS'17], who gave a canonical LOCAL algorithm for solving LCL problems on trees.