Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2108.11099

Towards Informed Partitioning for Load Balancing: a Proof-of-Concept

Anthony Boulmier·Nabil Abdennadher·Bastien Chopard
View original source ↗Read source document ↗Export metadata ↓

Abstract

Most parallel applications suffer from load imbalance, a crucial performance degradation factor. In particle simulations, this is mainly due to the migration of particles between processing elements, which eventually gather unevenly and create workload imbalance. Dynamic load balancing is used at various iterations to mitigate load imbalance, employing a partitioning method to divide the computational space evenly while minimizing communications. In this paper, we propose a novel partitioning methodology called ``informed partitioning''. It uses information based on the evolution of the computation to reduce the load balancing growth and the number of load balancing calls. We illustrate informed partitioning by proposing a new geometric partitioning technique for particles simulations. This technique is derived from the well-known recursive coordinate bisection and employs the velocity of the particles to guide the bisection axis. To properly compare the performance of our new method with existing partitioning techniques during application execution, we introduce an effort metric based on a theoretical model of load balanced parallel application time. We propose a proof-of-concept of informed partitioning, through a numerical study, on three N-Body simulations with various particle dynamics, and we discuss its performance against popular geometric partitioning techniques. Moreover, we show that our effort metric can be used to rank partitioning techniques by their efficiency at any time point during the simulation. Eventually, this could be used to choose the best partitioning on the fly. In the numerical study, we report that our novel concept increases the performance of two experiments out of three by up to 76% and 15%, while being marginally slower by only $3\%$ in one experiment. Also, we discuss the limitations of our implementation of informed partitioning and our effort metric.

Explore related subjects

cs.DC

Keep this discovery

BibTeXRIS

Anthony Boulmier, Nabil Abdennadher, Bastien Chopard. 2021-08-25. Towards Informed Partitioning for Load Balancing: a Proof-of-Concept. https://arxiv.org/abs/2108.11099

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

KEEP EXPLORING

Related papers

arXiv2026

Online Treasure Hunt in Vertex-Permuted Dynamic Rings

We study the problem of treasure hunt by a group of $k \geq 1$ agents in vertex-permuted dynamic rings (VP). In this model, the $n$ vertices remain on a ring but are permuted at each time step. We first show that treasure hunt is impossible for any $k \leq n-3$ agents, if there are no restrictions on the sequence of permutations used in the dynamic ring. We then study the $VP(\delta)$ setting, in which for every pair $i, j$ of vertices, the edge $(i, j)$ is guaranteed to appear within $\delta$ steps. We show that the class $VP(\delta)$ is feasible only for $\delta \geq \left\lceil \frac{n-1}{2}\right\rceil$. For the one-agent case, we show a tight bound of $\Theta(\delta n)$ on the worst-case search time as well as competitive ratio of any online algorithm for treasure hunt, provided $\delta \geq 2n$. We then give an optimal algorithm for $k$ agents, thereby showing that $k$ agents can obtain a speedup of $k$ on the worst-case search time. Finally, in the R-VP setting, in which in every step, the vertices are arranged as a ring according to a random permutation, we show that treasure hunt takes expected $\Theta(n)$ steps against an oblivious adversary and $\Theta(n \log n)$ steps against an adaptive adversary.

cs.DC↗
arXiv2026

The Computing Channel: How Modulation Programs the Airwaves

Distributed computing and distributed artificial intelligence require frequent exchanges of intermediate results, although many applications need only an aggregate rather than messages from individual devices. Conventional systems recover each message before computing the aggregate, whereas over-the-air computation (OAC) exploits simultaneous transmission to obtain it directly. However, dominant OAC implementations rely on analog signaling, creating a mismatch with finite-precision data and digital communication procedures. This article presents digital function-oriented communication, in which finite-alphabet symbol representations and receiver decisions are jointly designed so that multiple-access superposition encodes the desired function without recovering individual inputs. We introduce its computational-constellation principle, main design approaches, extensions, and implementation challenges. Federated edge learning illustrates how the framework can reduce user-dependent data-bearing resources while operating directly on quantized model updates.

cs.DC↗
arXiv2026

Can AI Remediate Backend Failures Safely? GuardedAct with Blast-Radius-Aware Sandboxing

Large Language Models (LLMs) have shown promising capabilities in generating remediation actions for microservice failures. However, directly executing AI-generated repair actions in production risks cascading collateral damage. We propose GuardedAct, a sandbox-first remediation framework that interposes a blast-radius-aware verification layer between the LLM action generator and the production environment. GuardedAct operates in four phases: (1) ingesting a diagnosis report together with the live system topology and recent telemetry, (2) prompting an LLM to produce a ranked list of candidate remediation actions, (3) simulating each action in a lightweight digital-twin sandbox that estimates the blast radius and assigns a risk label, and (4) enforcing a rollback-confidence gate that auto-executes only low-risk actions while escalating high-risk ones for human review. We evaluate GuardedAct on five fault scenarios injected into the DeathStarBench social-network application. Experimental results show that GuardedAct achieves an overall recovery rate of 87.4% while reducing collateral damage by 79.7% relative to direct LLM execution (from 25.6% to 5.2%), at the cost of a modest sandbox-induced increase in mean time to recovery (approximately 8 s). Ablation studies confirm that each component contributes meaningfully to the safety-speed trade-off.

cs.DC↗