Peer Revenue Publish or profit

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Distributed Computing

cs.DC Watch Distributed Computing

Thursday 27 August

Distributed Computing

VPP: Virtual Pipeline Parallelism for Efficient Chunked Prefill in Long-Context LLM Inference

Yan Shi, Xiaochao Wang, Jingchun Gao et al. (+7 more)

Chunked prefill pipeline parallelism (CPP) is a key technique for LLM inference. However, equal-size chunks exhibit imbalanced latency, as later chunks attend longer prefix KV caches and incur higher attention costs, leading to pipeline bubbles. Existing approaches mitigate this imbalance through dynamic chunk resizing (Dynamic CPP, DCPP), but our measurements show that this trades scheduling overhead for load balancing, which becomes unfavorable on long sequences. In this study, we propose Virtual Pipeline Parallelism (VPP), which keeps chunk sizes fixed and optimizes the pipeline layout through virtual stages. A V-shaped virtual-stage traversal overlaps each chunk's expensive middle stages with the lighter head and tail stages of its neighbors, while asynchronous communication and pipelined packing further reduce communication stalls and cross-request drain bubbles. We implement VPP in vLLM-Ascend and evaluate it on three MoE-based LLMs with sequences up to 1M tokens on 16 Ascend 910C NPUs. VPP improves throughput by up to 13.1% over DCPP on long sequences and 6.7% on mixed workloads, while preserving performance on short sequences. On a 512K-token DeepSeek-V3.1 prefill workload, VPP reduces the pipeline bubble ratio from 6.4% to 0.1%, achieving a 98.0% reduction compared with DCPP.
Read the full abstract

All authors Yan Shi, Xiaochao Wang, Jingchun Gao, Jintao Luo, Xinyi Zhou, Feng Liu, Kui Luo, Xushi Li, Xinjie Guo, Liangjun Feng

Wednesday 26 August

Distributed Computing · Cross-listed from cs.LG

Distributed Training using an Intelligent Network

Nihar Shah and Ben Blier

Distributed training across a wide area network (WAN) is challenging, as continuous parameter exchange by islands of compute is constrained by limited bandwidth, high latency, and uneven topology. We propose making the network an active participant in training. On the systems side, such networks should leverage (i) multicast technology to replicate outbound traffic and (ii) in-line FPGAs to aggregate inbound traffic, to ease egress and ingress bottlenecks. These technologies are used for training across workers within a data center, but this paper extends them to the WAN. On the algorithms side, we develop an optimization framework that produces rich synchronization schedules (namely, rotating cliques of islands) around the underlying network topology and these technologies, to maximize information exchange. Finally, we illustrate this on a nine-city topology modeled on the DoubleZero network, a live programmable WAN equipped with both technologies, and show how the optimal schedules shift with the network's capabilities. Together, these can narrow the gap to the gold standard of colocated training.
Read the full abstract

Distributed Computing

Multi-Dataset Inverse Problem Solving with Distributed Generative AI

Daniel Lersch, Steven Goldenberg, Johann Rudi et al. (+5 more)

Extracting a shared set of unknown, not directly measurable quantities from multiple, heterogeneous datasets is a common challenge across scientific domains. A prominent example is the combination of datasets obtained from different measurements with different settings (e.g. varying detector resolutions). Analyzing such datasets jointly, rather than independently or after naive merging, is essential for obtaining precise and unbiased estimates of the unknowns, but requires careful treatment of dataset heterogeneity and is computationally demanding. We present a generalized framework for simultaneously analyzing multiple heterogeneous datasets in the context of generative AI-based inverse problem solvers. Building on our recent Scalable Asynchronous Generative Inverse Problem Solver (SAGIPS) framework, we extend the well-established distributed data-parallel training paradigm to non-identically distributed datasets, where each dataset is controlled by the same set of unknown inference parameters but covers a different region of the available feature space. Each dataset is processed through its own forward operator and discriminator, providing complementary constraints that collectively guide a shared generator toward global parameter consistency. We validate the approach using a controlled setup inspired by a multi-detector scattering experiment. We provide numerical evidence that our framework is robust to different data fidelities, which arise from unknown detector systematics in the Rutherford experiment, and we show the scaling behavior on multi-GPU leadership computing systems. The results show that our approach is well suited for real-world multi-dataset analyses in which experimental conditions vary across measurements.
Read the full abstract

All authors Daniel Lersch, Steven Goldenberg, Johann Rudi, Markus Diefenthaler, Kevin Brager, Xingfu Wu, Yaohang Li, Nobuo Sato

Distributed Computing · Cross-listed from cs.AI

Agent Mesh: Reliability Primitives for Non-Idempotent Agent Delegation - Identity Adequacy and Evidence Adequacy

Mazhar Shaikh, Anurag Rajkumar Bombarde and Harshal Pathak

Autonomous agents increasingly perform bounded software tasks under an orchestrator that retries, resumes, and budgets them. The machinery such orchestrators reach for is the service mesh's: retry, timeout, and error-rate circuit breaking. We report a failure study of a production agentic software-delivery platform over 147 numbered incidents spanning 81 runs, each with a measured cost and, in most cases, a mutation proof reproducing the failure. All three assumptions those primitives rest on are violated in practice, and we quantify the consequences: a loop of fifty-four consecutive successful tool calls no error-rate breaker could see; a progress signal constant by construction, guaranteeing a false trip on the third repair round and driving one run from six of six components to three; twenty-one events accumulated across six invocations of one delegation, making a correct, idempotent component unwinnable; a misrouted failure that woke five components for a two-component fault, leaving three bystanders regressing working code; and twelve incidents in which the enforcement layer blocked correct work, the most expensive costing 107 agent turns and zero accepted writes. We find one cross-cutting cause and its dual. Identity adequacy: in five separate subsystems an identity that failed to discriminate produced a confident wrong answer, and two of them derived the corrective rule independently. Evidence adequacy: a reliability decision may be taken only on evidence capable of moving, attributable to what it measures, and deterministic under identical conditions. From the findings we derive seven reliability primitives whose enforcement unit is the delegation rather than the message, and specify the controlled evaluation the study motivates but does not constitute.
Read the full abstract

Distributed Computing · Cross-listed from cs.AR

MeshReduce-U: Compiler-Guided Communication Reduction for Irregular Neural Reductions on Mesh NoCs

Amirreza Khorasanian

Many irregular neural workloads induce skewed many-to-one reductions with repeated neighborhoods and nonlocal communication. Conventional NoC mappers optimize placement and routes for a fixed communication graph, even though associative reductions expose legal opportunities to eliminate traffic before routing. We present MeshReduce-U, a compiler-guided communication-reduction and routing framework for mesh-NoC-based spatial accelerators. MeshReduce-U coalesces colocated sources, forms local aggregation islands, blocks channels with compatible fan-in structure, selects capacity-feasible sinks, and routes the remaining fixed-width carriers using fused usage-aware costs. A deterministic route-replay model reports schedule-derived communication latency, total link usage (TLU), and fused link usage (FusedTLU) separately. Across a 20-workload lowerable neural-network zoo, MeshReduce-U reduces mean latency, TLU, and FusedTLU by 40.3%, 56.0%, and 48.7%, respectively, relative to an ABC-style source-aggregation baseline, improving all three metrics on every workload. Across 40 synthetic irregular reductions, it reduces mean latency and TLU by 12.3% and 19.7%. A new 30-instance pass-by-pass study further shows that the structural rewrites reduce the global carrier count by 60.9% and replay latency by 63.0%. These results show that rewriting reducible neural communication before routing can be more effective than searching harder over an unreduced traffic graph.
Read the full abstract

Distributed Computing · Cross-listed from eess.SP

Adaptive Peer Clustering with Hierarchical Random Linear Network Coding for Resilient Decentralized Wireless Networks

Navaneetha Krishnan Kamalakannan and Harinisri Velmurugan

Decentralized wireless collectives including vehicular swarms, IoT clusters, and edge AI networks require communication protocols that maintain robustness under dynamic topologies and heterogeneous link quality. While Random Linear Network Coding (RLNC) provides algebraic resilience against packet erasures, its performance degrades significantly when peers exhibit diverse channel conditions. This paper presents Adaptive Peer Clustering with Hierarchical RLNC (APC-RLNC), a system that dynamically groups peers by exponentially weighted moving average (EWMA) reliability metrics and applies multi-tier network coding within and across clusters. We formalize the clustering optimization problem, derive closed-form decoding probability bounds for Markov erasure channels, and prove O(sqrt(T)) regret for online reconfiguration under the Follow-the-Regularized-Leader (FTRL) framework. Our implementation includes both a high-fidelity network simulator and a proof-of-concept testbed deployment on Jetson Nano edge devices. Evaluation across diverse scenarios including high-mobility vehicular networks, burst-error channels, and adversarial interference demonstrates 5.2-9.8 percentage-point packet delivery ratio (PDR) improvements, 10-23% latency reductions, and up to 30% higher node retention compared to state-of-the-art baselines. The system exhibits linear scalability to 500+ nodes and maintains real-time reconfiguration overhead below 3%. APC-RLNC establishes adaptive clustering as a foundational primitive for AI-native 6G wireless systems.
Read the full abstract

Distributed Computing

Slasher: Power Flexibility for Cloud Datacenters

Liuzixuan Lin, Fiodar Kazhamiaka, Alok Gautam Kumbhare et al. (+10 more)

Datacenters consume many megawatts of power, and regularly encounter scenarios that require modulating their power draw. These scenarios include datacenter infrastructure failures, power grid failures, grid services, and more, spanning a diverse range of requirements in terms of the power magnitude, the scope of the reduction, the notice time, and other dimensions. To address these scenarios, we have built Slasher, a general system for modulating the power of \azure datacenters to handle scenarios ranging from individual racks to regional multi-datacenter grid events. Slasher coordinates datacenter resources with the goal of meeting power targets while minimizing negative impact on hosted workloads. In this paper, we review the main power modulation scenarios, characterize the power reduction levers using data from production cloud datacenters, describe Slasher's system architecture, and formulate the cloud datacenter power modulation control problem. We also develop a high-fidelity datacenter simulator and propose a workload impact model, using them to design and evaluate power control algorithms.
Read the full abstract

All authors Liuzixuan Lin, Fiodar Kazhamiaka, Alok Gautam Kumbhare, Chaojie Zhang, Jaylen Wang, Hassan Khan, Rodrigo L. Assis, Mariana Rodrigues, Kyle Woolcock, Nithish Mahalingam, Brijesh Warrier, Rodrigo Fonseca, Ricardo Bianchini

Distributed Computing

Scalable Multi-GPU Simulation of 3D Multicellular Growth with RNN-Based Workload Balancing

Matvey Moisseyev, Huijing Du, Dandan Zheng et al. (+2 more)

Detailed multicellular growth simulations based on subcellular element models (SEMs) can capture complex tissue development, but their element-level interactions impose substantial computational cost. This work presents a scalable multi-GPU framework for 3D multicellular growth simulation that combines GPU acceleration, spatial binning, domain decomposition, and workload-aware partitioning. Cell movement, growth, and division continuously reshape the spatial workload distribution, causing initially balanced partitions to become inefficient over time. To address this, we introduce an RNN-based load-balancing controller that observes recent per-rank execution times and partition states and learns residual corrections to a reactive boundary-adjustment rule. The controller is trained offline in a differentiable surrogate of the load-balancing loop with randomized workload dynamics, requiring no measured execution traces for training. We evaluate the framework in terms of single-GPU acceleration, multi-GPU computation scaling, controller-level load-balancing behavior, and end-to-end simulation performance, with comparisons against static partitioning, reactive load balancing, and conventional time-series prediction baselines. A representative embryonic epidermal development use case further demonstrates the type of spatially and temporally evolving workload targeted by the framework. In our evaluation, GPU acceleration with spatial binning accelerates the interaction computation by roughly three orders of magnitude over a serial CPU baseline. RNN-guided load balancing reduces the mean global imbalance from 11.3% under static partitioning to 3.5%, lowers end-to-end runtime by 9.0% relative to static partitioning, and reduces slice migration by 7.7x compared with the reactive baseline, showing that history-aware control can improve workload balance while avoiding unnecessary repartitioning.
Read the full abstract

All authors Matvey Moisseyev, Huijing Du, Dandan Zheng, Chi Zhang, Hongfeng Yu

Distributed Computing · Cross-listed from cs.LG

Cooperative Multi-Agent Reinforcement Learning for Adaptive Aggregation in Semi-Supervised Federated Learning with non-IID Data

Rene Glitza, Luca Becker and Rainer Martin

Federated Learning (FL) enables distributed training of machine learning models while preserving data privacy. However, FL struggles with heterogeneous, non-IID client data distributions, resulting in sub-optimal and biased global models. In this paper, we propose pFedMARL, a novel approach leveraging Multi-Agent Reinforcement Learning (MARL) with Twin Delayed Deep Deterministic Policy Gradient (TD3) to dynamically adapt aggregation strategies in FL settings. Our method employs a server-side agent adjusting client contributions to optimize global model robustness and client-side agents balancing global and local updates to personalize models effectively without pre-training. We demonstrate superior performance of pFedMARL for training a semi-supervised audio spectrogram transformer, matching or outperforming FedAvg, Ditto, and local training approaches across multiple non-IID scenarios and in the presence of adversarial clients. Our results indicate that pFedMARL actively improves accuracy, robustness, and fairness, making it suitable for real-world deployments.
Read the full abstract

Distributed Computing

Quantum Blackhole Learning-Optimized Hadamard Neural Network Model for Dynamic Resource Reservation in Industry Clouds

Deepika Saxena, Hari Mohan Gaur, Ashutosh Kumar Singh et al. (+1 more)

Accurate workload prediction and proactive resource reservation are crucial for industry clouds. However, the conventional machine learning (CML) models with limited learning capabilities often fail to predict diverse, high-dimensional workloads with sudden changes in resource demand, leading to excessive power consumption and resource management issues. In this context, this article proposes a novel Hadamard neural network with quantum blackhole (QB-HNN) optimization. This model combines the computational efficiency of quantum mechanics with the persuasive learning capability of neural networks (NNs). The workload information is transformed into qubits and propagated via a deep network of qubit neurons comprising a Hadamard-gated activation function to fetch superposition within the QB-HNN model for intuitive pattern learning. Furthermore, a novel quantum blackhole biphase optimization (QB-BiO) algorithm is introduced to train and optimize qubit neural weights. The performance of the proposed model is comprehensively evaluated and compared with five state-of-the-art approaches using six benchmark datasets of three heterogeneous varieties of cloud workloads. The prediction accuracy achieved for an extensive range of workloads confirms its influential performance by minimizing the prediction error up to 36.36% and 22.83% over existing LSTM- and EQNN-based prediction approaches, respectively.
Read the full abstract

All authors Deepika Saxena, Hari Mohan Gaur, Ashutosh Kumar Singh, Anand Mohan

Distributed Computing

REE-TM: Reliable and Energy-Efficient Traffic Management Model for Diverse Cloud Workloads

Ashutosh Kumar Singh, Deepika Saxena and Volker Lindenstruth

Diversity of workload demands lays a critical impact on efficient resource allocation and management of cloud services. The existing literature has either weakly considered or overlooked the heterogeneous feature of job requests received from wide range of internet services users. To address this context, the proposed approach named Reliable and Energy Efficient Traffic Management (REE-TM) has exploited the diversity of internet traffic in terms of variation in resource demands and expected complexity. Specifically, REE-TM incorporates categorization of heterogeneous job requests and executes them by selecting the most admissible virtual node (a software-defined instance such as a virtual machine or container) and physical node (an actual hardware server or compute host) within the cloud infrastructure. To deal with resource-contention-based resource failures and performance degradation, a novel workload estimator 'Toffoli Gate-based Quantum Neural Network' (TG-QNN) is proposed, wherein learning process or interconnection weights optimization is achieved using Quantum version of BlackHole (QBHO) algorithm. The proactively estimated workload is used to compute entropy of the upcoming internet traffic with various traffic states analysis for detection of probable resource-congestion. REE-TM is extensively evaluated through simulations using a benchmark dataset and compared with optimal and without REE-TM versions. The performance evaluation and comparison of REE-TM with measured significant metrics reveal its effectiveness in assuring higher reliability by up to 30.25% and energy-efficiency by up to 23% as compared without REE-TM.
Read the full abstract

Distributed Computing

A Spatially-Aware Publish-Subscribe Middleware for IoT Applications

Philipp Ungrund, Kurt Rothermel and Sukanya Bhowmik

Spatial and proximity awareness are critical enablers for efficient communication in Cyber-Physical Systems (CPS) and the Internet of Things (IoT). However, the lack of suitable middleware support and standardized mechanisms for spatial awareness significantly limits the pervasiveness of location-dependent applications. In this paper, we present a novel approach that extends topic-based publish-subscribe, a dominant messaging model in this domain, with integrated spatial filtering and routing capabilities. Our approach leverages a highly flexible geometric method for describing spatial filters, enabling expressive and efficient spatial subscriptions. A geometric world model operates together with an enhanced topic-based broker to realize spatially constrained message dissemination. Importantly, this functionality is achieved using the MQTT5 protocol without violating the standard or requiring any modifications to existing client implementations. Experimental results based on both synthetic and real-world data demonstrate strong system efficiency, minimal computational overhead at the broker, and the practical feasibility of deploying the approach on existing IoT devices.
Read the full abstract

Distributed Computing

An Oversubscription and Service Pricing Exploitation-Based Profit Maximization Framework for Industry Cloud Resource Management

Deepika Saxena and Ashutosh Kumar Singh

This article proposed a novel industry cloud resource management framework that exploits resource oversubscription and heterogeneous service pricing models to maximize profitability and operational efficiency for industry cloud providers. The framework proposes an adaptive ensemble machine learning driven prediction model for proactive estimation of resource utilization of Virtual Machines (VM)s based on previous resource utilization of respective users' VMs to minimize resource wastage due to oversubscription by them. Accordingly, the VMs having similar predicted resource usage are grouped using Fuzzy C means clustering. This helps to determine the required number of VMs with specific configuration to be deployed before executing user requests. Concurrently, the framework incorporates two distinct categories of cloud service pricing models, namely the Delay Sensitive Model and the Best-Effort Model. Accordingly, the user requests are classified and executed by selecting the most suitable VMs, with the goal of maximizing revenue and reducing electricity costs in cloud data centers (CDCs). Experimental simulation and comparison against state-of-the-art methods, using two benchmark VM traces, validates the performance of proposed framework. It significantly reduces electricity bills by 55.56 percentage, power consumption and active servers by up to 60.7 percentage and 51 percentage, respectively, while improving resource utilization and profits by up to 60 percentage and 51.18 percentage, respectively
Read the full abstract

Distributed Computing

psRL: Efficient Training for Agentic AI via Training-Time Prefix Sharing

Mianjie Yu, Zizhao Mo, Huanyu Qu et al. (+8 more)

In modern agentic AI training, the system bottleneck is shifting from rollout to update. Emerging sampling strategies such as tree-structured and step-wise RL greatly increase training sample volume while incurring relatively low marginal rollout cost, causing the update phase to dominate the end-to-end execution time. Crucially, this shift exposes a new optimization opportunity, as production traces reveal substantial prefix redundancy across training samples. In this paper, we propose psRL (prefix sharing for RL), a new training system for agentic AI designed to exploit prefix redundancy among training samples. Leveraging the global visibility and data immutability inherent to the update phase, psRL achieves efficient workload scheduling and memory management for distributed training. Specifically, psRL introduces two novel prefix-sharing mechanisms that enable flexible, fine-grained workload distribution across GPU workers, simultaneously optimizing prefix reuse and achieving load balancing. Moreover, psRL implements a new underlying KV cache manager that facilitates adaptable block-size allocation and dynamic KV caching, maximizing memory utilization while maintaining a high prefix hit rate. Evaluations using production traces demonstrate that psRL outperforms existing systems by up to 5.2x in throughput. The source code will be publicly available soon.
Read the full abstract

All authors Mianjie Yu, Zizhao Mo, Huanyu Qu, Zhirong Qian, Huanle Xu, Cen Li, Zifeng Zhao, Zhi Zhou, Jinhua Zhou, Jun Xie, Chengzhong Xu

Distributed Computing · Cross-listed from eess.SY

Throughput Maximization for MapReduce-Based Collaborative Computing over Energy-Harvesting Wireless Devices

Yuhang Li, Siqi Sun, Hongen Zheng et al. (+3 more)

This paper studies resource allocation for MapReduce-based collaborative computing over heterogeneous wireless devices powered by renewable energy harvesting. We formulate a long-run average throughput maximization problem that jointly optimizes computing load, phase time allocations, transmit power, and per-device energy consumption, subject to battery evolution, CPU frequency, and latency constraints. To solve this problem online without prior knowledge of channel states or energy arrivals, we propose a DDPG-CVX algorithm that couples Deep Deterministic Policy Gradient (DDPG) with convex programming. DDPG determines the per-slot energy budget for each device from observed battery and channel states; the remaining resource allocation variables are then resolved to global optimality by an embedded convex solver. This two-phase decomposition reduces the action-space dimensionality of DDPG while preserving per-slot solution quality. Simulations show that DDPG-CVX achieves 1.25$\times$$\sim$32.36$\times$ the throughput of representative benchmarks.
Read the full abstract

All authors Yuhang Li, Siqi Sun, Hongen Zheng, Xiaojing Chen, Shunqing Zhang, Yanzan Sun

Distributed Computing · Cross-listed from cs.LG

Resilient Decentralized Wireless Federated Learning via Gradient Tracking with AdamW

Nguyen Van Thieu, Ti Ti Nguyen, Ons Aouedi et al. (+2 more)

Wireless Internet-of-Things (IoT) edge networks require decentralized learning (DecL) methods that can operate reliably under both heterogeneous local data and communication-constrained wireless links. However, existing decentralized optimization schemes often incur substantial communication overhead and degraded performance when transmissions are constrained by strict airtime budgets, fading channels, and packet losses. This paper proposes QEF-GT-AdamW, a communication-efficient and outage-resilient algorithm for DecL over wireless communication (WCom) networks. The proposed method combines gradient tracking to mitigate the effect of non-IID data, AdamW-based adaptive optimization to improve training stability, and dual-stream biased quantization with error feedback to reduce communication payloads for both model and tracking exchanges. To address unreliable broadcast communication, the proposed framework further employs a local fallback strategy when scheduled packets are not successfully received. We explicitly model the effect of bandwidth, transmit power, airtime constraints, and fading channels on DecL performance, and establish convergence guarantees for the proposed algorithm under compressed and unreliable wireless communication. Experimental results on heterogeneous MNIST and CIFAR-10 settings show that QEF-GT-AdamW consistently improves robustness and convergence performance over representative DecL baselines while achieving favorable accuracy-communication trade-offs under limited wireless resources.
Read the full abstract

All authors Nguyen Van Thieu, Ti Ti Nguyen, Ons Aouedi, Vu Nguyen Ha, Symeon Chatzinotas

Distributed Computing

Hierarchical Shared Memory-Aware Optimization for TRSM on GPU Platforms

Xinzhe Chen, Haowei Li, Lijuan Hu et al. (+2 more)

Triangular Solve with Multiple Right-hand Sides (TRSM) is a fundamental BLAS Level-3 operation that underpins LU/Cholesky decomposition, sparse direct solvers, and matrix inversion. In the left-side lower-triangular case studied in this paper, efficient GPU implementation remains challenging because forward substitution introduces strict row-wise dependencies, and shared memory is too scarce to hold both operand matrices for wide data types such as double complex. This paper presents HSMA-TRSM, a hierarchical shared memory-aware optimization framework for left-side lower-triangular TRSM on NVIDIA A100, NVIDIA H800, and Hygon DCU Z100 accelerators. For the small-scale regime (m,n<=64), we design a pipelined compute-memory overlap mechanism through loop unrolling and instruction reordering, and propose a dual thread-group seven-stage pipeline strategy to address shared memory constraints for double complex types. For large-scale problems, we introduce a diagonal block decoupling optimization with an O(IB)shared-memory footprint for diagonal block inversion, enabling adaptive block size selection based on matrix scale and hardware characteristics. A compile-time configuration selection framework based on offline profiling and online lookup selects the optimal block size per platform with zero runtime overhead. Evaluated on NVIDIA A100, H800, and Hygon DCU Z100, HSMA-TRSM achieves peak speedups of 2.05xover cuBLAS and 2.06xover rocBLAS. The gains are strongest in shared-memory-constrained double-complex small cases and in large real-type cases where adaptive blocking improves GEMM-dominated updates, while mature vendor kernels leave less optimization headroom in some regimes.
Read the full abstract

All authors Xinzhe Chen, Haowei Li, Lijuan Hu, Wenjing Ma, Fangfang Liu

Tuesday 25 August

Distributed Computing · Cross-listed from cs.CE

Trade-Size-Aware Dynamic Fees for Impermanent Loss Mitigation in AMMs

Anton Ledrov, Ignat Melnikov, Irina Lebedeva et al. (+3 more)

Automated Market Makers enable decentralized trading but systematically expose liquidity providers to impermanent loss through arbitrage-driven rebalancing. While dynamic fee mechanisms offer a promising mitigation strategy, existing approaches remain largely reactive, adjusting costs based on historical signals rather than explicitly linking them to the structural risk imposed by individual trades. To address this limitation, we propose a novel fee formation framework built on three core innovations. First, we introduce a coupled market maker architecture in which fee dynamics are governed by a secondary invariant, allowing liquidity state and transaction costs to evolve jointly. Second, we develop an impermanent-loss trimming fee model that adaptively increases transaction costs for trades exceeding the liquidity providers' profitable region, effectively offsetting losses from large arbitrage executions while preserving baseline fees for smaller transactions. Third, we establish a unified evaluation methodology using performance profiles to systematically compare fee algorithms across diverse market conditions. By extending a heterogeneous trader model to derive optimal arbitrage strategies under state-dependent fees, we conduct extensive simulations on historical data spanning four distinct market regimes and three token pair categories. Our results demonstrate that the proposed fee enhancements improve liquidity provider yields by 6–24% in volatile markets and up to 119% in calm regimes, while maintaining uninformed user participation and reducing informed arbitrage profitability by 3–10%. Performance profile analysis confirms that ILT-enhanced algorithms dominate baseline counterparts across 60–75% of test scenarios.
Read the full abstract

All authors Anton Ledrov, Ignat Melnikov, Irina Lebedeva, Dmitrii Umnov, George Ovchinnikov, Yury Yanovich

Distributed Computing

When Does Learning Beat Heuristics? A Case Study in Kubernetes Scheduler Score Plugins

Wang Xuying and Zhibek Sarypbekova

Kubernetes scheduler plugins that score candidate nodes are, in production, hand-tuned heuristics. We ask whether a learned scoring function - trained on real placement decisions from a production cluster trace - can match or exceed these heuristics, and if not, why. We implement an external, HTTP-backed scoring plugin for a widely used scheduler simulator and evaluate two learned models (a Random Forest over engineered features, and a graph neural network encoder over per-job task-dependency graphs) trained on a large-scale production cluster trace. Under standard regression fit (R^2), both models improve modestly but monotonically across four feature-engineering iterations, reaching R^2 of about 0.042. However, on the metric that actually matters for scheduling - Top-1 ranking accuracy, whether the model scores the machine the production scheduler actually chose highest - both learned models are outperformed by a trivial single-feature heuristic (rank by free CPU: 74-84% vs. 65-66% for either model). We show this gap is best explained by objective mismatch: both models were trained with pointwise regression (MSE) rather than a ranking-specific objective, echoing a long-standing distinction in the learning-to-rank literature. This parallels prior evidence that RL-trained schedulers can outperform heuristics when the training objective is aligned with the deployment task, suggesting objective misalignment, not architecture, is the primary obstacle here. We further report an ablation of the occupancy reconstruction required to make offline trace data usable (naive features yield R^2 near 0), a controlled comparison isolating feature richness and data volume between the two model families, and a sensitivity analysis of inference latency and serving-container memory constraints. Code, data pipelines, and experiment scripts are released for reproducibility.
Read the full abstract

Distributed Computing

Probabilistic Performance Analysis of Parallel Signature Search Strategies in Multi-Level Tree Networks

Jingwei Li and Thomas G. Robertazzi

Hierarchical distributed search, locating a data pattern, or signature, across a tree-structured collection of files, underlies distributed index traversal, deep packet inspection and sequence alignment. A practitioner must decide how much parallelism to employ: scan each layer sequentially, fan out within subtrees, or launch the whole tree at once. Existing analyses answer this only partially: they characterize every node by the statistics of a signature-holding file and, for multi-signature files, need quantities revealed only at run time. We develop a probabilistic framework predicting the completion time of five search strategies, spanning sequential to full-tree parallelism, before any file is read. Node scan times are modeled as a mixture over signature presence, layer times as order statistics, and parallel subtree scans by extreme-value arguments; when signature counts are known, occupancy under capacity constraints is treated by generating functions. Each performance formula carries an exactness label: exact (or exact-in-regime), plug-in, asymptotic or bound, with each approximation quantified against Monte Carlo simulation and its regime identified. A multicore prototype reproduces the coarse separation between full-tree, layer- and subtree-level parallelism, but shows that synchronization overhead can erase the predicted separation between close strategies. The framework delivers a priori completion-time predictions with explicit accuracy regimes and negligible computational cost, the design example evaluated in under a millisecond; these timing models can support subsequent resource-cost optimization.
Read the full abstract

Distributed Computing · Cross-listed from cs.AR

FLINT: Efficiently Leveraging High Bandwidth Flash for Capacity-Scalable LLM Inference Acceleration

Geraldo F. Oliveira, Arash Tavakkol, Xiangyu Zhu et al. (+10 more)

LLM inference is increasingly constrained by accelerator memory capacity rather than compute throughput. This constraint is especially acute in single-accelerator and small-node inference systems, where limited on-package memory capacity restricts the size of deployable models. HBF is an emerging 3D-stacked NAND flash technology that provides multi-terabyte near-accelerator capacity, making it a promising capacity tier for storing LLM weights. However, existing HBF-based proposals face three adoption challenges: they (1) rely on coarse-grained static prefetching for LLM weights aiming to hide the microsecond-level read latency of the NAND flash device while maximizing HBF's read throughput, (2) expose NAND flash management tasks (e.g., refresh operations) to the accelerator-visible critical inference path, and (3) miss optimization opportunities to specialize and optimize the flash-management mechanisms to the workload behavior. Our goal is to design an efficient HBF substrate that integrates HBF as a memory-capacity tier alongside HBM while addressing these three challenges. To this end, we propose FLINT, a workload-driven HBF substrate for capacity-scalable LLM inference. FLINT introduces three mechanisms: (1) a hardware burst-buffer controller that dynamically coalesces and pipelines HBF reads aiming to utilize existing NAND flash buffers while sustaining high HBF bandwidth, (2) a phantom-plane refresh mechanism, which removes refresh from the critical inference path by moving refresh-related NAND flash operations outside the read foreground back via low-cost resource duplication, and (3) a read-only FTL, which replaces SSD-class support for arbitrary writes with a compact table that translates logical weight bursts to physical HBF locations.
Read the full abstract

All authors Geraldo F. Oliveira, Arash Tavakkol, Xiangyu Zhu, Ahmet Caner Yüzügüler, Vamanan Arulchelvan, Lukas Cavigelli, Renzo Andri, Mohammad Sadrosadati, Jia Xinglei, Onur Mutlu, Zhou Ke, Shai Bergman, Ji Zhang

Distributed Computing · Cross-listed from cs.AR

Hydra: Phase-Aware Workload Characterization of LLM Inference across Edge SoC Generations, Backends, and Quantization Levels

Amir Taherin, Sana Taghipour Anvari, Charles Amante et al. (+9 more)

Edge LLM deployment is shaped by more than model size and precision: inference backend, hardware platform, memory traffic, and power management all affect latency and efficiency. We present Hydra, a common-schema, phase-aware workload characterization framework for LLM inference on edge SoCs. Hydra instruments HuggingFace Transformers and llama.cpp with a shared per-prompt timing schema and fuses those records with hardware telemetry, enabling a multi-dimensional characterization of performance, system-resource utilization, and efficiency across prefill and decode phases. Using Hydra, we evaluate three consecutive edge System-on-Chip (SoC) generations (AGX Xavier, AGX Orin, and AGX Thor), 13 instruction-tuned LLMs from seven families, five execution formats, and consider input/output-length sensitivity. The resulting artifact contains roughly 107K per-prompt records and is publicly released with Hydra. Our analysis shows that aggregate latency alone hides key deployment effects: backend structure changes where latency is introduced, quantization reduces memory traffic and energy but does not predict power monotonically, and SoC generation changes how utilization and efficiency should be interpreted. By connecting phase-level timing with system-resource utilization and efficiency metrics, Hydra enables reproducible, phase-aware characterization of edge LLM inference. Hydra's source code and the collected per-prompt trace corpus are available open-source at: https://github.com/amirtaherin/hydra
Read the full abstract

All authors Amir Taherin, Sana Taghipour Anvari, Charles Amante, Yixiao Chen, Ruben Noroian, Zlatan Feric, Nicolas Bohm Agostini, Pu Zhao, José Cano, Bin Ren, Yanzhi Wang, David Kaeli

Distributed Computing · Cross-listed from cs.IT

Next-generation O-RAN Edge: Energy-aware Joint Placement and Migration of Cloud-Native Functions

Nguyen Phuc Tran, Brigitte Jaumard and Oscar Delgado

The transition toward Open Radio Access Networks (O-RANs) is reshaping how cellular infrastructure is deployed, managed, and optimized. This paper investigates the energy-aware joint placement and migration of cloud-native functions (CNFs) in an O-RAN edge cloud. We consider both a Single-CU-UP association model and a slice-aware Multi-CU-UP relaxation, in which distinct slice-flow groups of the same distributed unit (DU) may be assigned to different Centralized Unit User Plane (CU-UP) processing targets under one Centralized Unit Control Plane (CU-CP). For brevity, these scenarios are referred to as Single-CU and Multi-CU, respectively; Multi-CU never denotes multiple CU-CP associations. We formulate the problem as a Mixed-Integer Linear Program (MILP) that minimizes server, transmission, wake-up, and migration energy while satisfying server-resource capacities and one-way delay requirements over the F1 user-plane interface (F1-U) between each DU and its selected CU-UP in a fat-tree edge data center. To improve computational scalability, we also develop a deterministic k-means-based heuristic that approximates the MILP decisions without requiring repeated exact optimization. Over the evaluated 24-hour workload, the theoretical Multi-CU relaxation reduces modeled energy consumption by 5.7% relative to the Single-CU baseline. For the Multi-CU case, the proposed heuristic remains within approximately 9.7% of the proposed MILP, demonstrating a favorable trade-off between energy efficiency and computational tractability.
Read the full abstract

Distributed Computing · Cross-listed from cs.AR

Maia 200: A Software Defined Dataflow System for Large-scale AI Acceleration

Sherry Xu, Marco Heddes, Jackson Peng et al. (+14 more)

We introduce Maia 200, an advanced AI accelerator delivering high performance-10 145 Tflop/s FP4 and 5072 Tflop/s FP8 within a 750W TDP and 7 TB/s HBM bandwidth. Maia exemplifies a new class of Software Defined Locally Accessed Dataflow Architectures (SDLA), which explicitly program dataflow engines to orchestrate highly specialized memories and data movement engines. This approach shifts the focus from today's thread-centric to data-movement-centric architecture, improving efficiency and scalability. Our taxonomy of data management, inspired by Flynn's classification, highlights how SDLA addresses challenges in modern AI computing. Maia 200 achieves significant cost and energy savings while supporting massive parallelism for AI inference workloads, making it a compelling solution for next-generation high-performance computing systems.
Read the full abstract

All authors Sherry Xu, Marco Heddes, Jackson Peng, Tom Savell, Monica Tang, Prashant Ranjan, Jesse Benson, Ofer Dekel, Saurabh Dighe, Anupama Kurpad, Artour Levin, Matthew Mattina, George Petre, Cheng Tang, Yuan Yu, Li Zhang, Torsten Hoefler

Distributed Computing

Asynchronous Verifiable Information Dispersal with Low Space and Communication Complexity

Thomas Locher and Yvonne-Anne Pignolet

The primary goal of a distributed storage system is to ensure that clients can both write and read data in a reliable and consistent manner, even in the presence of failures. While existing asynchronous verifiable information dispersal (AVID) protocols achieve optimal space complexity for storage and communication complexity for data retrieval in a Byzantine setting, the crucial operations of data dispersal and node recovery have received less attention. We propose an efficient AVID protocol that simultaneously guarantees low complexities for dispersal, storage, retrieval, and recovery. At the core of the proposed protocol lies a novel mechanism to encode data in a two-dimensional matrix and a bespoke dispersal algorithm. The protocol maintains an optimal communication complexity for retrieval while substantially improving upon the state of the art for recovery. Additionally, we describe how the protocol can be modified in order to reduce the space complexity and the communication complexity for dispersal, at the expense of a higher communication complexity for retrieval and recovery. As the proposed protocols strike a balance across all considered metrics, they are suitable for a broad range of real-world use cases.
Read the full abstract

Distributed Computing

Scalable datacenter replication with mostly-synchronous consensus on hardware

Davide Rovelli, Philipp Berdesinski, Rodrigo Otoni et al. (+1 more)

Consistent replication of data among distributed processes – a task involving the well-known consensus problem – is notoriously expensive and hard to scale, affecting especially datacenter services with stringent performance requirements. To mitigate this problem, we introduce scalable replication in-hardware ( scarHW ): a network card design that improves throughput and latency of consistent replication even when increasing the number of replicas, whereas current systems operate at a small scale or with relaxed consistency guarantees. At the heart of scarHW is our novel POPUC consensus algorithm, implemented in an FPGA smartNIC to take full advantage of the "mostly synchronous" behavior of programmable network devices in the datacenter. Unlike widely-adopted "mostly asynchronous" coordination protocols such as Paxos or leaderless alternatives, POPUC implements a generalized variant of consensus dubbed collaborative consensus which allows for several simultaneous decisions, achieving great scalability without compromising availability. POPUC preserves safety guarantees in the presence of process crash-stop and message send/receive omission failures (capturing incidental asynchrony) and has been formally specified and verified in TLA+. Our FPGA prototype improves throughput and latency of widely-used services Redis and Zookeeper by up to two orders of magnitude compared to the state of the art. scarHW-based services also achieve zero downtime upon failure of a minority of replicas, offering a highly-robust, wire-speed, scalable replication system.
Read the full abstract

All authors Davide Rovelli, Philipp Berdesinski, Rodrigo Otoni, Patrick Eugster

Distributed Computing

SatDL: Jointly Optimizing Data Redistribution and Training for Satellite-Based Distributed Learning

Hao Wu, Kin Whye Chew, Yizhan Han et al. (+2 more)

Satellite-based distributed learning promises to train machine-learning models directly in orbit using massive, globally dispersed sensor data, thereby avoiding large-scale data downloads to ground servers. However, training convergence is significantly slowed by severe non-IID data, specifically label imbalance, as each satellite observes different geographic regions with distinct labels. This imbalance extends training duration and increases energy consumption for solar-powered satellites. Existing approaches either fully redistribute data to enforce IID conditions - accelerating convergence but incurring substantial communication delays - or avoid redistribution entirely by modifying local learning algorithms to mitigate the impact of label imbalance, which, however, still prolong training and increase energy use. Both extremes result in excessive total end-to-end learning time (data-transfer delay plus training time) and thus elevated onboard energy consumption. We present SatDL, a data-redistribution framework designed to minimize total end-to-end learning time. At its core, SatDL develops a Distributor-Critic framework that jointly models and optimizes data-transfer delay and training time. Evaluations through trace-driven simulations of a 1,584-satellite Starlink constellation and hardware emulations using NVIDIA Jetson and A100 GPUs across five datasets show SatDL reduces total end-to-end learning time by up to 18.6% and onboard energy consumption by 12.23-88.00%, while maintaining inference accuracy within a few percentage points of state-of-the-art baselines.
Read the full abstract

All authors Hao Wu, Kin Whye Chew, Yizhan Han, Han Li, Jingxian Wang

Distributed Computing · Cross-listed from cs.NE

On Scaling Coordinate-Based Neuroevolution: The Quadtree Bottleneck in ES-HyperNEAT

Romain Claret, Michael O'Neill, Paul Cotofrei et al. (+1 more)

ES-HyperNEAT evolves substrate topology through adaptive quadtree subdivision; to our knowledge, no implementation with full population-level GPU parallelization exists. We present JAX-ESHN, a JAX-based implementation targeting GPU parallelization with batched CPPN queries, and benchmark it against the CPU-based PUREPLES Baseline across five tasks: XOR, Parity-3, circle classification, sine regression, and CartPole. The core limitation is structural: each CPPN discovers a unique set of substrate positions, preventing population-level vectorization via vmap. On XOR, the CPU Baseline's runtime scales exponentially with depth while JAX-ESHN's construction cost on GPU (compilation plus first-generation evaluation) plateaus at deep substrates, so JAX-ESHN solves reliably where the Baseline rarely succeeds, with lower runtime variance. A CPU-vs-CPU multi-benchmark control reproduces the same scaling divergence across Boolean, continuous, and control task types, confirming it is a property of the substrate-discovery implementation, not of GPU hardware. An alternative data structure (Hierarchical Spatial Hash Grid) fails not because it precomputes positions but because it applies the variance test independently per position, discarding the quadtree's parent-gated filtering and with it the adaptive sparsity essential to ES-HyperNEAT. These findings define the structural constraints any substrate-discovery method must satisfy to scale coordinate-based neuroevolution; the companion EMR-HyperNEAT reformulation, which replaces adaptive subdivision with eager evaluation of a static multi-resolution grid, satisfies them and resolves the bottleneck this paper characterizes.
Read the full abstract

All authors Romain Claret, Michael O'Neill, Paul Cotofrei, Kilian Stoffel

Distributed Computing

Mixed-Precision SEM-Based CFD Simulations on GPUs: A Taylor-Green Vortex case

Yanxiang Chen, Manuel Münsch and Roman Iakymchuk

Mixed precision is a promising approach for reducing the computational cost and energy consumption of Computation Fluid Dynamics (CFD) simulations, but its effectiveness depends strongly on where precision is reduced within the full simulation pipeline. In this work, we study Taylor-Green vortex case using Neko, a matrix-free CFD solver based on the spectral element method (SEM). Profiling shows that the fluid time step is not dominated by Krylov convergence alone: the velocity and pressure solvers require only a small number of iterations per step, while a substantial fraction of runtime is spent in other SEM operators and solver components. Motivated by this structure, we propose a three-level hierarchical mixed-precision control model. Two groups of configurations are evaluated in environments bounded by 64-bit floating-point (fp64) and 32-bit floating-point (fp32) precision, respectively. The fp64-bounded group identifies accuracy sensitive components and shows that SEM-focused fp32 computation is a promising direction for future optimization. The fp32-bounded group provides the main practical benefit. For the high Reynolds number case studied, selected configurations reduce both time- and energy-to-solution by about 34% relative to the fp64 baseline, while improving robustness compared with global fp32. Targeted fp16 kernel overrides are also explored, showing potential for selected operations but increased sensitivity in gradient-based quantities such as enstrophy. Overall, these results indicate that mixed-precision for matrix-free SEM-based CFD should be treated as a simulation-level control problem rather than solely as a Krylov-solver optimization.
Read the full abstract

Distributed Computing

An HPC Approach to Accelerate Tensor Decompositions

Markus Hellgren, Erna Begovic Kovac, Hans O. Karlsson et al. (+1 more)

Quantum systems grow in complexity so rapidly that even modest models become difficult to simulate, creating a strong need for methods that can handle high-dimensional data, also known as tensors. In this work, we investigate a novel Jacobi-type tensor algorithm for tensor decomposition and develop a CUDA-based algorithm that supports tensors of arbitrary order on a single GPU. We test the implementation on NVIDIA H100 GPUs and show that the algorithm converges correctly for diagonalizable tensors up to nine dimensions, with runtime scaling in a predictable way as tensor order grows. Finally, our general algorithm outperforms the original MATLAB reference by more than two orders of magnitude.
Read the full abstract

All authors Markus Hellgren, Erna Begovic Kovac, Hans O. Karlsson, Roman Iakymchuk