Distributed Inference

Follow this topic Ask about this topic

Momentum

8 papers in the last four weeks, against 2 the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 55

Jun 2, 2026cs.DC

UltraEP: Unleash MoE Training and Inference on Rack-Scale Nodes with Near-Optimal Load Balancing

Large-scale expert parallelism (EP) is becoming pivotal for training and serving frontier MoE models, but it also amplifies device-level expert load imbalance into compute stragglers, token all-to-all bottlenecks, and activation-memory spikes. Existing balancers redistribute experts periodically based on historical load, which becomes unreliable for production deployments with non-stationary load patterns. We present UltraEP, the first exact-load, real-time balancer for large-EP MoE training and serving prefill on rack-scale nodes (RSNs). Leveraging the extended scale-up connectivity among dozens of GPUs within RSNs, UltraEP rebalances every microbatch and layer on critical paths, which requires nontrivial co-design of plan solving and expert replication communication to minimize exposed overhead. To this end, UltraEP eagerly reacts to post-gating load with an efficient quota-driven planner, and executes the resulting irregular expert-state transfers with RSN-native persistent tile streaming and relay-based fan-out mitigation. We evaluate UltraEP in a multi-RSN deployment of up to 256 GPUs, using cutting-edge MoE models from 106B to 671B parameters. Averaged across training and serving, UltraEP achieves 94.3% of the force-balanced ideal throughput, delivering 1.49

\times

improvement over no-balancing, while reducing the final inter-rank imbalance from 1.30

-

4.01 to 1.01

-

1.04.

Xinming Wei, Chao Jin, Tuo Dai +10

School of Computer Science, Peking University · Xiaohongshu Inc. · Shanghai AI Laboratory +1

Expert Parallelism Expert Load Balancing High-Performance Computing Mixture-of-Experts Inference Distributed Training Mixture-of-Experts Models+1

Jun 2, 2026cs.PF

NetKV: Network-Aware Decode Instance Selection for Disaggregated LLM Inference

Disaggregated LLM inference forces the KV cache to traverse the datacenter network before decoding begins, so transfer time enters directly into the Time to First Token (TTFT) budget. Current schedulers route on compute load and prefix-cache locality alone, ignoring the topological distance and dynamic congestion between prefill and decode instances. We close this gap with a thin operator-to-scheduler interface, the network cost oracle, and we prove that ignoring the network term renders cache-aware-only scheduling arbitrarily suboptimal as context length grows. NetKV, the O(|D|) per-request greedy that consumes this oracle, has tier rankings that are provably robust to stale telemetry. On a 64-GPU four-tier fat-tree simulator driven by Mooncake traces, NetKV reduces mean TTFT by up to 21.2% over round-robin and 17.6% over a tuned cache+load-aware scheduler, lifts SLO attainment by up to 20.1 percentage points, and keeps the Time Between Tokens overhead below 0.5 ms in every condition tested, with no changes to the transport, inference engine, or hardware.

Mubarak Adetunji Ojewale

Cloud Competency Centre, National College of Ireland

LLM Inference KV Caching LLM Inference Scheduling Cost-Aware Inference Distributed Inference

Jun 2, 2026cs.RO

CADET: A Modular Platform for Evaluating Distributed Cooperative Autonomy in Connected Autonomous Vehicles

Deep learning models are increasingly central to autonomous vehicle (AV) pipelines, yet their integration has traditionally followed a monolithic design where perception, planning, and control execute on a single onboard computer. This design overlooks the emerging paradigm of cooperative autonomy, where vehicles interact with roadside units (RSUs), edge servers, and cloud-hosted intelligence through vehicle-to-everything (V2X) connectivity. Cooperative perception and control improve safety and efficiency, but also introduce systems-level challenges: network latency, compute heterogeneity, and multi-tenant contention, all critically affect real-time decision-making. These challenges are further amplified by the increasing reliance on large foundation models, whose scale necessitates cloud deployment. We present CADET (Cooperative Autonomy through Distributed Experimentation Toolkit), a modular platform for systematic and reproducible evaluation of distributed cooperative autonomy systems under realistic deployment conditions. CADET decouples the AV stack into composable modules that can be flexibly deployed across vehicles, infrastructure, and edge/cloud tiers. The framework integrates state-of-the-art models, incorporates trace-driven network and workload emulation, and provides synchronized model-, system-, and task-level instrumentation. Through V2V and V2I experiments, we show that distributed deployment choices fundamentally shape safety, with V2V intent packets outperforming cloud-based perception and RSU-assisted perception sustaining safety until overloaded by concurrent requests. Although designed for AV pipelines, CADET also supports dataset-driven experimentation, enabling systems and ML researchers to benchmark distributed inference workloads independently of full vehicle simulation. CADET is open source, with code and demo available at https://nesl.github.io/cadet-web.

Pragya Sharma, Brian Wang, Mani Srivastava

University of California Los Angeles

Edge Computing Autonomous Driving V2X Communication Cooperative Perception Distributed Inference

May 31, 2026cs.DC

Move the Query, Not the Cache: Characterizing Cross-Instance Latent Attention Redistribution Across GPU Fabrics

Frontier LLMs increasingly decide what a query attends to with a sparse-attention indexer that picks a few KV-cache blocks per query: attention's unit is now a small, reusable chunk. Agentic workloads hammer it: many sub-agents query one large codebase, reusing the same blocks. When that corpus outgrows one GPU it is partitioned across instances, so a query and the blocks it selects often sit on different GPUs: answering it means attention across instances. The reflex of prior cross-instance KV systems is to move the cache: pull the selected blocks to the requester. Multi-head Latent Attention inverts the arithmetic, compressing each token's key and value into one narrow vector, so a routed query row is only ~1 KB, smaller than the chunk it attends; routing the query is then often cheaper than moving the cache. Which primitive wins, over which fabric and request shape, is uncharted, least of all on device-initiated RDMA that makes per-request cross-node transfers cheap. We characterize cross-instance MLA attention on a real multi-node H100 cluster, distilling two reusable artifacts: a topology-aware cost model (probe / transfer / compute / return / merge) and a closed-form route/fetch/local predicate, whose constants we measure on real IBGDA, where the model tracks batched round-trips to within ~7%. At decode it routes the query, trading the cost of moving the cache (a ~3 ms re-adaptation splice for a contiguous chunk, or a scattered gather under selection) for a tens-of-microsecond round trip, and picks the fabric by probe latency, not peak bandwidth. We instantiate the cost model and predicate for MLA, but neither is MLA-specific: they apply wherever compression or sparse selection shrinks attention to small chunks (DeepSeek-V3.2, V4, and GLM-5.1 today). Extending them to a new architecture requires measuring just two coefficients: the routed payload and fetch's move-the-cache cost.

Bole Ma, Jan Eitzinger, Harald Köstler +1

LLM Inference KV Caching Sparse Attention Attention Mechanisms Distributed Inference Multi-Head Latent Attention

May 31, 2026cs.LG

Beyond Task-Agnostic: Task-Aware Grouping for Communication-Efficient Multi-Task MoE Inference

Sparsely activated Mixture-of-Experts (MoE) models scale capacity via conditional computation, but distributed inference suffers from cross-GPU expert communication and routing-induced load imbalance. Existing placement methods reduce this cost by co-locating frequently co-activated experts; however, they derive a single deployment plan from globally aggregated routing traces, thereby averaging away the heterogeneous, task-specific co-activation patterns that actually drive communication in multi-task serving. We observe that expert co-activation is strongly task-conditioned: pairs tightly coupled in one task family are often uncorrelated in another, so effective deployment should group experts by task-aware co-activation rather than by a task-agnostic average. Based on this insight, we propose \emph{Task-Aware Coactivation Grouping} (TACG), a deployment-time framework that uses family-specific dispatch and co-activation traces to derive per-expert task-family preferences, reweights the co-activation graph so that intra-family locality dominates grouping, and assigns each expert to a primary GPU under exact capacity constraints. To keep the static placement robust under online workload skew, we further introduce \emph{Generic Expert Shared Replication} (GESR), a lightweight companion that identifies generic experts with consistently central co-activation profiles, replicates them across a small set of secondary GPUs, and applies locality- and load-aware selection at serving time. Experiments on three representative open-source MoE models demonstrate that our framework reduces the average communication cost by 31.39% over the baseline, while preserving an average Jain fairness index of 0.9975. This advantage persists even under severe distribution shifts in the inference data, consistently outperforming strong baselines.

Zhiyao Xu, Aoxue Liu, Zhanjie Ding +3

Tsinghua Shenzhen International Graduate School, Shenzhen, China · Pengcheng Laboratory, Shenzhen, China

Mixture-of-Experts Inference Mixture-of-Experts Models Distributed Inference

May 31, 2026cs.DC

Lodestar: An Online-Learning LLM Inference Router

Efficiently serving large language model (LLM) inference tasks is crucial both for user-perceived latency such as time-to-first-token (TTFT) and for GPU utilization. However, LLM request routing, that is, assigning each inference request to a GPU instance, is particularly challenging: execution is highly input-dependent; batching and KV-cache reuse create strong cross-request coupling; and latency responds nonlinearly to context length, model/engine settings, and heterogeneous accelerators. As a result, simple traditional load balancing algorithms, and even heuristics tailored for LLM inference, fail to achieve good performance. We present Lodestar, a novel learning-based request routing system for distributed GPU clusters. Lodestar continuously collects a snapshot of the cluster at per-request level, including real-time instance state, request characteristics, and observed performance, and trains an online reward predictor that it uses to route inference requests to the instance that will maximize given reward (e.g., minimizing TTFT). Lodestar is cloud-native and works seamlessly with existing serving stacks (vLLM). With continuous online adaptation to changing workloads and infrastructure conditions, Lodestar achieves 1.41x lower average TTFT and 1.47x lower P99 TTFT on average (up to 2.15x/1.86x on homogeneous and 4.38x/4.42x on heterogeneous clusters) compared to a state-of-the-art prefix cache and load-aware heuristic, and learns these efficient routing strategies within about 5 minutes, based on experiments in a public cloud GPU cluster.

Gangmuk Lim, Wanyu Zhao, Brighten Godfrey +3

UIUC · Bytedance · University of Edinburgh

Inference-Time Optimization LLM Inference Scheduling LLM Inference Serving LLM Routing Distributed Inference Online Learning

May 27, 2026cs.LG

How Far Can Disaggregation Go? A Design-Space Exploration of Attention-FFN Disaggregation for Efficient MoE LLM Serving

Modern large language model (LLM) inference has progressively disaggregated to keep pace with growing model sizes and tight TTFT and TPOT service-level objectives: from chunked-prefill aggregation, to prefill-decode (P/D) disaggregation, and most recently to operator-level Attention-FFN Disaggregation (AFD). This trend is especially important for mixture-of-experts (MoE) models, where memory-bound attention, compute-intensive expert FFNs, and MoE dispatch/combine communication create distinct resource demands. AFD further exposes this heterogeneity by placing attention and MoE-FFN execution on separate GPU groups. Each level of disaggregation deepens the scheduling design space across workload characteristics, resource allocation, and interconnect topology, raising the central question: when does each level actually pay off? We systematically characterize this trade-off for MoE inference across realistic workloads spanning input/output sequence lengths, prefix-KV reuse, and per-user latency constraints. Using chunked-prefill and P/D disaggregation as baselines, we study the benefits and limits of AFD at scale through a framework that fuses on-device kernel measurements with high-fidelity network simulation. Under strict TTFT/TPOT SLOs, AFD sustains around 4k tokens/s of system throughput on DeepSeek-V3.2 across chat, coding, and agentic-coding workloads, where non-AFD deployments are infeasible. We distill concrete takeaways for jointly optimizing throughput and interactivity, including how to partition attention and FFN across GPUs as a function of workload and model architecture, providing design principles for current rack- and cluster-scale deployments as well as future disaggregated AI infrastructure.

Hanjiang Wu, Abhimanyu Rajeshkumar Bambhaniya, Sarbartha Banerjee +9

1Georgia Institute of Technology · 5Infravana · 2Intel +2

LLM Serving LLM Inference Disaggregated LLM Serving Mixture-of-Experts Inference Distributed Inference

May 22, 2026cs.LG

Unextractable Protocol Models: Collaborative Training and Inference without Weight Materialization

We consider a decentralized setup in which the participants collaboratively train and serve a large neural network, and where each participant only processes a subset of the model. In this setup, we explore the possibility of unmaterializable weights, where a full weight set is never available to any one participant. We introduce Unextractable Protocol Models (UPMs): a training and inference framework that leverages the sharded model setup to ensure model shards (i.e., subsets) held by participants are incompatible at different time steps. UPMs periodically inject time-varying, random, invertible transforms at participant boundaries; preserving the overall network function yet rendering cross-time assemblies incoherent. On Qwen-2.5-0.5B and Llama-3.2-1B, 10,000 transforms leave FP32 perplexity unchanged (

Δ

PPL

< 0.01

; Jensen-Shannon drift

< 4 \times 10^{-5}

), and we show how to control growth for lower precision datatypes. Applying a transform every 30s adds 3% latency, 0.1% bandwidth, and 10% GPU-memory overhead at inference, while training overhead falls to 1.6% time and

< 1

% memory. We consider several attacks, showing that the requirements of direct attacks are impractical and easy to defend against, and that gradient-based fine-tuning of stitched partitions consumes

\geq 60

% of the tokens required to train from scratch. By enabling models to be collaboratively trained yet not extracted, UPMs make it practical to embed programmatic incentive mechanisms in community-driven decentralized training.

Alexander Long, Chamin Hewa Koneputugodage, Thalaiyasingam Ajanthan +5

Pluralis Research

LLM Inference Decentralized Learning Privacy-Preserving ML Distributed Inference

May 20, 2026cs.DC

Frontier: Towards Comprehensive and Accurate LLM Inference Simulation

Modern LLM serving is no longer homogeneous or monolithic. Production systems now combine disaggregated execution, complex parallelism, runtime optimizations, and stateful workloads such as reasoning, agents, and RL rollouts. Simulation is attractive for exploring this growing design space, yet existing simulators lack the architectural completeness and decision-grade fidelity it demands. Their monolithic-replica abstractions are ill-suited to disaggregated serving, while average-case analytical proxies can distort SLA predictions and even reverse optimization conclusions. We present Frontier, a discrete-event simulator for modern LLM inference serving. Frontier features a disaggregated abstraction. It captures the structure and dynamics of modern serving systems by modeling co-location, Prefill-Decode Disaggregation (PDD), and Attention-FFN Disaggregation (AFD) with role-specific cluster workers, incorporating key runtime optimizations (e.g., CUDA Graphs, speculative decoding) within the scheduler-batch-engine loop, and supporting stateful requests for emerging workloads. It further provides accurate and generalizable predictions of computation, communication, and memory costs across diverse serving scenarios with complex workload compositions. On 16-H800 GPU testbed, Frontier achieves an average throughput error below 4%. Compared with state-of-the-art simulators, it reduces end-to-end latency error from 44.9% to 6.4% under co-location and from 51.7% to 2.6% under disaggregation. It scales to over 1K GPUs on commodity CPUs and enables new use cases such as SLA-dependent Pareto frontier exploration, heterogeneous disaggregated allocation, agentic reasoning scheduling validation, and RL post-training reconfiguration. We release Frontier at https://github.com/NetX-lab/Frontier.

Yicheng Feng, Xin Tan, Yangtao Deng +3

The Chinese University of Hong Kong · Anuttacon · StepFun

LLM Serving Disaggregated LLM Serving LLM Inference Scheduling LLM Inference Serving Distributed Inference

May 18, 2026cs.LG

Federated Martingale Posterior Samping

Federated Bayesian neural networks require fixing a prior on the model parameters together with a likelihood. Eliciting meaningful priors on the weight space of modern overparameterized models is notoriously difficult, and misspecification of either component can severely degrade accuracy and calibration. Motivated by the rapid progress of predictive models such as large language models, the martingale posterior, also known as predictive Bayes, replaces the prior--likelihood pair with a predictive distribution and recovers parameter uncertainty by repeatedly drawing predictive samples and refitting the model. A direct federated implementation, however, would require clients to share the local data sets. This letter proposes {federated martingale posterior} (FMP) sampling, a one-shot embarrassingly parallel protocol in which each client uploads a small set of trainable data embeddings and the server runs the predictive sampler centrally. Experiments on MNIST, CIFAR-10, and CIFAR-100 show that FMP closely matches the centralized counterpart and significantly improves calibration over consensus-style baselines.

Boning Zhang, Matteo Zecchin, Mingzhao Guo +2

School of Computing, University of Glasgow, Glasgow G12 8QQ, U.K. · Communication Systems Department, EURECOM, 06904 Sophia Antipolis, France · Institute for Intelligent Networked Systems, Northeastern University London, London E1 8PH, U.K.

Bayesian Neural Networks Posterior Sampling Distributed Inference Federated Learning

May 13, 2026cs.LG

Byzantine-Robust Distributed Sparse Learning Revisited

We revisit Byzantine robust distributed estimation for high-dimensional sparse linear models. By combining local

\ell_1

-regularized robust estimation with robust aggregation at the server, the framework applies to pseudo-Huber regression, quantile regression, and sparse SVM. We show that the resulting estimators yield non-asymptotic guarantees and attain near-optimal statistical rates under mild conditions, while remaining communication-efficient. Simulations confirm strong robustness in estimation, support recovery and classification accuracy under various Byzantine attacks.

Yuxuan Wang, Lixin Zhang, Kangqiang Li

School of Mathematical Sciences, Zhejiang University, Hangzhou, 310027, Zhejiang, China · School of Statistics and Mathematics, Zhejiang Gongshang University, Hangzhou, 310018, Zhejiang, China · Information Center, Hubei Provincial Tobacco Monopoly Administration, Wuhan, 430030, Hubei, China

Robust Regression Distributed Optimization Linear Regression Distributed Training Distributed Inference Robust Estimation

May 11, 2026cs.DC

ChunkFlow: Communication-Aware Chunked Prefetching for Layerwise Offloading in Distributed Diffusion Transformer Inference

Layerwise offloading reduces the GPU memory footprint of large diffusion transformer (DiT) inference by prefetching upcoming layers from host memory, but its effectiveness hinges on hiding prefetch latency behind per-layer computation. This assumption breaks down when the per-GPU compute workload is small. Moreover, on PCIe-only nodes, prefetch and inter-GPU collective communications such as all-reduce and all-to-all contend on the shared PCIe path, exposing prefetch latency even when compute would otherwise hide it. We revisit layerwise offloading as a co-scheduling problem between prefetch and communication, guided by a first-order analytical model that predicts when prefetch can be hidden by computation. Building on this model, we design ChunkFlow, a communication-aware, chunk-granular offloading runtime that adaptively yields to collective communication and smoothly trades GPU memory for prefetch volume. On three representative diffusion transformers running on two H100 GPUs over PCIe with Ulysses sequence parallelism, ChunkFlow delivers up to 1.28x step-time speedup over SGLang's existing layerwise offloading, reduces peak GPU memory by up to 49% over the no-offload baseline at near-identical step time once the workload is large enough, and exposes a tunable memory-latency tradeoff that recovers near-zero step-time overhead in the small-workload regime.

Han Meng, Danny Willow Liu, Dong Li

University of California, Merced · University of Chicago · University of California, Merced and Yotta Labs +1

Diffusion Transformer GPU Acceleration Memory-Efficient Inference Distributed Inference Diffusion Model Inference Acceleration

May 11, 2026cs.DC

ShardTensor: Domain Parallelism for Scientific Machine Learning

Scientific Machine Learning (SciML) faces unique challenges for extreme-resolution data, with mitigations that often fail to scale or degrade the accuracy of trained models. While some specialized methods have achieved remarkable results in training models or performing inference on massive spatial datasets with bespoke techniques, there is no generalized framework for parallelization over input data below batch size one per device. In this work we introduce ShardTensor: a novel paradigm of domain parallelism that enables flexible scaling of input data to arbitrary sizes. By decoupling the spatial dimensionality of input data from hardware constraints, ShardTensor enables scientific machine learning workloads to reach new levels of high fidelity training and inference. We demonstrate both strong and weak scaling of workloads during training and inference, showing improved latency with strong scaling and demonstrating the capacity to process higher data sizes with weak scaling. Additionally, we demonstrate multiple dimensions of parallelization, removing barriers to SciML on extreme-scale inputs.

Corey Adams, Peter Harrington, Akshay Subramaniam +4

NVIDIA · Santa Clara, CA, USA

Scientific ML Distributed Training Distributed Inference

May 10, 2026cs.DC

PoHAR: Understanding Hyperlocal Human Activities with Pollution Sensor Networks

Low-cost air quality sensors are becoming ubiquitous in our daily lives as public awareness of air pollution continues to grow, and people take measures to monitor and improve the air they breathe indoors. Besides the standard operation of these sensors, fluctuations in environmental parameters can be leveraged to understand human behavior and activities in indoor spaces. Unlike traditional audio-visual, Radio Frequency, and inertial sensors, air quality sensors are easily scalable to a household, are privacy-preserving, and more economical. Such distributed sensor networks must jointly make decisions to monitor indoor occupants for downstream smart home and healthcare applications. However, due to low processing power, memory, and energy, they often struggle to maintain distributed data consensus and identify activity-affected sensor groups for accurate on-device inference. In this paper, we propose PoHAR framework that implements: (i) a conflict-free replicated data primitive for data sharing, (ii) a hierarchical clustering for ESP32 to detect activity-affected sensor groups with a self-supervised distance metric, and (iii) a leader-based group inference with off-the-shelf ML classifiers, enabling the sensor network to collaboratively detect hyperlocal indoor activities. Our extensive experiments demonstrated on-device activity detection, achieving 97.41% accuracy for indoor activity and 99.68% for cooking activity, using off-the-shelf ML models with latency below 34 microseconds.

Prasenjit Karmakar, Karthik Reddy, Sandip Chakraborty

Department of Computer Science and Engineering, Indian Institute of Technology Kharagpur, India.

Human Activity Recognition Internet of Things Distributed Inference

May 7, 2026cs.DC

Towards Distributed Inference of LLMs on a P2P Network

Prefix caching can reduce LLM inference latency by reusing KV caches across requests with shared prompts, but cluster-scale reuse is challenging because caches are partitioned across nodes. We propose a decentralized, prefix-cache-aware routing scheme for peer-to-peer LLM serving. Each node maintains a local radix tree of its own cached prefixes and asynchronously refreshed estimates of peer caches using periodic anti-entropy. Requests are routed to the node with the longest estimated prefix match, without centralized coordination or KV-cache transfer. Stale metadata only causes cache misses, not incorrect outputs, making weak consistency sufficient for correctness. Evaluation on simulated MMLU workloads show that decentralized routing improves latency under low communication delay and skewed prefix distributions, while high network latency and affinity-induced hotspots limit its benefits.

Shabari S Nair, Krishanu Saini

Department of Computer Science, The University of Texas at Austin.

LLM Inference KV Caching LLM Inference Serving Prefix Caching LLM Routing Distributed Inference

May 7, 2026cs.LG

Federation of Experts: Communication Efficient Distributed Inference for Large Language Models

Mixture of experts has emerged as the primary mechanism for making Large Language Models (LLMs) computationally efficient. However, in distributed settings, communicating token embeddings between experts is a significant bottleneck. We present the novel Federation of Experts (FoE) architecture. FoE restructures the MoE block of a transformer layer into multiple MoE clusters. Each cluster is responsible for only one of the KV heads and expert parallelism is applied between those experts. Between clusters, a sum synchronizes the post-attention residuals, which then drives routing and dispatch for the next MoE block. In a single-node setting, FoE completely eliminates all-to-all communication as all experts within a group are contained on the same GPU. In multi-node settings, FoE confines all-to-all communication to the intra-node fabric, thus significantly reducing communication overhead. An implementation of FoE finds that on LongBench, FoE significantly improves inference throughput and latency in both single-node and multi-node settings, reducing the end-to-end forward-pass latency by up to 5.2x, TTFT by 3.62x, and TBT by 1.95x. It does so while achieving comparable generation quality to a mixture of experts model of the same size and training configuration.

Muhammad Shahir Abdurrahman, Chun Deng, Azalia Mirhoseini +1

Stanford University

Expert Parallelism Mixture-of-Experts Inference LLM Inference Acceleration Distributed Inference

May 7, 2026cs.LG

Enabling Federated Inference via Unsupervised Consensus Embedding

Cooperative inference across independently deployed machine learning models is increasingly desirable in distributed environments, as there is a growing need to leverage multiple models while keeping their data and model parameters private. However, existing cooperative frameworks typically rely on sharing input data, model parameters, or a common encoder, which limits their applicability in privacy-sensitive or cross-organizational settings. To address this challenge, we propose Consensus Embedding-based Federated Inference (CE-FI), a framework that enables pretrained models to cooperate at inference time without sharing model parameters or raw inputs and without assuming a common encoder. CE-FI introduces two components: a Consensus Embedding (CE) layer that maps heterogeneous intermediate representations into a common embedding space, and a Cooperative Output (CO) layer that produces predictions from these embeddings. Both layers are trained using shared unlabeled data only, so the cooperative stage does not require additional labeled data. Experiments on image classification benchmarks -- CIFAR-10 and CIFAR-100 -- under diverse non-IID conditions show that CE-FI consistently outperforms solo inference and performs comparably to conventional methods that require stronger sharing assumptions. Additional evaluations on text and time-series tasks indicate applicability beyond image classification, although performance depends on the ensemble strategy. Further analysis identifies representation alignment as the primary bottleneck.

Yui Hashimoto, Takayuki Nishio, Yuichi Kitagawa +1

School of Engineering, Institute of Science Tokyo, Tokyo 152-8550, Japan · Research & Development Group, Hitachi Ltd., Tokyo 185-8601, Japan

Ensemble Learning Privacy-Preserving ML Distributed Inference Representation Alignment Federated Learning

May 3, 2026cs.LG

MoE-Prefill: Zero Redundancy Overheads in MoE Prefill Serving

Production LLM workloads increasingly serve discriminative tasks, such as classification, recommendation, and verification, whose answers are read from the logits of a single prefill pass with no autoregressive decoding. Serving these prefill-only workloads on mixture-of-experts (MoE) models is bottlenecked not by compute but by the distributed execution required to fit the model: existing parallel strategies (tensor, expert, and pipeline parallelism) trade memory pressure for redundant computation, communication, and synchronization, severely degrading MoE prefill serving efficiency. We observe that these overheads stem from coupling expert placement with synchronous activation routing -- a design inherited from the decoding era. The long, compute-bound forward passes of large-batch prefill open a per-layer window wide enough to stream expert weights in the background, replacing per-layer activation AllToAll with asynchronous weight AllGather fully overlapped with computation. We propose MoE-Prefill, a prefill-only serving system whose backend, AsyncEP (Asynchronous Expert Parallelism), gathers experts by weight rather than routing them by activation, and whose frontend co-enforces a physically-derived saturation threshold through prefix-aware routing and true-FLOPs load tracking. On Qwen3-235B-A22B across four hardware/precision configurations, MoE-Prefill delivers 1.35-1.37x throughput over the strongest distributed baseline on real-world workloads and up to 1.59x on long-context synthetic workloads, sustaining 29.8-36.2% per-GPU model FLOPs utilization.

Zhaoyuan Su, Olatunji Ruwase, Karthik Ganesan +5

University of Virginia · 2Snowflake AI Research · 3Harvard University

Expert Parallelism LLM Serving Mixture-of-Experts Inference Distributed Inference

May 1, 2026cs.DC

SpaceMoE: Realizing Distributed Mixture-of-Experts Inference over Space Networks

Leveraging continuous solar energy harvesting at high efficiency, space data centers are envisioned as a promising platform for executing energy-intensive large language models (LLMs). Recognizing this advantage, space and AI conglomerates (e.g., SpaceX, Google) are actively investing in this vision. One key challenge, however, is the efficient distributed deployment of a large-scale LLM in a satellite network due to the limited onboard computing and communication resources. This gives rise to a placement problem that involves partitioning and mapping model components to satellites such that the fundamentally different model architecture and network topology can be reconciled to ensure low-latency token generation. To address this problem, we present the Space Network of Mixture-of-Experts (SpaceMoE) framework targeting the distributed execution of a popular mixture-of-experts (MoE) model in space. The proposed placement strategies are two-level: (1) layer placement, which assigns MoE layers to satellite subnets; and (2) intra-layer expert placement, which assigns individual experts to satellites associated with the same layer/subnet. For layer placement, we exploit the ring-like communication pattern of autoregressive inference to partition the satellite constellation along the orbiting direction into subnets arranged on a ring, each hosting one MoE layer. Based on this architecture, we formulate and solve an optimization problem for intra-layer expert placement to map experts with heterogeneous activation probabilities onto satellites. The derived strategy reveals an intuitive principle: a frequently activated expert should be mapped to a satellite on a routing path with low expected latency. Experiments over a thousand-satellite constellation show that SpaceMoE achieves at least a threefold latency reduction compared with conventional random and ablation-based placement strategies.

Zhanwei Wang, Huiling Yang, Min Sheng +2

Department of Electrical and Computer Engineering, The University of Hong Kong (HKU), Hong Kong SAR, China · State Key Laboratory of Integrated Service Networks, Institute of Information Science, Xidian University (XDU), Xi’an, Shaanxi, China · Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology (HKUST), Hong Kong SAR, China

Mixture-of-Experts Language Models Mixture-of-Experts Inference LEO Satellite Networks Distributed Inference

Apr 30, 2026cs.DC

AI Inference as Relocatable Electricity Demand: A Latency-Constrained Energy-Geography Framework

AI inference is becoming a persistent and geographically distributed source of electricity demand. Unlike many traditional electrical loads, inference workloads can sometimes be executed away from the user-facing service location, provided that latency, state locality, capacity, and regulatory constraints remain acceptable. This paper studies when such digital relocation of computation can be interpreted as latency-constrained relocation of electricity demand. We develop an energy-geography framework for geo-distributed AI inference. The framework models a three-layer architecture of clients, service nodes, and compute nodes, and formulates inference placement as a constrained optimization problem over electricity prices, marginal carbon intensity, power usage effectiveness, compute capacity, network latency, and migration frictions. The key object is the energy-latency frontier: the marginal cost and carbon benefit unlocked by relaxing inference latency budgets. The paper makes four contributions. First, it distinguishes physical electricity transmission from digital relocation of electricity-consuming computation. Second, it formulates a geo-distributed inference placement model with feasibility masks and migration frictions. Third, it introduces operational metrics, including relocatable inference demand, energy return on latency, carbon return on latency, and a relocation break-even condition. Fourth, it provides a transparent stylized simulation over representative global compute regions to show how heterogeneous latency tolerance separates workloads into local, regional, and energy-oriented execution layers. The results show that latency relaxation expands feasible geography, while migration frictions, egress costs, state locality, legal constraints, and capacity limits can sharply reduce realized benefits.

Xubin Luo, Cheng Yang

School of Computer Science and Artificial Intelligence, Southwestern University of Finance and Economics, Chengdu, China

Energy-Efficient ML Cost-Aware Inference Distributed Inference

Apr 27, 2026cs.DC

Spark Policy Toolkit: Semantic Contracts and Scalable Execution for Policy Learning in Spark

Custom policy-learning pipelines in Spark fail for two coupled systems reasons: rowwise Python execution makes inference impractical, and driver-side candidate materialization makes split search fragile at feature scale. We present Spark Policy Toolkit, a semantics-governed systems toolkit for scalable policy learning in Spark. The toolkit provides two Spark-native primitives: partition-initialized vectorized inference through mapInPandas and mapInArrow, and collect-less split search that scores candidates on executors. Both primitives are governed by one fixed-input semantic contract: the same rows, feature order, treatment vocabulary, preprocessing manifest, and split boundaries must preserve per-row score vectors, best-split decisions, and end-to-end learned policy outputs. The evaluation combines practical baseline ladders, backend parity checks, measured split-search scale results, synthetic and Hillstrom end-to-end policy preservation, missingness stress, partition and order perturbation tests, quantile-boundary sensitivity, and a concrete adversarial failure catalog. On a 40-worker Databricks cluster, mapInArrow reaches 4.72M rows/s at 10M matched rows and 7.23M rows/s at 50M rows, while collect-less split search remains valid from F = 10 through F = 1000 with 124000 candidate rows, where the driver-collect baseline is intentionally skipped. Across 24 backend-ablation settings, mapInArrow wins 18 while mapInPandas wins 6, so the paper treats backend choice as workload-dependent rather than universal. Once the fixed-input lock is enforced, all six tested repartition/coalesce/shuffle perturbations preserve identical signatures; before lock, all six drift. The central result is not speed alone: throughput and collect-less execution are the mechanisms that let policy semantics survive at Spark scale.

Zeyu Bai

Efficient Inference Policy Learning Distributed Inference

Apr 25, 2026cs.LG

Scaling Multi-Node Mixture-of-Experts Inference Using Expert Activation Patterns

Most recent state-of-the-art (SOTA) large language models (LLMs) use Mixture-of-Experts (MoE) architectures to scale model capacity without proportional per-token compute, enabling higher-quality outputs at manageable serving costs. However, MoE inference at scale is fundamentally bottlenecked by expert load imbalance and inefficient token routing, especially in multi-node deployments where tokens are not guaranteed to be routed to local experts, resulting in significant inter-node all-to-all communication overhead. To systematically characterize these challenges, we profile SOTA open-source MoE models, including Llama 4 Maverick, DeepSeek V3-671B, and Qwen3-230B-A22B, on various datasets and collected over 100k real expert activation traces. Upon studying the expert activation patterns, we uncover various persistent properties across all the frontier MoE models: variable expert load imbalance, domain-specific expert activation where expert popularity shifts across task families (code, math, chat, general), and a strong correlation between prefill and decode expert activations. Motivated by these findings, we propose workload-aware micro-batch grouping and an expert placement strategy to maximize token locality to the destination expert, thereby reducing inter-node communication. Across models and datasets, these optimizations help reduce all2all communication data up to 20, resulting in lower MoE decode latency and better accelerator utilization.

Abhimanyu Bambhaniya, Geonhwa Jeong, Jason Park +6

1Georgia Institute of Technology, USA · 2Meta Platforms, Inc.

Expert Load Balancing Mixture-of-Experts Inference LLM Routing Mixture-of-Experts Models Distributed Inference

Apr 23, 2026cs.AI

Time, Causality, and Observability Failures in Distributed AI Inference Systems

Distributed AI inference pipelines rely heavily on timestamp-based observability to understand system behavior. This work demonstrates that even small clock skew between nodes can cause observability to become causally incorrect while the system itself remains functionally correct and performant. We present controlled experiments on a multi-node AI inference pipeline, where clock skew is introduced at a single stage. Results show that no violations are observed under synchronized conditions and up to 3 ms skew, while clear causality violations emerge by 5 ms. Despite this, system throughput and output correctness remain largely unaffected. We further observe that violation behavior is not strictly static. In longer runs, negative span rates may stabilize or decrease over time, indicating that effective skew evolves due to relative clock drift between nodes. Experiments were conducted using Kafka and ZeroMQ transports, with consistent results across both. Aeron is under active exploration but is not yet included in the completed validation set. These findings suggest that observability correctness depends not only on system functionality but also on precise time alignment, and that timing must be treated as a first-class concern in distributed AI systems.

Ankur Sharma, Deep Shah, David Lariviere +1

Unified Intelligent Infrastructure workstream at the Open Compute Project (OCP)

Temporal Consistency Distributed Inference

Apr 16, 2026cs.RO

NeuroMesh: A Unified Neural Inference Framework for Decentralized Multi-Robot Collaboration

Deploying learned multi-robot models on heterogeneous robots remains challenging due to hardware heterogeneity, communication constraints, and the lack of a unified execution stack. This paper presents NeuroMesh, a multi-domain, cross-platform, and modular decentralized neural inference framework that standardizes observation encoding, message passing, aggregation, and task decoding in a unified pipeline. NeuroMesh combines a dual-aggregation paradigm for reduction- and broadcast-based information fusion with a parallelized architecture that decouples cycle time from end-to-end latency. Our high-performance C++ implementation leverages Zenoh for inter-robot communication and supports hybrid GPU/CPU inference. We validate NeuroMesh on a heterogeneous team of aerial and ground robots across collaborative perception, decentralized control, and task assignment, demonstrating robust operation across diverse task structures and payload sizes. We plan to release NeuroMesh as an open-source framework to the community.

Yang Zhou, Yash Shetye, Long Quang +8

New York University, New York, USA · U.S. Army Combat Capabilities Development Command, Army Research Laboratory, Adelphi, MD 20783, USA · Vanderbilt University, Tennessee, USA +1

Efficient Neural Network Inference Multi-Robot Systems Multi-Robot Coordination Multi-Agent Communication Distributed Inference

Apr 7, 2025cs.DC

Prima.cpp: Fast 30-70B LLM Inference on Heterogeneous and Low-Resource Home Clusters

On-device inference offers privacy, offline use, and instant response, but consumer hardware restricts large language models (LLMs) to low throughput and capability. To overcome this challenge, we present prima.cpp, a distributed on-device inference system that runs 30-70B LLMs on consumer home clusters with mixed CPUs/GPUs, insufficient RAM/VRAM, slow disks, Wi-Fi links, and heterogeneous OSs. We introduce pipelined-ring parallelism (PRP) to overlap disk I/O with compute and communication, and address the prefetch-release conflict in mmap-based offloading. We further propose Halda, a heterogeneity-aware scheduler that co-optimizes per-device CPU/GPU workloads and device selection under RAM/VRAM constraints. On four consumer home devices, a 70B model reaches 674 ms/token TPOT with <6% memory pressure, and a 32B model with speculative decoding achieves 26 tokens/s. Compared with llama.cpp, exo, and dllama, our proposed prima.cpp achieves 5-17 times lower TPOT, supports fine-grained model sizes from 8B to 70B, ensures broader cross-OS and quantization compatibility, and remains OOM-free, while also being Wi-Fi tolerant, privacy-preserving, and hardware-independent. The code is available at https://github.com/OpenCPIL/prima.cpp.

Zonghang Li, Tao Li, Wenjiao Feng +8

MBZUAI · UESTC · City University of Hong Kong +1

On-Device Language Model Inference LLM Inference LLM Inference Scheduling LLM Inference Acceleration Pipeline Parallelism Distributed Inference