LLM Inference Efficiency

LLM: Large Language Model

Ask about this topic

Overview

The field studies how to serve large language models faster and more cheaply while preserving answer quality, especially for long contexts, reasoning, and multi-step agents. Systems approaches improve batching, scheduling, speculative decoding, and key-value cache storage through methods such as quantization, eviction, and reuse. Model-side approaches adapt reasoning budgets, route queries between models, and compress prompts or context to avoid unnecessary computation. A central challenge is that gains depend on the model, workload, hardware, and latency or energy goals, so evaluations increasingly compare quality, cost, latency, and energy together on tasks such as LongBench, GSM8K, and MATH.

Latest papers

CardsList

352 papers40 papers in the last four weeks, down 30% on the four weeks before. 0.2% of all new papers.

Mar 31, 2026

The Energy Footprint of LLM-Based Environmental Analysis: LLMs and Domain Products

As large language models (LLMs) are increasingly used in domain-specific applications, including climate change and environmental research, understanding their energy footprint has become an important concern. The growing adoption of retrieval-augmented (RAG) systems for climate-domain specific analysis raises a key question: how does the energy consumption of domain-specific RAG workflows compare with that of direct generic LLM usage? Prior research has focused on standalone model calls or coarse token-based estimates, while leaving the energy implications of deployed application workflows insufficiently understood. In this paper, we assess the inference-time energy consumption of two LLM-based climate analysis chatbots (ChatNetZero and ChatNDC) compared to the generic GPT-4o-mini model. We estimate energy use under actual user queries by decomposing each workflow into retrieval, generation, and hallucination-checking components. We also test across different times of day and geographic access locations. Our results show that the energy consumption of domain-specific RAG systems depends strongly on their design. More agentic pipelines substantially increase inference-time energy use, particularly when used for additional accuracy or verification checks, although they may not yield proportional gains in response quality. While more research is needed to further test these initial findings more robustly across models, environments and prompting structures, this study provides a new understanding on how the design of domain-specific LLM products affects both the energy footprint and quality of output.

Alicia Bao, Jiamian He, Angel Hsu +3

James · Department of Computer Science, University of North Carolina at Chapel Hill, Chapel Hill, NC 27516 · Arboretica, Rotterdam, The Netherlands +2

LLM Inference EfficiencyLLM InferenceRetrieval-Augmented Generation
AskPDF
Mar 30, 2026

Privacy Guard & Token Parsimony by Prompt and Context Handling and LLM Routing

The large-scale adoption of Large Language Models (LLMs) forces a trade-off between operational cost (OpEx) and data privacy. Current routing frameworks reduce costs but ignore prompt sensitivity, exposing users and institutions to leakage risks towards third-party cloud providers. We formalise the "Inseparability Paradigm": advanced context management intrinsically coincides with privacy management. We propose a local "Privacy Guard" -- a holistic contextual observer powered by an on-premise Small Language Model (SLM) -- that performs abstractive summarisation and Automatic Prompt Optimisation (APO) to decompose prompts into focused sub-tasks, re-routing high-risk queries to Zero-Trust or NDA-covered models. This dual mechanism simultaneously eliminates sensitive inference vectors (Zero Leakage) and reduces cloud token payloads (OpEx Reduction). A LIFO-based context compacting mechanism further bounds working memory, limiting the emergent leakage surface. We validate the framework through a 2x2 benchmark (Lazy vs. Expert users; Personal vs. Institutional secrets) on a 1,000-sample dataset, achieving a 45% blended OpEx reduction, 100% redaction success on personal secrets, and -- via LLM-as-a-Judge evaluation -- an 85% preference rate for APO-compressed responses over raw baselines. Our results demonstrate that Token Parsimony and Zero Leakage are mathematically dual projections of the same contextual compression operator.

Alessio Langiu

National Research Council of Italy - Institute of Marine Sciences (CNR-ISMAR)

Privacy Protection for LLM AgentsPrivacy Leakage in Language ModelsLLM RoutingLLM Context ManagementPrivacy-Preserving Language ModelsLLM Inference Efficiency+1
AskPDF
Mar 27, 2026

When Perplexity Lies: Generation-Focused Distillation of Hybrid Sequence Models

Converting a pretrained Transformer into a more efficient hybrid model through distillation offers a promising approach to reducing inference costs. However, achieving high-quality generation in distilled models requires careful joint design of both the student architecture and the distillation process. Many prior distillation works evaluate downstream multiple-choice benchmarks by ranking candidate answers with log-likelihood rather than requiring autoregressive generation, which can obscure important differences in model quality. For example, on overlapping benchmarks, we show that a 7B distilled model that nearly matches its teacher to within 0.2 pp under log-likelihood scoring falls behind by 20.8 pp when it must generate answers autoregressively. We investigate this phenomenon with GenDistill, a multi-stage pipeline we designed for distilling a pretrained Transformer into an efficient Hybrid Kimi Delta Attention (Hybrid-KDA) student. Using it as a controlled testbed on Qwen3-0.6B, we systematically ablate six design axes (training objective, loss masking, training duration, dataset selection, parameter freezing, and architecture choice) and evaluate every choice under both log-likelihood and generation-based protocols. We find that log-likelihood-based evaluation consistently underestimates the gap between teacher and student, and can in some cases reverse the ranking of design choices, so conclusions drawn from perplexity-only evaluation may be misleading. Among the factors we study, dataset selection, completion-only masking, and freezing attention layers during post-training have the largest impact on generation quality. Our best distillation recipe, using a Hybrid-KDA model as the student, retains 86-90% of teacher accuracy on knowledge benchmarks while reducing KV cache memory by up to 75% and improving time-to-first-token by 2-4x at 128K-token contexts.

Juan Gabriel Kostelec, Qinghai Guo

Huawei Switzerland · ACS Lab, Huawei Technologies

LLM Inference EfficiencyLLM EvaluationLanguage Model Generation EvaluationLanguage Model DistillationSequence ModelingGenerative AI Evaluation
AskPDF
Mar 26, 2026

Runtime Burden Allocation for Structured LLM Routing in Agentic Expert Systems: A Full-Factorial Cross-Backend Methodology

Structured LLM routing is often treated as a prompt-engineering problem. We argue that it is, more fundamentally, a systems-level burden-allocation problem. As large language models (LLMs) become core control components in agentic AI systems, reliable structured routing must balance correctness, latency, and implementation cost under real deployment constraints. We show that this balance is shaped not only by prompts or schemas, but also by how structural work is allocated across the generation stack: whether output structure is emitted directly by the model, compressed during transport, or reconstructed locally after generation. We evaluate this formulation through a comprehensive full-factorial benchmark covering 48 deployment configurations and 15,552 requests across OpenAI, Gemini, and Llama backends. Our central finding is consequential: there is no universal best routing mode. Instead, backend-specific interaction effects dominate performance. Modes that remain highly reliable on Gemini and OpenAI can suffer substantial correctness degradation on Llama, while efficiency gains from compressed realization are strongly backend-dependent. Rather than presenting another isolated model comparison, this work contributes a deployable framework for reasoning about structured routing under heterogeneous backend conditions. We provide a cross-backend evaluation methodology and practical deployment guidance for navigating the correctness-cost-latency frontier in production-grade agentic expert systems.

Zhou Hanlin, Chan Huah Yong

School of Computer Sciences, Universiti Sains Malaysia (USM), Penang, Malaysia · Xiamen Software Vocational & Technical College, Xiamen, China

LLM RoutingLLM EvaluationLLM Inference EfficiencyLLM Serving
AskPDF
Mar 26, 2026

EcoThink: A Green Adaptive Inference Framework for Sustainable and Accessible Agents

As the Web transitions from static retrieval to generative interaction, the escalating environmental footprint of Large Language Models (LLMs) presents a critical sustainability challenge. Current paradigms indiscriminately apply computation-intensive strategies like Chain-of-Thought (CoT) to billions of daily queries, causing LLM overthinking, a redundancy that amplifies carbon emissions and operational barriers. This inefficiency directly undermines UN Sustainable Development Goals 13 (Climate Action) and 10 (Reduced Inequalities) by hindering equitable AI access in resource-constrained regions. To address this, we introduce EcoThink, an energy-aware adaptive inference framework designed to reconcile high-performance AI intelligence with environmental responsibility. EcoThink employs a lightweight, distillation-based router to dynamically assess query complexity, skipping unnecessary reasoning for factoid retrieval while reserving deep computation for complex logic. Extensive evaluations across 9 diverse benchmarks demonstrate that EcoThink reduces inference energy by 40.4% on average (up to 81.9% for web knowledge retrieval) without statistically significant performance loss. By mitigating algorithmic waste, EcoThink offers a scalable path toward a sustainable, inclusive, and energy-efficient generative AI Agent.

Linxiao Li, Zhixiang Lu

The University of Sydney Sydney, Australia · University of Liverpool Liverpool, United Kingdom

LLM Inference EfficiencyEnergy-Efficient MLAdaptive ComputationOverthinking in Language Models
AskPDF
Mar 24, 2026

LLM Inference at the Edge: Mobile, NPU, and GPU Performance Efficiency Trade-offs Under Sustained Load

Deploying large language models on-device for always-on personal agents demands sustained inference from hardware tightly constrained in power, thermal envelope, and memory. We benchmark Qwen 2.5 1.5B (4-bit quantised) across four platforms: a Raspberry Pi 5 with Hailo-10H NPU, a Samsung Galaxy S24 Ultra, an iPhone 16 Pro, and a laptop NVIDIA RTX 4050 GPU. Using a fixed 258-token prompt over 20 warm-condition iterations per device, we measure throughput, latency, power, and thermal behaviour. For mobile platforms, thermal management supersedes peak compute as the primary constraint: the iPhone 16 Pro loses nearly half its throughput within two iterations, and the S24 Ultra suffers a hard OS-enforced GPU frequency floor that terminates inference entirely. On dedicated hardware, distinct constraints dominate: the RTX 4050 is bounded by its battery power ceiling, while the Hailo-10H is limited by on-module memory bandwidth. The RTX 4050 sustains 131.7 tok/s at 34.1 W; the Hailo-10H sustains 6.9 tok/s at under 2 W with near-zero variance, matching the RTX 4050 in energy proportionality at 19x lower throughput. Results should be interpreted as platform-level deployment characterisations for a single model and prompt type, reflecting hardware and software combined, rather than general claims about hardware capability alone.

Pranay Tummalapalli, Sahil Arayakandy, Ritam Pal +1

Conscious Engines

On-Device Language Model InferenceLLM Inference Efficiency
AskPDF
Mar 24, 2026

StepCache: Step-Level Reuse with Lightweight Verification and Selective Patching for LLM Serving

We address LLM serving workloads where repeated requests share a common solution structure but differ in localized constraints, such as output schema, variable names, or numeric constants. Prior caching approaches typically reuse either full responses (semantic caching) or model-internal KV/prefix states, which are respectively brittle under partial changes or tightly coupled to specific backends. We present StepCache, a backend-agnostic step-level reuse layer that segments outputs into ordered steps, retrieves the best-matching cached request, verifies steps using lightweight task-aware checks, and regenerates only failing regions via selective patching. StepCache additionally supports strict structured-output enforcement for JSON, including single-step extraction, required-key constraints, and one-shot repair, as well as conservative skip-reuse fallbacks for semantic changes. For linear equations, StepCache promotes verification into correction via a bounded repair loop with a deterministic fallback that guarantees correctness when the backend model fails. In a CPU-only perturbation-heavy micro-benchmark on math and JSON variants, averaged over three seeds, StepCache reduces mean latency from 2.13 s to 0.67 s, median latency from 2.42 s to 0.01 s, and p95 latency from 3.38 s to 3.30 s. It also reduces total token usage from 36.1k to 27.3k and improves end-to-end correctness from 72.5% to 100% under task-specific checks and a stitched-output integrity check. Across requests, 79.7% take the reuse-only fast path, 5.4% require patching, and 14.9% trigger skip-reuse.

Azam Nouri

Department of Science, Technology & Mathematics, Lincoln University, USA

LLM ServingLLM Inference EfficiencyLLM Answer Verification
AskPDF
Mar 24, 2026

Knowledge Access Beats Model Size: Memory Augmented Routing for Persistent AI Agents

Production AI agents frequently receive user-specific queries that are highly repetitive, with up to 47% being semantically similar to prior interactions, yet each query is typically processed with the same computational cost. We argue that this redundancy can be exploited through conversational memory, transforming repetition from a cost burden into an efficiency advantage. We propose a memory-augmented inference framework in which a lightweight 8B-parameter model leverages retrieved conversational context to answer all queries via a low-cost inference path. Without any additional training or labeled data, this approach achieves 30.5% F1, recovering 69% of the performance of a full-context 235B model while reducing effective cost by 96%. Notably, a 235B model without memory (13.7% F1) underperforms even the standalone 8B model (15.4% F1), indicating that for user-specific queries, access to relevant knowledge outweighs model scale. We further analyze the role of routing and confidence. At practical confidence thresholds, routing alone already directs 96% of queries to the small model, but yields poor accuracy (13.0% F1) due to confident hallucinations. Memory does not substantially alter routing decisions; instead, it improves correctness by grounding responses in retrieved user-specific information. As conversational memory accumulates over time, coverage of recurring topics increases, further narrowing the performance gap. We evaluate on 152 LoCoMo questions (Qwen3-8B/235B) and 500 LongMemEval questions. Incorporating hybrid retrieval (BM25 + cosine similarity) improves performance by an additional +7.7 F1, demonstrating that retrieval quality directly enhances end-to-end system performance. Overall, our results highlight that memory, rather than model size, is the primary driver of accuracy and efficiency in persistent AI agents.

Xunzhuo Liu, Bowei He, Xue Liu +3

vLLM Semantic Router Project · MBZUAI · McGill University +3

Memory-Augmented Language Model AgentsConversational MemoryLLM Inference EfficiencyLLM Routing
AskPDF
Mar 22, 2026

The Workload-Router-Pool Architecture for LLM Inference Optimization: A Vision Paper from the vLLM Semantic Router Project

Over the past year, the vLLM Semantic Router project has released a series of work spanning: (1) core routing mechanisms -- signal-driven routing, context-length pool routing, router performance engineering, policy conflict detection, low-latency embedding models, category-aware semantic caching, user-feedback-driven routing adaptation, hallucination detection, and hierarchical content-safety classification for privacy and jailbreak protection; (2) fleet optimization -- fleet provisioning and energy-efficiency analysis; (3) agentic and multimodal routing -- multimodal agent routing, tool selection, CUA security, and multi-turn context memory and safety; (4) governance and standards -- inference routing protocols and multi-provider API extensions. Each paper tackled a specific problem in LLM inference, but the problems are not independent; for example, fleet provisioning depends on the routing policy, which depends on the workload mix, shifting as organizations adopt agentic and multimodal workloads. This paper distills those results into the Workload-Router-Pool (WRP) architecture, a three-dimensional framework for LLM inference optimization. Workload characterizes what the fleet serves (chat vs. agent, single-turn vs. multi-turn, warm vs. cold, prefill-heavy vs. decode-heavy). Router determines how each request is dispatched (static semantic rules, online bandit adaptation, RL-based model selection, quality-aware cascading). Pool defines where inference runs (homogeneous vs. heterogeneous GPU, disaggregated prefill/decode, KV-cache topology). We map our prior work onto a 3x3 WRP interaction matrix, identify which cells we have covered and which remain open, and propose twenty-one concrete research directions at the intersections, each grounded in our prior measurements, tiered by maturity from engineering-ready to open research.

Huamin Chen, Xunzhuo Liu, Bowei He +5

vLLM Semantic Router Project · MBZUAI · McGill University +3

LLM RoutingLLM Inference ServingLLM Inference Efficiency
AskPDF
Mar 22, 2026

ConsRoute:Consistency-Aware Adaptive Query Routing for Cloud-Edge-Device Large Language Models

Large language models (LLMs) deliver impressive capabilities but incur substantial inference latency and cost, which hinders their deployment in latency-sensitive and resource-constrained scenarios. Cloud-edge-device collaborative inference has emerged as a promising paradigm by dynamically routing queries to models of different capacities across tiers. In this paper, we propose ConsRoute, a lightweight, semantic-aware, and adaptive routing framework that significantly improves inference efficiency while minimizing impact on response quality. Unlike prior routing methods that rely on predicting coarse-grained output quality gaps, ConsRoute leverages a reranker to directly assess the semantic consistency between responses generated by models at different tiers, yielding fine-grained soft supervision signals for routing. To minimize device-side overhead, ConsRoute reuses hidden states from the LLM prefilling stage as compact query representations, avoiding additional encoders or inference passes. Furthermore, these representations are clustered, and Bayesian optimization is employed to learn cluster-specific routing thresholds that dynamically balance quality, latency, and cost under heterogeneous query distributions. Extensive experiments demonstrate that ConsRoute achieves near-cloud performance (>=95%) while reducing end-to-end latency and inference cost by nearly 40%, consistently outperforming existing routing baselines in both response quality and system efficiency.

Haoyu Qiao, Hao Zhang, Shanwen Mao +2

the Department of Computer Science and Technology, Harbin Institute of Technology, Harbin 150001, China · the National Key Laboratory of Smart Farm Technologies and Systems, Harbin 150001, China · the Department of Computer Science and Technology, Harbin Institute of Technology (ShenZhen), ShenZhen 518055, China

Edge-Cloud Collaborative InferenceLLM RoutingAdaptive Model RoutingLLM Inference EfficiencyCost-Aware Inference
AskPDF
Mar 21, 2026

Clinical Note Bloat Reduction for Efficient LLM Use

Background: Clinical notes contain extensive duplicated text from templates, copy-paste, and auto-populated fields ("note bloat"), diluting clinical signal, limiting longitudinal context, and increasing large language model (LLM) costs. Methods: TRACE removes note bloat using note-level EHR metadata to identify templated and copied content, with frequency-based de-duplication when metadata are unavailable. We evaluated TRACE using blinded physician span review and gold-standard templated-text annotations across four cohorts spanning liver transplant, obstetrics, and inpatient populations at multiple health systems (5.3M notes). We compared zero-shot LLMs and embedding-based classifiers using original and TRACE-processed notes for 20 information extraction tasks and prediction of 5-year survival, postpartum hemorrhage, and 30-day readmission. Results: Only 0.3-6.6% of removed text was flagged as author-generated; TRACE captured 86% of annotated templated characters. Information extraction F1 differences averaged by cohort ranged from -0.009 to +0.004; task-specific prediction F1 differences ranged from -0.011 to +0.018. Among 1,000 randomly sampled Stanford Health Care patients, TRACE reduced chart text by 47.3% (742.7M characters), averaging 220,167 fewer tokens per patient. Using 2024 encounter volumes at a large tertiary academic center and one query per encounter, projected three-year net savings ranged from 1.00Mto1.00M to 13.58M across evaluated model pricing schemes, including initial and annual TRACE processing costs. Conclusion: TRACE substantially reduces clinical note redundancy while preserving information extraction and prediction performance. Underused EHR metadata can reduce LLM inference costs, expand usable longitudinal context, and support scalable clinical AI.

Jordan L. Cahoon, Chloe Stanwyck, Asad Aali +6

Department of Biomedical Data Science, Stanford University, Stanford, CA · Department of Pathology, Stanford University, Stanford, CA · Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University, Stanford, CA +6

LLM Inference EfficiencyElectronic Health Records
AskPDF
Mar 20, 2026

AE-LLM: Adaptive Efficiency Optimization for Large Language Models

Large Language Models (LLMs) have achieved remarkable success across diverse applications, yet their deployment remains challenging due to substantial computational costs, memory requirements, and energy consumption. Recent empirical studies have demonstrated that no single efficiency technique is universally optimal; instead, the effectiveness of methods such as efficient attention mechanisms, mixture-of-experts (MoE), parameter-efficient fine-tuning, and quantization varies significantly depending on task characteristics, resource constraints, and model scales. Building upon these insights, we propose AE-LLM, a unified framework that automatically selects and combines optimal efficiency techniques tailored to specific deployment scenarios. Our approach introduces a multi-objective optimization framework that jointly considers accuracy, latency, memory footprint, and energy consumption, while accounting for hardware constraints and task requirements. We develop an efficient search algorithm that explores the combinatorial space of efficiency techniques across architecture, fine-tuning, and inference stages, identifying Pareto-optimal configurations. Extensive experiments across 15 models (0.5B-70B parameters) and 10 diverse tasks demonstrate that AE-LLM achieves an average of 2.8×2.8\times improvement in efficiency metrics while maintaining competitive accuracy (within 1.2% of baseline), compared to static efficiency configurations. Furthermore, our framework generalizes effectively to vision-language models, achieving similar efficiency gains. Our contributions provide practitioners with an automated tool for navigating the complex trade-off landscape of LLM efficiency optimization.

Kaito Tanaka, Masato Ito, Yuji Nishimura +2

SANNO University

LLM Inference EfficiencyEnergy-Efficient ML
AskPDF
Mar 20, 2026

KV Cache Optimization Strategies for Scalable and Efficient LLM Inference

The key-value (KV) cache is a foundational optimization in Transformer-based large language models (LLMs), eliminating redundant recomputation of past token representations during autoregressive generation. However, its memory footprint scales linearly with context length, imposing critical bottlenecks on GPU memory capacity, memory bandwidth, and inference throughput as production LLMs push context windows from thousands to millions of tokens. Efficient KV cache management has thus become a first-order challenge for scalable LLM deployment. This paper provides a systematic review of recent KV cache optimization techniques, organizing them into five principal directions: cache eviction, cache compression, hybrid memory solutions, novel attention mechanisms, and combination strategies. For each category we analyze the underlying mechanisms, deployment trade-offs, and empirical performance across memory reduction, throughput, and model accuracy metrics. We further map techniques to seven practical deployment scenarios, including long-context single requests, high-throughput datacenter serving, edge devices, multi-turn conversations, and accuracy-critical reasoning, providing actionable guidance for practitioners selecting among competing approaches. Our analysis reveals that no single technique dominates across all settings; instead, the optimal strategy depends on context length, hardware constraints, and workload characteristics, pointing toward adaptive, multi-stage optimization pipelines as a promising direction for future research.

Yichun Xu, Navjot K. Khaira, Tejinder Singh

Dell Technologies, Hopkinton, MA 01748, USA · Dell Technologies, Santa Clara, CA 95054, USA

KV-Cache ManagementLLM Inference EfficiencyKV-Cache CompressionKV-Cache EvictionKV-Cache Offloading
AskPDF
Mar 20, 2026

PoC: Performance-oriented Context Compression for Large Language Models via Performance Prediction

While context compression can mitigate the growing inference costs of Large Language Models (LLMs) by shortening contexts, existing methods that specify a target compression ratio or length suffer from unpredictable performance degradation, hindering their reliable deployment. We introduce a paradigm shift to Performance-oriented Context Compression (PoC), where developers specify an acceptable performance floor instead of a compression ratio. PoC employs a lightweight performance predictor to automatically find the most aggressive compression ratio that satisfies this constraint before steering an off-the-shelf compressor. We design and compare two predictor variants: a simple context-agnostic predictor and a more sophisticated context-aware one that considers the input's inherent compressibility. On both question-answering and summarization benchmarks, the context-aware predictor consistently achieves lower performance prediction error than the context-agnostic predictor, while the resulting context-aware PoC attains a superior overall performance. Our work paves the way for a more reliable, efficient, and performance-aware deployment of context compression for LLMs.

Runsong Zhao, Shilei Liu, Jiwei Tang +8

Northeastern University, China · Future Living Lab of Alibaba · Tsinghua University

Language Model Performance PredictionLLM Context CompressionContext CompressionLLM Context ManagementLLM Inference Efficiency
AskPDF
Mar 18, 2026

RAMP: Reinforcement Adaptive Mixed Precision Quantization for Efficient On Device LLM Inference

Post training quantization is essential for deploying large language models (LLMs) on resource constrained hardware, yet state of the art methods enforce uniform bit widths across layers, yielding suboptimal accuracy efficiency trade offs. We present RAMP (Reinforcement Adaptive Mixed Precision), an off policy Soft Actor Critic framework that learns per layer bit width assignments to minimize perplexity under a global bit budget. The policy conditions on an 11 dimensional embedding of activation statistics, weight properties, and structural descriptors, enabling zero shot transfer across model families and scales. To enable stable sub 4 bit quantization, we introduce Scale Folding, a preconditioning technique that migrates activation outliers into weights via per channel scaling and normalization layer compensation. A quality prioritized reward with asymmetric penalties and budget cliffs drives rapid convergence. On Llama 2 7B, RAMP achieves 5.54 perplexity at 3.68GB (3.65 effective bits), outperforming uniform 4 bit AWQ (5.60 at 3.90 GB) and GPTQ by 6% in size and 1% to3% in quality. Critically, a policy trained only on Llama 2 7B generalizes zero shot to Llama 2 13B and Mistral 7B, often surpassing target specific training, supporting the hypothesis that quantization sensitivity is primarily architectural. The HALO pipeline exports allocations to GGUF format for kernel free inference on CPUs, GPUs, and edge devices, retaining 99.5% of FP16 commonsense reasoning performance.

Arpit Singh Gautam, Saurabh Jha

LLM QuantizationNeural Network QuantizationLLM Inference EfficiencyMixed-Precision QuantizationPost-Training Quantization
AskPDF
Mar 17, 2026

Resource Consumption Threats in Large Language Models

Given limited and costly computational infrastructure, resource efficiency is a key requirement for large language models (LLMs). Efficient LLMs increase service capacity for providers and reduce latency and API costs for users. Recent resource consumption threats induce excessive generation, degrading model efficiency and harming both service availability and economic sustainability. This survey presents a systematic review of threats to resource consumption in LLMs. We further establish a unified view of this emerging area by clarifying its scope and examining the problem along the full pipeline from threat induction to mechanism understanding and mitigation. Our goal is to clarify the problem landscape for this emerging area, thereby providing a clearer foundation for characterization and mitigation.

Yuanhe Zhang, Xinyue Wang, Zhican Chen +8

Beijing University of Posts and Telecommunications · Hangzhou Dianzi University · Nanyang Technological University +1

LLM SecurityLLM Inference EfficiencyAdversarial Attacks on LLMs
AskPDF
Mar 13, 2026

Token-Budget-Aware Pool Routing for Cost-Efficient LLM Inference

Production vLLM fleets provision every instance for worst-case context length, wasting 4-8x concurrency on the 80-95% of requests that are short and simultaneously triggering KV-cache failures -- OOM crashes, preemption storms, and request rejections. Both problems share a single root cause: configuration-traffic mismatch. We propose token-budget-aware pool routing: estimate each request's total token budget using a self-calibrating per-category bytes-per-token ratio, then dispatch it to one of two vLLM pools -- a high-throughput short pool or a high-capacity long pool -- each right-sized for its workload class. The ratio is learned online via exponential moving average from usage.prompt_tokens feedback, requiring no tokenizer. A closed-form cost model, savings = alpha * (1 - 1/rho), predicts fleet-level GPU savings from two observable quantities: the short-traffic fraction alpha and the throughput gain ratio rho. On traces from the Azure LLM Inference Dataset and LMSYS-Chat-1M serving Llama-3-70B on A100 GPUs, token-budget routing reduces GPU instances by 17-39% ($1.2-2.0M/yr at 1,000 req/s), with savings verified by a self-contained discrete-event simulator. A case study projecting Qwen3-235B-A22B on AMD MI300X at 10,000 req/s shows $15.4M/yr in savings. The algorithm adds O(1) dispatch overhead, self-calibrates across content types without a tokenizer, and composes with PagedAttention, continuous batching, and prefill-decode disaggregation.

Huamin Chen, Xunzhuo Liu, Junchen Jiang +2

vLLM Semantic Router Project · Tensormesh Inc / UChicago · MBZUAI / McGill University

LLM Inference EfficiencyLLM RoutingCost-Aware InferenceLLM Request Scheduling
AskPDF
Mar 12, 2026

TaxBreak: Unmasking the Hidden Costs of LLM Inference Through Overhead Decomposition

Large Language Model (LLM) inference is widely used in interactive assistants and agentic systems. In latency-sensitive deployments, inference time can become dominated by host-side overheads. Existing approaches typically expose this cost only as an aggregate residual or a launch/queue metric, which is often insufficient to identify which execution layer should be optimized. This work presents TaxBreak, a trace-driven methodology for decomposing host-visible orchestration overhead into three components: framework translation time, CUDA library translation time, and kernel launch-path time. We validate TaxBreak on NVIDIA H100 and H200 systems and use it to derive our proposed Host-Device Balance Index (HDBI), a boundedness summary index that relates device-active execution to host-visible orchestration. Across representative dense and mixture-of-experts workloads in both prefill and decode, we show that aggregate latency, GPU inactivity, or boundedness ratios alone can obscure the dominant optimization target. TaxBreak instead distinguishes cases where optimization should reduce software-stack overhead from cases where the primary win comes from reducing device-side work. We further show that MoE models dispatch 8-11x more kernels per output token than dense models, and that for such host-bound workloads, CPU single-thread performance is a first-order parameter: a faster host CPU reduces orchestration overhead by 10-29% and improves end-to-end latency by up to 14%, even when paired with a slower-clocked GPU. These results position TaxBreak as a diagnostic tool for assessing whether optimization effort should target the software stack or the device-side workload execution.

Prabhu Vellaisamy, Shreesh Tripathi, Vignesh Natarajan +3

NeuroAI Computer Architecture Lab (NCAL), Electrical and Computer Engineering Department, Carnegie Mellon University

LLM Inference EfficiencyLLM Inference
AskPDF
Mar 11, 2026

CacheSolidarity: Preventing Prefix Caching Side Channels in Multi-Tenant LLM Serving Systems

Large Language Models (LLMs) rely on optimizations like Automatic Prefix Caching (APC) to accelerate inference. APC works by reusing previously computed states for the beginning part of a request (prefix), when another request starts with the same text. While APC improves throughput, it introduces timing side channels: cache hits are faster than misses, creating observable latency differences. In multi-tenant systems, attackers can exploit these differences to infer sensitive information, e.g., by incrementally reconstructing another user's request by observing hit/miss patterns. Current defenses take a sledgehammer approach: they disable APC and cache sharing, isolating users, and sacrificing efficiency for regular users. This paper presents CacheSolidarity, a system that secures multi-tenant LLM serving systems against APC side channels without sacrificing performance and efficiency. CacheSolidarity monitors cache reuse across users, flags suspicious sharing, and selectively isolates prefixes, restricting their reuse only when necessary. Evaluation shows that CacheSolidarity enables up to 70% higher cache reuse and 30% lower inference latency compared to existing user-isolation defenses. CacheSolidarity's lightweight design demonstrates how LLM serving can be secured without sacrificing performance or incurring unbearable overheads.

Panagiotis Georgios Pennas, Konstantinos Papaioannou, Marco Guarnieri +1

IMDEA Software Institute Universidad Politécnica de Madrid · IMDEA Software Institute

LLM SecurityPrefix CachingLLM Inference Efficiency
AskPDF
Mar 9, 2026

Quantifying the Accuracy and Cost Impact of Design Decisions in Budget-Constrained Agentic LLM Search

Agentic Retrieval-Augmented Generation (RAG) systems combine iterative search, planning prompts, and retrieval backends, but deployed settings impose explicit budgets on tool calls and completion tokens. We present a controlled measurement study of how search depth, retrieval strategy, and completion budget affect accuracy and cost under fixed constraints. Using Budget-Constrained Agentic Search (BCAS), a model-agnostic evaluation harness that surfaces remaining budget and gates tool use, we run comparisons across six LLMs and three question-answering benchmarks. Across models and datasets, accuracy improves with additional searches up to a small cap, hybrid lexical and dense retrieval with lightweight re-ranking produces the largest average gains in our ablation grid, and larger completion budgets are most helpful on HotpotQA-style synthesis. These results provide practical guidance for configuring budgeted agentic retrieval pipelines and are accompanied by reproducible prompts and evaluation settings.

Kyle McCleary, James Ghawaly

Division of Computer Science & Engineering, Louisiana State University, Baton Rouge, Louisiana, USA

Agentic SearchAgentic RAGRAG EvaluationLLM Inference Efficiency
AskPDF
Mar 9, 2026

Ares: Adaptive Reasoning Effort Selection for Efficient LLM Agents

Modern agents powered by thinking LLMs achieve high accuracy through long chain-of-thought reasoning but incur substantial inference costs. While many LLMs now support configurable reasoning levels (e.g., high/medium/low), static strategies are often ineffective: using low-effort modes at every step leads to significant performance degradation, while random selection fails to preserve accuracy or provide meaningful cost reduction. However, agents should reserve high reasoning effort for difficult steps like navigating complex website structures, while using lower-effort modes for simpler steps like opening a target URL. In this paper, we propose Ares, a framework for per-step dynamic reasoning effort selection tailored for multi-step agent tasks. Ares employs a lightweight router to predict the lowest appropriate reasoning level for each step based on the interaction history. To train this router, we develop a data generation pipeline that identifies the minimum reasoning effort required for successful step completion. We then fine-tune the router to predict these levels, enabling plug-and-play integration for any LLM agents. We evaluate Ares on a diverse set of agent tasks, including TAU-Bench for tool use agents, BrowseComp-Plus for deep-research agents, and WebArena for web agents. Experimental results show that Ares reduces reasoning token usage by up to 52.7% compared to fixed high-effort reasoning, while introducing minimal degradation in task success rates.

Jingbo Yang, Bairu Hou, Wei Wei +2

Department of Computer Science, University of California, Santa Barbara · Center for Advanced AI, Accenture

Efficient Language Model ReasoningLLM Inference EfficiencyLLM Agent Workflow Optimization
AskPDF
Mar 6, 2026

The Compression Paradox in LLM Inference: Provider-Dependent Energy Effects of Prompt Compression

The rapid proliferation of Large Language Models has created an environmental paradox: the very technology that could help solve climate challenges is itself becoming a significant contributor to global carbon emissions. We test whether prompt compression improves inference energy efficiency in 28,421 successful API trials (28,428 planned) across three providers (OpenAI GPT-4o-mini, Anthropic Claude-3.5-Sonnet, and DeepSeek-Chat), five benchmarks (HumanEval, MBPP, GSM8K, MATH, MMLU), and four compression ratios (r in {1.0, 0.7, 0.5, 0.3}). Energy is estimated with a token-based proxy calibrated against local direct measurements, and quality is tracked with benchmark pass rates. Compression produced substantial quality loss (overall pass rate 26.0% at baseline vs. 1.5% at r=0.7) and strongly provider-dependent energy behavior. DeepSeek exhibited output expansion under compression (21 to 798 tokens at r=0.3), corresponding to energy increases up to +2,140%, while GPT-4o-mini showed mixed effects including a reduction at r=0.5. These results indicate that input-token reduction alone is not a reliable energy optimization strategy in production inference. For the evaluated settings, model selection and output-length control provided more consistent energy-quality tradeoffs than prompt compression.

Warren Johnson

Plexor Labs Principal Researcher

LLM Inference EfficiencyEnergy-Aware LLM InferenceEfficient Language Model InferencePrompt Compression
AskPDF
Mar 6, 2026

Compression Method Matters: Benchmark-Dependent Output Dynamics in LLM Prompt Compression

Prompt compression is often evaluated by input-token reduction, but its real deployment impact depends on how compression changes output length and total inference cost. We present a controlled replication and extension study of benchmark-dependent output dynamics under aggressive compression, covering 5,400 API calls across three benchmarks and multiple providers. To explain conflicting prior observations, we formalize instruction survival probability (Psi), a structural metric that captures whether task-critical prompt segments remain after truncation. Results show a strong benchmark effect: under r=0.3, DeepSeek exhibits severe output expansion on MBPP (56x, Psi approx 0.15) but substantially lower expansion on HumanEval (5x, Psi approx 0.72), while GPT-4o-mini is comparatively stable across benchmarks. This reconciles the apparent discrepancy between previously reported extreme explosion and lower replication effects by identifying prompt structure, not provider identity alone, as the primary moderator. We introduce the Compression Robustness Index (CRI) for cross-benchmark evaluation and show that single-benchmark assessments can produce misleading conclusions about compression safety and efficiency. To contextualize energy claims, we incorporate companion direct NVML measurements from rented RunPod GPUs and show that token savings can overstate joule savings. These findings motivate benchmark-diverse testing and structure-aware compression policies for reliable, energy-conscious LLM deployment.

Warren Johnson

Plexor Labs Principal Researcher

Prompt CompressionLLM Inference EfficiencyEfficient Language Model InferenceLLM Evaluation
AskPDF
Mar 6, 2026

MoEless: Efficient MoE LLM Serving with Serverless Experts

Large Language Models (LLMs) increasingly adopt Mixture-of-Experts (MoE) architectures to scale efficiently under stringent resource constraints. However, MoE's sparse activation causes severe expert load imbalance, where a few experts become stragglers while others remain underutilized, leading to inflated inference latency and cost. Existing solutions assume static, serverful model deployments, limiting expert elasticity and often incurring costly expert swapping or degraded output quality. We present MoEless, an efficient serverless MoE serving framework that mitigates expert load imbalance via elastic expert execution. MoEless leverages lightweight, layer-aware predictors to estimate incoming expert load distributions and proactively identify stragglers. We design optimized scaling and placement strategies to improve function locality, GPU utilization, and cross-expert load balance. MoEless is prototyped on top of Megatron-LM and deployed on an eight-GPU testbed. Experiments with open-source MoE models and real-world workloads show that MoEless reduces inference latency by 43% and inference cost by 84% compared to state-of-the-art solutions.

Hanfei Yu, Bei Ouyang, Shwai He +2

University of Maryland College Park

LLM Inference EfficiencyExpert Load BalancingLLM ServingMixture-of-Experts Inference
AskPDF
Mar 5, 2026

Sparse-BitNet: 1.58-bit LLMs are Naturally Friendly to Semi-Structured Sparsity

Semi-structured N:M sparsity and low-bit quantization (e.g., 1.58-bit BitNet) are two promising approaches for improving the efficiency of large language models (LLMs), yet they have largely been studied in isolation. In this work, we investigate their interaction and show that 1.58-bit BitNet is naturally more compatible with N:M sparsity than full-precision models. To study this effect, we propose Sparse-BitNet, a unified framework that jointly applies 1.58-bit quantization and dynamic N:M sparsification while ensuring stable training for the first time. Across multiple model scales and training regimes (sparse pretraining and dense-to-sparse schedules), 1.58-bit BitNet consistently exhibits smaller performance degradation than full-precision baselines at the same sparsity levels and can tolerate higher structured sparsity before accuracy collapse. Moreover, using our custom sparse tensor core, Sparse-BitNet achieves substantial speedups in both training and inference, reaching up to 1.30X. These results highlight that combining extremely low-bit quantization with semi-structured N:M sparsity is a promising direction for efficient LLMs. Code available at https://github.com/AAzdi/Sparse-BitNet

Di Zhang, Xun Wu, Shaohan Huang +9

Microsoft Research · Peking University · South China University of Technology

LLM Inference EfficiencyLLM QuantizationLow-Bit QuantizationStructured Sparsity
AskPDF
Mar 5, 2026

Beyond the Context Window: A Cost-Performance Analysis of Fact-Based Memory vs. Long-Context LLMs for Persistent Agents

Persistent conversational AI systems face a choice between passing full conversation histories to a long-context large language model (LLM) and maintaining a dedicated memory system that extracts and retrieves structured facts. We compare a fact-based memory system built on the Mem0 framework against long-context LLM inference on three memory-centric benchmarks - LongMemEval, LoCoMo, and PersonaMemv2 - and evaluate both architectures on accuracy and cumulative API cost. Long-context GPT-5-mini achieves higher factual recall on LongMemEval and LoCoMo, while the memory system is competitive on PersonaMemv2, where persona consistency depends on stable, factual attributes suited to flat-typed extraction. We construct a cost model that incorporates prompt caching and show that the two architectures have structurally different cost profiles: long-context inference incurs a per-turn charge that grows with context length even under caching, while the memory system's per-turn read cost remains roughly fixed after a one-time write phase. At a context length of 100k tokens, the memory system becomes cheaper after approximately ten interaction turns, with the break-even point decreasing as context length grows. These results characterize the accuracy-cost trade-off between the two approaches and provide a concrete criterion for selecting between them in production deployments.

Natchanon Pollertlam, Witchayut Kornsuwannawit

Bricks Technology

Persistent Memory for Language ModelsLong-Term Conversational MemoryLong-Context LLM InferenceLLM Inference Efficiency
AskPDF
Mar 5, 2026

SLO-Aware Compute Resource Allocation for Prefill-Decode Disaggregated LLM Inference

Prefill-Decode (P/D) disaggregation has emerged as a widely adopted optimization strategy for Large Language Model (LLM) inference. However, there currently exists no well-established methodology for determining the optimal number of P/D hardware resources, subject to constraints on total throughput, service level objectives (SLOs), and request characteristics - specifically input and output lengths. To address this gap, we propose a hybrid approach that combines theoretical modeling with empirical benchmarking. First, we present a theoretical model for calculating P/D resource counts, which is based on total throughput requirements, request input and output lengths, as well as prefill and decode throughput. Then, to obtain the actual prefill and decode throughput under SLO constraints, we model the prefill process using M/M/1 queuing theory, deriving the achieved prefill throughput from the benchmarked maximum prefill throughput and Time-To-First-Token (TTFT). For the decode phase, we determine the decode batch sizes that meet Time-Per-Output-Token (TPOT) requirements and obtain the corresponding decode throughput through empirical measurements. Our experimental results demonstrate that the proposed method can accurately predict optimal P/D resource allocation in real-world LLM inference scenarios.

Luchang Li, Dongfang Li, Bozhao Gong +1

Kingsoft Cloud

Disaggregated LLM ServingPrefill-Decode DisaggregationLLM Inference EfficiencyLLM Inference Scheduling
AskPDF
Mar 4, 2026

Beyond Test-Time Compute Strategies: Advocating Energy-per-Token in LLM Inference

Large Language Models (LLMs) demonstrate exceptional performance across diverse tasks but come with substantial energy and computational costs, particularly in request-heavy scenarios. In many real-world applications, the full scale and capabilities of LLMs are often unnecessary, as Small Language Models (SLMs) can provide accurate responses for simpler text generation tasks. When enhanced with advanced reasoning strategies, such as Chain-of-Thought (CoT) prompting or Majority Voting, SLMs can approach the performance of larger models while reducing overall computational requirements. However, these strategies can also introduce additional energy costs, creating an energy-accuracy trade-off. Our analysis examines these trade-offs in test-time compute strategies for smaller models compared to larger ones, using the MMLU benchmark. Additionally, we explore the input-output token dynamics of transformer architectures, which result in nonlinear hardware energy operation curves for LLMs. To bridge AI research with its physical impact, we propose energy efficiency metrics, including Energy-per-Token, as complements to traditional accuracy benchmarks. Beyond model selection, we propose controlled reasoning in CoT token generation, using operating curves to regulate reasoning depth dynamically. This vision integrates a energy-aware routing mechanism, ensuring that model selection and inference strategies balance accuracy for sustainable AI deployment.

Patrick Wilhelm, Thorsten Wittkopp, Odej Kao

BIFOLD, Technische Universität Berlin, Berlin, Germany

Energy-Aware LLM InferenceLLM Inference EfficiencyEfficient Language Model ReasoningSmall Language ModelsCoT Reasoning
AskPDF
Mar 3, 2026

Parallel Test-Time Scaling with Multi-Sequence Verifiers

Parallel test-time scaling, which generates multiple candidate solutions for a single problem, is a powerful technique for improving large language model performance. However, it is hindered by two key bottlenecks: accurately selecting the correct solution from the candidate pool, and the high inference latency from generating many full solutions. We argue that both challenges are fundamentally linked to verifier calibration, as a well-calibrated verifier improves answer selection and enables early-stopping strategies to reduce latency. However, existing non-generative verifiers are limited as they score each candidate in isolation, overlooking rich contextual information across the set of candidates. To address this, we introduce the Multi-Sequence Verifier (MSV), a lightweight verifier that predicts each candidate's correctness conditioned on the full sampled set. MSV achieves improved calibration, which directly enhances best-of-N selection performance and empowers a novel early-stopping framework. Across challenging mathematical reasoning benchmarks, MSV improves best-of-64 accuracy by up to 6% relative to strong baselines, and in the early-stopping setting reaches the same accuracy as baselines with less than half the latency.

Yegon Kim, Seungyoo Lee, Chaeyun Jang +2

Graduate School of AI, KAIST, Seoul, South Korea

LLM Answer VerificationTest-Time ScalingLLM Inference Efficiency
AskPDF
Mar 2, 2026

Same Pieces, Different Servers: A Tetris Benchmark for AI Agents as Served

An agent meets a model as served: through an endpoint with a price card, a shared cache and other tenants, or on whatever hardware a self-hosted model runs. Benchmarks rank the weights. We introduce a Tetris benchmark that measures what agents get from models as served: every move is scored against an oracle, and every agent receives the same pieces. In five pre-specified experiments with nine open-weight models on one serverless provider, plus an open decision model self-hosted on a CPU, we find that price and size do not predict decision quality; that resending history costs almost nothing when cached input is free, would cost eleven to twelve times more if it were not, and makes play worse; that deployments keep between 4 and more than 64 agent contexts warm, in line with their throughput rather than the model's KV-cache size; that an agent's own long requests slow its slowest short decisions more than tenfold, which a simple admission rule cuts by 59% without hurting play; and that the decision model plays mid-pack but takes 27 s per move on a CPU, about 650 times longer than reported on a GPU. Architecture predicts some of what an agent sees; the deployment sets the rest.

Haochuan Wang

Massachusetts Institute of Technology

LLM Inference EfficiencyLLM Agent EvaluationLLM Inference ServingAI Agent BenchmarksLLM Agent Serving
AskPDF