arXiv subjects

Stephen Abkin

Publications and source records attributed to Stephen Abkin.

5 recordsLinked to original sources

arXiv2026

Understanding the Energy Cost of Quantized Agents on Shared GPUs

Quantization can lower the energy of language-model calls and free memory for serving more agents, but its benefit per completed task remains unclear when agents share a GPU. Measurements of fixed requests or isolated agents leave open how the agents' generated work interact with competition for compute and cache. We present a systematic study of precision and admission limits for two agent models on A100 deployments with 40 or 80 GB of memory, using coding and data-analysis tasks with independently verified outcomes. We replay recorded calls across precisions to isolate execution energy and cap the cache to test the value of freed memory. For both models with four admitted agents where requests fit in the cache, 4-bit weights with 16-bit activations execute the same calls with 10-25% less energy than 8-bit weights and activations. Despite this saving, extra work and fewer completions in live runs raise their energy per verified completion by 49%. Where requests compete for cache on the 40 GB deployment, 4-bit weights cost less, and capping their cache raises energy per verified completion by 34% with 16 agents. These results show how generated work and cache demand change the energy benefit of quantization, providing a practical basis for choosing precision and admission limits together.

cs.PF↗

arXiv2026

Measuring the Energy Cost of Completing Tasks with Quantized Language-Model Agents

Quantization reduces the memory and energy needed to serve large language models (LLMs) and AI agents for coding and data analysis can now run on locally hosted LLMs. In practice, this benefit is commonly benchmarked per output token. Yet an agent calls its model repeatedly, runs tools, and revises its work, and quantization can change both how much the agent generates and whether it succeeds. Existing energy studies either compare quantized models on single requests or measure agents at one fixed precision, overlooking how a change in precision plays out across a complete agent task. To address this gap, this paper presents the first characterization of the task-level energy trade-offs of model quantization in locally hosted agents. We develop 29 agent tasks with automatic pass/fail checks in six families, ranging from single-answer SQL analytics to multi-step repository bug fixing, and measure CPU and GPU energy per output token and per completed task for 4-bit and 8-bit versions of two open-weight models on bare-metal NVIDIA A100 nodes with llama$.$cpp. Our findings reveal that 4-bit precision generally saves energy per token, though whether it saves energy per completed task is task dependent and changes with the model and the execution budget (i.e., limits on turns, tokens, and time). We trace this discrepancy to the extra tokens needed to complete a task in full, which translates into using up to 3.5 times as much energy per completed task. These findings offer actionable guidance for developers and operators choosing model precision for agents, showing how tokens per completed task, set against the per-token saving, reveal whether lower precision completes the work with less energy.

cs.PF↗

arXiv2026

Energy per Successful Goal: Goal-Level Energy Accounting for Agentic AI Systems

Current AI energy benchmarks measure consumption per model invocation. For agentic systems, where a single user goal may trigger multistep orchestration, tool calls, retries, and failure recovery, the invocation count is an implementation artifact rather than a task property, and per inference normalization misrepresents the energy cost of completing a goal. We present A-LEMS (Agentic LLM Energy Measurement System), a cross layer framework that redefines the unit of AI energy accounting as Energy per Successful Goal (EpG): total workflow energy across all attempts, including failures and retries, divided by successfully completed goals. A-LEMS combines a temporal boundary model, a five layer attribution pipeline mapping hardware signals (RAPL on x86; SPBM and DCGM on ARM) to workflow energy, and a reproducibility protocol binding every measurement to hardware, environment, and runtime configuration. Building on EpG, we define the Orchestration Overhead Index (OOI), the ratio of agentic to linear EpG on matched goals under identical success criteria. Across nine task families on Intel x86 and NVIDIA Grace GB10 ARM, CPU only OOI ranges 0.35x to 11.07x. Including GPU dynamic energy, which accounts for 83.8% of agentic EpG in local inference, widens this to 0.4x to 26.48x, so CPU only measurement understates overhead about twofold for reasoning tasks. For remote inference, network wait energy shifts OOI from 2.5x to 2.35x. The overhead arises from orchestration structure (planning, retries, recovery) that induces additional inference calls, not from costlier individual calls. The metric is not biased upward: single tool calculator dispatch inverts to OOI = 0.61x (95% CI [0.35, 0.97]) on x86. Goal level accounting via EpG and OOI provides a measurement foundation for benchmarking agentic AI.

cs.AI↗

arXiv2026

Sign-switched antithetic coupling in gradient random-walk methods for convection-diffusion equations

We propose an antithetic coupling that reduces the computational cost of gradient random-walk methods for one-dimensional convection-diffusion equations. The methods represent the solution's spatial derivative by particles with positive or negative masses, whose cumulative sums reconstruct the solution, and model diffusion by Gaussian displacements. The particles reorder and, for a nonlinear flux, interact. Antithetic sampling negates the displacements of a second simulation, which positively correlates the contributions of particles matched by position when their masses have opposite signs. Our coupling instead gives matched particles of opposite sign identical displacements and retains opposite displacements for equal-sign pairs. We prove that each simulation keeps the distribution of an uncoupled run as the matching changes, so the averaged solution keeps the bias of one simulation. For fixed particle configurations and matching, the coupling minimizes each matched pair's conditional variance contribution after diffusion among all couplings with the same Gaussian marginal distributions. From the same configurations, its variance reduction over antithetic sampling at the final diffusion stage has a closed form for any flux. For the heat equation, a martingale decomposition extends both results to each step's contribution to the terminal variance. In a Burgers' equation test, the coupling meets a tolerance for the spatially integrated mean squared error with 4 simulations, against 34 for antithetic sampling and 65 for independent sampling. In a second test, it meets a common error tolerance in about 40% less time than pairing particles only within each sign.

math.NA↗

arXiv2026

Error attribution in gradient random walk methods for parabolic equations

Gradient random walk (GRW) methods represent a solution's spatial derivative by weighted particles and recover the solution by cumulative summation. We examine their accuracy for the one-dimensional heat, scalar cubic reaction-diffusion, and Burgers' equations. Thirty-seed ensembles, paired reconstructions, and deterministic controls identify the sources of two apparent accuracy limits. For the heat equation, comparing cumulative bin sums at bin centers introduces a first-order alignment error, which the exact reflected-particle distribution predicts and bin-edge comparison removes. An identity for the expected squared error predicts the remaining floor and the particle count for a target root-mean-square accuracy. With unequal weights, a count chosen in advance gives an error within 1% of its prediction. Reaction-front position, speed, and shape errors decrease under refinement. For Burgers' equation, exact transformed boundary data reduce the deterministic Cole-Hopf error about 490-fold, while fixed-endpoint particle runs approach their deterministic reference.

math.NA↗