Night-Window Batching versus Carbon-Aware Scheduling for Clinical AI GPU Workloads
2026-06-01 • Distributed, Parallel, and Cluster Computing
Distributed, Parallel, and Cluster ComputingEmerging Technologies
AI summaryⓘ
The authors used computer simulations to test different ways of scheduling GPU-powered machine learning jobs in hospitals, focusing on balancing urgent cases and reducing carbon emissions from electricity use. They found that running non-urgent jobs overnight captures most of the carbon savings of a more complex scheduling rule, while causing fewer urgent deadline misses. Strict carbon-focused scheduling led to many urgent jobs missing deadlines, showing it can disrupt clinical priorities. Their results suggest simpler overnight batching can be an effective, low-impact way to reduce carbon without hurting urgent job processing.
GPU schedulingmachine learningcarbon footprintjob prioritizationqueueing simulationurgent deadlinescarbon-aware computingbatch processinghospital computing
Authors
Nishi Doshi, Shrey Shah
Abstract
Hospitals run more machine learning on GPUs while the carbon footprint of grid electricity rises and falls through the day. Using a computer simulation, we compare $13$ scheduling rules on mixed GPU hardware, with synthetic patient-style jobs, urgency tiers, and time-of-day carbon traces. We do not study patient outcomes; every percentage we report is a simulator queue number, not a clinical finding. We ask whether running non-urgent jobs overnight is almost as good as a richer rule that mixes urgency and carbon (CUCA at weight 0.45, written CUCA$_{0.45}$). The comparison keeps carbon reduction secondary to clinical priority and deadline compliance, so each policy is judged on both average kg CO$_2$e and missed-deadline behavior. CarbonGreedy and CarbonShift are carbon-first stress tests that demonstrate how poorly wrong vendor presets can disrupt clinical priorities, and are not meant for production. Numbers are averages over many test settings, with wide run-to-run spread and no statistical adjustment, so headline ratios are exploratory. On an eight-GPU baseline, the overnight rule closes about $78\%$ of the carbon gap between urgency-only and CUCA$_{0.45}$ while missing fewer urgent deadlines than either. CarbonShift lets about $46\%$ of the most urgent jobs miss their deadline; this is simulated queueing, not bedside harm. At $48$ jobs per hour, the carbon footprints almost tie, yet the overnight rule still misses fewer urgent deadlines. A geography test, where regions share one daily carbon shape with only timezone shifts, trims under one percentage point of average carbon; a twelve-hour routine window saves a little carbon for CUCA$_{0.45}$ but raises overall missed deadlines. Overnight batching stays competitive on average modelled carbon; carbon-only rules belong only in stress tests.