Power capping is illusory in LLM decode as memory-bound operation leaves power headroom untouched on 700 W GPUs, while SM clock locking saves up to 32% energy and three DVFS classes appear across attention types.
Polca: Power oversubscription in llm cloud providers
4 Pith papers cite this work. Polarity classification is still indexing.
4
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 4verdicts
UNVERDICTED 4roles
background 1polarities
background 1representative citing papers
A classifier using NVML telemetry identifies ML training workloads at 98.2% accuracy and retains 43-87% accuracy against the strongest tested adversarial evasions across 9 GPUs and 5 iteration rounds.
nvPAX is a three-phase hybrid QP/LP policy that allocates power in hierarchical multi-tenant datacenters to achieve 98.92% mean satisfaction ratio on production GPU traces while outperforming static and greedy baselines.
CompPow makes the case that component-aware power management inside GPUs can yield 10% higher energy efficiency and 5% better performance for ML workloads.
citing papers explorer
-
The Illusion of Power Capping in LLM Decode: A Phase-Aware Energy Characterisation Across Attention Architectures
Power capping is illusory in LLM decode as memory-bound operation leaves power headroom untouched on 700 W GPUs, while SM clock locking saves up to 32% energy and three DVFS classes appear across attention types.
-
Detecting Hidden ML Training With Zero-Overhead Telemetry
A classifier using NVML telemetry identifies ML training workloads at 98.2% accuracy and retains 43-87% accuracy against the strongest tested adversarial evasions across 9 GPUs and 5 iteration rounds.
-
nvPAX: Constrained Optimization for Dynamic Power Allocation in Hierarchical and Multi-Tenant Systems
nvPAX is a three-phase hybrid QP/LP policy that allocates power in hierarchical multi-tenant datacenters to achieve 98.92% mean satisfaction ratio on production GPU traces while outperforming static and greedy baselines.
-
CompPow: A Case for Component-level GPU Power Management
CompPow makes the case that component-aware power management inside GPUs can yield 10% higher energy efficiency and 5% better performance for ML workloads.