Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
QEIL v2 optimizes LLM deployment on heterogeneous edge devices by replacing static heuristics with physics-grounded, runtime-adaptive models using DASI, CPQ, and Phi metrics to form a unified energy equation. It employs PGSAM for multi-objective optimization (energy, latency, device utilization) and EAC/ARDE with CSVET for inference-time verification. Evaluated on various benchmarks and model families, QEIL v2 achieves a 2.86x improvement over standard inference, surpassing the IPW=1.0 empirical reference mark with a 4-bit Llama-3.1-8B model, while also reducing energy consumption and latency.
Edge LLM orchestration finally beats the "1.0 IPW" barrier, thanks to physics-aware resource allocation that slashes energy by 75% and latency by 38%.
Deploying large language models (LLMs) on heterogeneous edge devices demands frameworks that jointly optimize energy efficiency, inference quality, and reliability. Our prior QEIL v1 (Kumar&Jha, 2026) achieved 4.82x IPW improvement but relied on static efficiency factors, greedy optimization, and unverified candidate selection. QEIL v2 replaces every static heuristic with physics-grounded, runtime-adaptive models. We introduce three device-workload metrics: DASI (roofline-derived compute utilization), CPQ (memory pressure from allocation theory), and Phi (thermal yield from CMOS leakage physics), forming a unified energy equation with every coefficient traceable to semiconductor physics. For optimization, PGSAM (Pareto-Guided Simulated Annealing with Momentum) simultaneously minimizes energy, latency, and device underutilization. At inference time, the EAC/ARDE selection cascade with CSVET early stopping provides progressive verification among repeated samples. Evaluated on WikiText-103, GSM8K, and ARC-Challenge across seven model families (125M-8B parameters, including one pre-quantized variant), QEIL v2 achieves 75.7% pass@k at 63.8W (IPW=0.9749), a 2.86x improvement over standard inference. When applied to a 4-bit Llama-3.1-8B, QEIL v2's physics-grounded routing achieves IPW=1.024 at 54.8W -- the first edge orchestration system to surpass the IPW=1.0 empirical reference mark, with the gain attributable entirely to QEIL v2's workload-adaptive device allocation on a model with reduced memory bandwidth requirements. Total energy drops 75.6% vs. standard with 38.3% latency reduction, zero thermal throttling, and 100% fault recovery across all benchmarks and model families.