Combinational Equivalence Checking (CEC) has become an increasingly critical and active research topic in modern Electronic Design Automation (EDA) flows, as it plays a key role in guaranteeing functional correctness for large-scale and highly optimized designs. However, the Exact Probability-based Simulation (EPS) phase in CEC is extremely computation-intensive and has emerged as the dominant performance bottleneck, suffering from massive data dependencies, irregular memory access patterns, and low hardware utilization on conventional platforms. To address these challenges, we propose a dataflow-driven GPU acceleration framework for EPS, featuring a dataflow-driven execution model and architecture-aware optimizations. First, we reformulate the EPS process as a dataflow-driven computational graph, enabling fully parallel and asynchronous signature propagation on GPU architectures. Second, we design a loop-unrolled CUDA kernel integrated with a lightweight shared-memory LRU caching mechanism to improve data locality and reduce redundant memory accesses. Third, we introduce a hierarchical memory coalescing and constant-broadcast strategy to enhance memory alignment, bandwidth efficiency, and inter-thread data reuse. Experimental results show that the proposed framework achieves up to 12.9 & times; speedup over optimized CPU-based solutions and 3.5 & times; improvement compared to existing GPU-based approaches.
The environmental impact of electricity consumption is commonly assessed through its carbon footprint (CF), while water-related impacts are often overlooked despite the strong interdependence between energy and water systems. This is particularly relevant for electricity-intensive activities such as data centre (DC) operations, where both carbon emissions and water use occur largely off-site through electricity consumption. In this work, we present Wattnet, an open-source tool that jointly assesses the CF and water footprint (WF) of electricity consumption across Europe with high temporal resolution. Wattnet implements an electricity flow-tracing methodology that accounts for local generation mixes, as well as for cross-border electricity imports and exports at a 15 min resolution. Operational and life-cycle impact factors are used to quantify and compare local (generation-based) and global (consumption-based) footprints for multiple European regions during 2024. Wattnet includes a 72 h forecasting module to facilitate informed energy-aware decision-making. The results demonstrate that neglecting electricity flows and temporal variability can lead to significant misestimations of both CF and WF, particularly in countries with high levels of electricity trade or hydropower dependence. Furthermore, the joint analysis reveals trade-offs between decarbonization and water use, highlighting the prominent role of reservoir-based hydropower in increasing WF even in low-carbon systems. Wattnet facilitates informed energy-aware decision-making, while also enhancing transparency regarding the environmental impacts of electricity consumption for end users and policymakers.
Smart urban building management systems require decision support under high-dimensional, heterogeneous indicators that depend on weather conditions, energy consumption patterns, and occupant behavior, while reliable ground-truth labels for optimal interventions are typically unavailable. Conventional soft set and multi-attribute decision frameworks assume parameter independence and rely on correlation-or variance-based dimensionality reduction, which can distort decisions when prerequisite relationships exist among building energy indicators. This paper presents a dependency-aware soft-set framework that explicitly models prerequisite-type relationships among energy parameters and enforces them through an order-preserving, non-inflating dependency propagation operator. The framework extends classical soft sets by integrating a directed dependency graph and dependency functions, ensuring logical consistency and convergence to a unique dependency-consistent closure for each initial membership matrix. It further enables decision-preserving redundancy reduction, in which parameters are removed only when their omission leaves dependency-consistent memberships, rankings, and final decisions unchanged. The framework is evaluated on 300 commercial buildings using 50 indicators derived from electricity consumption and outdoor temperature data, grouped into thermal, behavioral, and flexibility-related dimensions, and validated through repeated out-of-sample train-test splits in which dependency graphs learned on training buildings are applied unchanged to held-out test buildings. Dependency relations are acquired directly from data through a combined criterion of statistical association and prerequisite-violation constraints, without expert-defined assumptions. Results show stable, interpretable, and dependency-consistent building-energy prioritization with substantial dimensionality reduction and no loss of decision integrity, providing an interpretable portfolio-screening tool for building-energy decision making under uncertainty and limited data.
