Graphs are a fundamental representation for a wide range of real-world systems, including social networks Nettleton [2013], molecular structures Trinajs-tic [2018], knowledge graphs Hogan et al. [2021], and recommender systems Wu et al. [2022], all of which naturally take the form of nodes connected by edges. Geometric learning, which generalizes deep learning to non-Euclidean domains Bronstein et al. [2017], especially on graph-structured data Xia et al. [2021], Wu et al. [2020], has emerged as a powerful paradigm for capturing relational and structural information. Graph neural network (GNN) models such as 1 graph convolutional networks (GCNs) Kipf and Welling [2017], graph attention networks (GATs) Veliˇ ckovi c et al. [2018], Brody et al. [2022], Lee et al. [2023], and GraphSAGE Hamilton et al. [2017] have demonstrated success across tasks from node classification Xiao et al. [2022] to link prediction Zhang and Chen [2018] and recommendation Lee et al. [2024]. While Nvidia's CUDA-enabled graphics processing units (GPUs) have become the de facto standard for accelerating deep learning workloads, there is a growing ecosystem of alternative processing units that offer different trade-offs in cost, power efficiency, and architectural specialization Lee et al. [2025]. Non-CUDA processing units, such as Google's tensor processing unit (TPUs) and Intel's Gaudi HPUs, are gaining traction in both industry and research. Supporting geometric learning workloads on these non-CUDA hardware platforms is essential to broaden hardware choice, optimize for diverse deployment scenarios, and leverage specialized hardware capabilities. Intel's Gaudi HPUs exemplify non-CUDA accelerators tailored for deep learning.

With recently available football match event data that record the details of football matches, analysts and researchers have a great opportunity to develop new performance metrics, gain insight, and evaluate key performance. However, most sports sequential events modeling methods and performance metrics approaches could be incomprehensive in dealing with such large-scale spatiotemporal data (in particular, temporal process), thereby necessitating a more comprehensive spatiotemporal model and a holistic performance metric. To this end, we proposed the Transformer-Based Neural Marked Spatio Temporal Point Process (NM-STPP) model for football event data based on the neural temporal point processes (NTPP) framework. In the experiments, our model outperformed the prediction performance of the baseline models. Furthermore, we proposed the holistic possession utilization score (HPUS) metric for a more comprehensive football possession analysis. For verification, we examined the relationship with football teams' final ranking, average goal score, and average xG over a season. It was observed that the average HPUS showed significant correlations regardless of not using goal and details of shot information. Furthermore, we show HPUS examples in analyzing possessions, matches, and between matches.