GPU-Accelerated Computational Infrastructure for Biomedically-Driven Applications - Project Summary/Abstract Computation and high-performance computing are essential to advancing biomedical research and healthcare applications, driving progress in areas such as wearable health technology, medical image analysis, and health informatics. These fields require substantial computational power to process and analyze large datasets from diverse sources, including sensor data, medical images, and electronic health records (EHR). The ability to utilize GPU acceleration for deep learning model development and real-time data processing is critical for addressing the growing complexity of data-driven healthcare applications. As such, the proposal is to acquire a dedicated GPU server to support biomedical research and educational efforts across two UNT colleges: the College of Engineering and the College of Information Science. The proposed Dell PowerEdge XE9680 Rack Server, equipped with 8 NVIDIA HGX H100 80GB SXM5 GPUs, will provide the necessary computational resources to enable cutting-edge biomedical research and enhance student education in health technologies. This GPU cluster will support research projects focused on AI-driven health monitoring systems, including wearable devices for diabetes management and hydration monitoring, as well as applications in medical image processing and natural language processing for health informatics. The server will be used in various graduate-level courses in both colleges, including those focused on medical image acquisition, processing, and AI applications in clinical settings. The GPU cluster will facilitate the development and deployment of complex machine learning models, including vision transformers (ViTs), convolutional neural networks (CNNs), and transformer-based models, for tasks such as disease detection, image classification, and biomedical data analysis. The requested GPU infrastructure is crucial for enabling high-performance computing in biomedical and health informatics research. It will enable students and researchers to process large datasets in real-time, accelerate the training of deep learning models, and support the integration of multimodal data streams from medical images and sensor data for comprehensive health monitoring solutions. The server will also enhance the hands-on learning experience for students by providing access to cutting-edge computational tools used in AI-driven healthcare systems. This will better prepare students for careers in health technology, research, and digital medicine. Given the loss of access to Talon 3.0, the proposed GPU server will fill the gap in computational resources, ensuring that UNT’s research teams and students have access to the necessary infrastructure to support innovative research and training in health technology and biomedical applications. The availability of this powerful system will continue to drive advancements in biomedical sciences, contributing to breakthroughs in areas such as skin cancer detection, kidney tissue analysis, and clinical decision support systems. The integration of this GPU-powered infrastructure will position UNT as a leader in AI and health informatics, providing critical support for interdisciplinary research and educational programs across the College of Engineering and the College of Information Science.
