오윤호 연구실

CoreCraft focuses on revolutionizing the design and functionality of modern processors to address the increasing demands of high-performance and parallel computing. This area explores innovative techniques to enhance computational efficiency, and improve overall processor performance. Central to this research is the development of advanced compiler techniques and hierarchical register file management that streamline data access and storage. These innovations reduce conflicts and inefficiencies, allowing GPUs and CPUs to process data more swiftly and effectively. Our CoreCraft aims to create more powerful, efficient, and versatile processors that can meet the demands of modern computing tasks, from deep learning to complex simulations. By integrating these cutting-edge advancements, this research is paving the way for the next generation of high-performance computing technologies. VitesseAI research seeks to push the boundaries of AI performance, enabling faster and more efficient AI solutions. This field is essential for advancing AI technologies in various industries, from autonomous systems to big data analytics, where high-speed and high-efficiency computation is paramount. VitesseAI focuses on developing innovative methods and architectures to enhance the performance of AI workloads. This domain aims to design both hardware and software to achieve maximum efficiency and speed in AI processing. Our key innovations include designing new software for GPUs or new GPU architectures to handle AI tasks effectively, ensuring that AI workloads run concurrently with minimal resource contention, and designing scalable AI accelerator architectures. Prof. Yunho Oh completed the research visit at KTH through the Digital Futures Programme. Prof. Yunho Oh has given a talk at EPFL,