This project aims to develop next-generation programming language abstractions and tools that will significantly simplify and enhance GPU programming. GPUs have become essential in various fields such as scientific computing, machine learning, and computer graphics. Unfortunately, the complexity of current programming language models often poses a barrier to efficient usage.

Traditional programming language models are not designed to optimally harness the massive parallel processing capabilities of GPUs. Therefore, there is an urgent need for new abstractions that allow programmers to work at a higher level, reducing the learning curve and development times. Our project focuses on designing language constructs and tools that make complex parallel algorithms comprehensible while ensuring maximum performance and efficiency.