The OSCAR intelligent compilation optimization system, independently crafted by the "Dual-Engine Intelligent Translation" research team at Dalian University of Technology, has effectively tackled the core challenges in compilation optimization for domestic chips. This breakthrough substantially boosts the high-performance compilation capabilities of domestic processors. The system creatively combines large model adaptive optimization, intelligent swarm search algorithms, and a multi-level intelligent decision-making mechanism. Together, they form a multi-granularity system tuning framework that significantly enhances the efficiency of harnessing the underlying computing power of domestic chips. Under typical computational scenarios, the OSCAR system exhibits remarkable performance improvements compared to traditional compilation baselines. It intelligently navigates optimization spaces and crafts customized compilation solutions for diverse application scenarios. Capitalizing on this achievement, the team secured the Special Prize in the "Solution to Key Challenges" category at the 19th "Challenge Cup" China Youth Science and Technology Innovation Competition, standing out as the sole undergraduate team to be honored with this accolade. Presently, the team has established a company focused on ongoing technological refinement and industrial deployment. Looking ahead, they aim to adapt and broaden the technology to encompass more domestic processor architectures, while further deepening the integration and innovation of AI and compilation technologies.