Iterating Pointers - Enabling Static Analysis for Loop-based Pointers

Learn about groundbreaking research in pointer analysis and optimization through this technical presentation from ETH Zurich's Scalable Parallel Computing Lab. Explore an innovative approach to analyzing and optimizing pointer-based code by separating pointers into data containers and offsets, enabling more effective static analysis and compiler optimizations. Discover how this methodology successfully improves performance in real-world applications, including OpenSSL and the Mantevo benchmark suite, achieving up to 11x speedup in the PBKDF2 algorithm and outperforming reference implementations by 18%. Presented by Andrea Lepori at ACM Transactions on Architecture and Code Optimization, the talk demonstrates practical solutions for handling pointer-based code analysis challenges in C programming, ultimately enhancing program optimization and auto-parallelization capabilities.