This episode explores why GCC consistently outperformed Clang in a Game of Life simulation, which was unexpected. Through extensive testing with various optimization levels (O0, O1, O2, O3, Os, Og) and compiler versions, it became clear that O3 delivered the fastest code for both compilers, challenging some recent assertions. While GCC generally showed better performance, a simple change—switching from compactly sized indices to int64_t—dramatically boosted Clang's speed, allowing it to eclipse GCC. The cause of this notable performance improvement remains unclear, leading to further investigation into compiler optimizations and unexpected interactions between data type sizes and compiler behavior.
