N64 RSP Assembly Optimization in F3DEX3 Custom Microcode

You asked for this! I go over the basics of optimizing microcode assembly for the N64's RSP, and how I achieved two significant performance optimizations compared to Nintendo's F3DEX2: saving 10 cycles on small RDP commands, and 7 cycles on triangles. F3DEX3 is in beta: https://github.com/HackerN64/F3DEX3