How does dyld3 improve app launch performance compared to dyld2?

dyld3 uses prebuilt closure caches to avoid runtime symbol resolution work done in dyld2.

dyld2 resolves symbols at runtime by parsing Mach-O headers, binding symbols, and loading dependent libraries dynamically on every launch. dyld3 introduces a shared cache and closure model where most linking work is precomputed during app install or OS update. This reduces cold start time by shifting expensive parsing and symbol resolution offline.