Deep C/Clang compiler optimization internals — register allocation, instruction selection (ISel), loop-invariant code motion (LICM), vectorization failure triage, BOLT post-link optimization, and pass interaction. Use when -O3 fails to vectorize a hot loop, when diagnosing register pressure spills, when planning PGO/BOLT deployment, or when understanding why passes interact unexpectedly. For LLVM tooling (opt/llc/IR) → See `lobor/cc-skills-clang@clang-llvm`; for high-level optimization methodology → See `lobor/cc-skills-clang@clang-performance`.
