Intermediate Representations
Analysis of Sea-of-Nodes, MLIR dialects, and polyhedral optimization frameworks for high-performance computing.
Register Allocation
Linear scan vs graph coloring allocators in modern JIT engines, V8 TurboFan, and LLVM backend pipelines.
Instruction Scheduling
Pipelining techniques, superscalar scheduling, and microarchitectural hazard mitigation algorithms.
Garbage Collection
State-of-the-art concurrent generational collectors, ZGC / Shenandoah compaction, and pauseless runtime design.