Zero-Overhead Compiler Pessimization
summary
Shows that simple microbenchmarks vanish under optimization: the optimizer removes the increment and leaves only timing calls unless the code forces observable effects. Uses Compiler Explorer, standard clock calls and benchmark tricks from Google Facebook Celero as examples, and demonstrates adding a fake read/call so the increment survives between two calls to clock time. Result is concrete: getting reliable microbenchmarks is tricky and requires careful fences.
AI-written summary · about
related talks
01h 03mHidden Overhead of a Function API
01h 00mJust-in-Time Compilation
01h 50mUnderstanding Compiler Optimization
08mOn "simple" Optimizations
48m[a bit noisy] Optimizing Generic Algorithms
01h 04mWhat Can Compiler Benchmarks Reveal About Metaprogramming Implementation Strategies
01h 05mThe Most Important Optimizations to Apply in Your C++ Programs
01h 04mCPU design effects