Cache Miss, TLB Miss & False Sharing — The Performance Killers in 3 Minutes

Same code, different cache layout, 10× performance gap. This video runs three benchmarks back-to-back to show how cache misses, TLB misses, and false sharing each show up in perf counters — and why none of them are visible in a flame graph.

Part of the CoreTracer project: the kind of bottleneck that you can only debug after you know to look for it.

Related

• Blog post: cache miss / TLB miss / false sharing — what perf shows you
• GitHub: CoreTracer