Comment by rpy

Does anyone understand why the arm64 instance types are showing roughly half the performance versus the comparable x64 instances in this? Other benchmarks I've seen put them within 20% of each other. Is this just a hyperthreading vs full core thing?