when I run pods in kubernetes do I lose cache efficiency since it is constantly evicting the pod from the pipeline in a fair time scheduler (therefore constantly evicting and moving in cache tables)? Does the CPU retain some cache locality throughout the process?

e.g. Can containers retain some cache locality or for security reasons is this impossible (all containers are calling the same .so in separate containers but these containers have the same base image where the .so exists-- could that stay in cache across the scheduler)?