Skip to content

Two cores, one cache

A real-time task runs on a dedicated core. Another core, running unrelated work, shares the last-level cache. The real-time task's execution time becomes variable.

Why?