Abstract

In a system with volatile hardware components (like CPU registers, processor caches, etc.), simply issuing a store operation does not guarantee that the update is durably written to PM. For instance, when a store operation is issued by the CPU, modern processors perform the update to a write-back cache. The write-back cache then lazily drains the update to PM, for example, when a cache line conflict occurs. Any subsequent failure erases the volatile write-back cache (or generally, any volatile component) and the update is lost. To prevent data loss on failure, hardware systems must guarantee that the updates reach PM.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call