L3 Latency

Metric Description

This metric shows an aggregate fraction of cycles where CPU accessed L3 cache for all load requests, while there was no contention/sharing with a sibling core. Avoiding cache misses (L2 misses/L3 hits) improves the latency and increases performance.