摘要 |
A second cache (e.g., L2 cache) is enabled or disabled based at least in part on an utilization of a first cache (e.g., L1 cache). The utilization of the first cache may be interpreted as an estimation of the likely utilization of the second cache. The utilization may be estimated by monitoring, for example, the eviction rates of the first cache or other caches, by monitoring the proportion of ways of the first cache used by an instruction stream, or by monitoring the duration between certain instructions executed by the processor, such as the duration between suspend operations.
|