ADJUSTING PAGE SHARING SCAN RATES BASED ON ESTIMATION OF PAGE SHARING OPPORTUNITIES WITHIN LARGE PAGES

摘要

Memory performance in a computer system that implements large page mappings is improved by dynamically tuning the page scan rate at which a memory sharing module (e.g., in a hypervisor) performs small page scanning operations that identify and exploit potential small page sharing opportunities within large pages. In operation, when free memory is relatively low, the hypervisor adjusts the page scan rate based on a statistical estimate of the percentage of virtual small pages that are mapped to physical large pages that are shareable. In this fashion the hypervisor dynamically tunes the sharing rate to reflect memory usage of applications. Further, unlike conventional approach to page sharing, the hypervisor proactively breaks large pages before resorting to more expensive memory reclamation techniques, such as ballooning and host swapping.

主权项

1. A method of determining a page scan rate of a memory that is partitioned and accessed as large physical pages and small physical pages to identify potential page sharing opportunities, comprising:
scanning at least one context for duplicated small virtual pages at a configurable page scan rate, wherein each of the at least one context has a virtual memory divided into a plurality of small virtual pages that are mappable to the physical pages; statistically estimating a percentage of small virtual pages mapped to large physical pages that are shareable while performing small virtual page scanning operations at a first page scan rate; and determining that the amount of free physical memory has decreased below a high memory threshold and, in response thereto, determining a second page scan rate based on the percentage of small virtual pages mapped to large physical pages that are shareable and performing small virtual page scanning operations at the second page scan rate instead of the first page scan rate.