A memory swap operation comprises writing information about a process in which a page fault occurred, into a temporary memory using a processor of a host, copying a page in which the page fault occurred, from a memory device recognized as a swap memory into a main memory of the host, and after completing the copying of the page, resuming the process in which the page fault occurred, using the information about the process, written in the temporary memory.

A proposed prefetcher may operate at a cache level where accesses are conducted using physical addresses. The proposed prefetcher may include one or more prefetch engines. Similar to conventional prefetchers, a prefetch engines of the proposed prefetcher may train on access patterns of a memory page to predict future accesses and perform prefetches based on the training. But unlike the conventional prefetchers, the trained prefetch engine may be reused for prefetching even when a request for a new page is received without requiring the prefetch engine to be newly trained on the new page. This can lower access latencies and lower cumulative training time.

An apparatus (2) comprises processing circuitry (4) for performing data processing in response to instructions. The processing circuitry (4) supports a cache maintenance instruction (50) specifying a virtual page address (52) identifying a virtual page of a virtual address space. In response to the cache maintenance instruction, the processing circuitry (4) triggers at least one cache (18, 20, 22) to perform a cache maintenance operation on one or more cache lines for which a physical address of the data stored by the cache line is within a physical page that corresponds to the virtual page identified by the virtual page address provided by the cache maintenance instruction.

Embodiments are described for methods and systems for mapping virtual memory pages to physical memory pages by analyzing a sequence of memory-bound accesses to the virtual memory pages, determining a degree of contiguity between the accessed virtual memory pages, and mapping sets of the accessed virtual memory pages to respective single physical memory pages. Embodiments are also described for a method for increasing locality of memory accesses to DRAM in virtual memory systems by analyzing a pattern of virtual memory accesses to identify contiguity of accessed virtual memory pages, predicting contiguity of the accessed virtual memory pages based on the pattern, and mapping the identified and predicted contiguous virtual memory pages to respective single physical memory pages.

Memory controllers and related methods for implementing an address-based dynamic page close policy are described. An example method includes using a first tracker, having entries indexed by a hash of system memory addresses for a physical memory region, tracking page hits associated with the physical memory region to develop a historical record of page hits to the physical memory region. The method further includes using a second tracker, having entries indexed by a hash of system memory addresses for the physical memory region, tracking page hits to the physical memory region to develop a current record of page hits to the physical memory region. The method further includes predicting whether to close a page or keep the page open based on a comparison of the historical record of page hits to the physical memory region with the current record of page hits to the physical memory region.

Prefetching is permitted to cross from one physical memory page to another. More specifically, if a stream of access requests contains virtual addresses that map to more than one physical memory page, then prefetching can continue from a first physical memory page to a second physical memory page. The prefetching advantageously continues to the second physical memory page based on the confidence level and prefetch distance established while the first physical memory page was the target of the access requests.

Memory controllers and related methods for implementing an address-based dynamic page close policy are described. An example method includes using a first tracker, having entries indexed by a hash of system memory addresses for a physical memory region, tracking page hits associated with the physical memory region to develop a historical record of page hits to the physical memory region. The method further includes using a second tracker, having entries indexed by a hash of system memory addresses for the physical memory region, tracking page hits to the physical memory region to develop a current record of page hits to the physical memory region. The method further includes predicting whether to close a page or keep the page open based on a comparison of the historical record of page hits to the physical memory region with the current record of page hits to the physical memory region.