摘要 |
The specification and drawings present a new apparatus, method and software related product for using a cache/central cache module/device (instead of e.g., system DRAM) which can serve multiple memory modules/devices. Each memory/IO module/device connected to the same memory network (e.g., via hub, bus, etc.) may utilize memory resources of this cache module/device either in a fixed manner using pre-set allocation of resources per the memory module/device, or dynamically using run-time allocation of new resources to an existing module/device per its request or to a new module/device connecting to the memory network (e.g., comprised in a host device) and possibly requesting memory resources. |
主权项 |
1. A method, comprising:
reserving, by at least one cache module of a host device, one or more memory resource allocations of memory resources in the at least one cache module, the one or more memory resource allocations being reserved for one or more memory modules; providing, by the at least one cache module, an identification of the at least one cache module to at least one memory module; receiving, by the at least one cache module and from the at least one memory module, a request to use at least part of the memory resources in the at least one cache module; and implementing, by the at least one cache module, the request using at least one memory resource of the memory resources in the at least one cache module, wherein the at least one memory resource of the memory resources in the at least one cache module is:
previously reserved for the at least one memory module based at least in part on the one or more memory resource allocations, ordynamically identified by the at least one cache module. |