Three level memory hierarchy using write and share flags
AssigneeInternational Business Machines Corporation
InventorsRobert P. Fletcher, David M. Stein, Irving Wladawsky-Berger
Life of this patent
Estimating the term…
Loading patent…
AssigneeInternational Business Machines Corporation
InventorsRobert P. Fletcher, David M. Stein, Irving Wladawsky-Berger
Estimating the term…
A multiprocessing three level memory hierarchy implementation is described which uses a "write" flag and a "share" flag per page of information stored in a level three main memory. These two flag bits are utilized to communicate from main memory at level three to private and shared caches at memory levels one and two how a given page of information is to be used. Essentially, pages which can be both written and shared are moved from main memory to the shared level two cache and then to the shared level one cache, with the processors executing from the shared level one cache. All other pages are moved from main memory to the private level two and level one caches of the requesting processor. Thus, a processor executes either from its private or shared level one cache. This allows several processors to share a level three common main memory without encountering cross interrogation overhead.
A method of operating a multiprocessor system comprised of at least first and second processors, a system control element and a three level memory hierarchy, including a main memory at memory level three, connected to said system control element, in which pages of information are stored at addressable locations, with each page of information being comprised of lines of information, including WRITE and SHARE flags which may be in one of two binary states, with the state of the WRITE flag determining if a page of information is writeable or read only by a processor, and the state of the SHARE flag determining if a page of information is shareable or non-shareable by all processors, with each flag being set in a predetermined state, at least first and second private caches at memory level two, connected to said systems control element, which are exclusively operative with said at least first and second processors, respectively, memory, at least third and fourth private caches at memory level one which are exclusively operative with said first and second processors, respectively, and are connected to said first and second private caches, respectively, and first and second shared caches at memory levels one and two, respectively, which are connected together, with the first shared cache being connected to said first and second processors and the second shared cache being connected to said system control element, with pages of information being stored in the caches at memory level two, and lines of information being stored in the caches at memory level one, with said information being accessed with respect to said addressable locations in said main memory, said method comprising the steps of:
In a multiprocessor system, including a three level memory hierarchy, the combination comprising:
In a multiprocessor system, including a three level memory hierarchy, the combination comprising:
In a multiprocessor system, including a three level memory hierarchy, the combination comprising:
In a multiprocessor system, including a three level memory hierarchy, the combination comprising:
Dependencies are inferred from the claim text.