摘要:
In some embodiments, the invention includes a snoop filter, wherein entries in the snoop filter are allocated in response to initial accesses of local cache lines by a remote node, but entries in the snoop filter are not allocated in response to accesses of the local cache lines by a local node. Other embodiments are described and claimed.
摘要:
A method and apparatus for cache coherency states is disclosed. In one embodiment, a cache accessible across two interfaces, an inner interface and an outer interface, may have a joint cache coherency state. The joint cache coherency state may have a first state for the inner interface and a second state for the outer interface, where the second state has higher privilege than the first state. In one embodiment this may promote speculative invalidation. In other embodiments this may reduce snoop transactions on the inner interface.
摘要:
An example embodiment of a computer system utilizing a central snoop filter includes several nodes coupled together via a switching device. Each of the nodes may include several processors and caches as well as a block of system memory. All traffic from one node to another takes place through the switching device. The switching device includes a snoop filter that tracks cache line coherency information for all caches in the computer system. The snoop filter has enough entries to track the tags and state information for all entries in all caches in all of the system's nodes. In addition to the tag and state information, the snoop filter stores information indicating which of the nodes has a copy of each cache line. The snoop filter serves in part to keep snoop transactions from being performed at nodes that do not contain a copy of the subject cache line, thereby reducing system overhead, reducing traffic across the system interconnect busses, and reducing the amount of time required to perform snoop transactions.
摘要:
In some embodiments, a computer system includes nodes connected through conductors. At least some of the nodes each include memory and processing circuitry to receive snoop requests in a node reception order and to initiate snoops of the memory in the node before the snoop requests are in a global order. The at least some nodes also each include an ordering buffer to receive the snoop requests and provide them at an output of the ordering buffer in the global order.
摘要:
Methods, techniques, systems and apparatuses for utilizing reserved space for error correcting functionality. A cache line (“reserved line”) in a plurality of cache lines to store error correcting code (ECC) data is utilized for storing ECC data corresponding to other cache lines within the plurality of cache lines when a memory device has failed.
摘要:
A semiconductor device and a method of manufacturing the device is disclosed. In one aspect, a method includes providing a substrate, providing a first epitaxial semiconducting layer on top of the substrate, and forming a one- or two-dimensional repetitive pattern, each part of the pattern having an aspect ratio in the range of about 0.1 to 50.
摘要:
Disclosed is an apparatus and a method to inject errors to a memory. In one embodiment, a dedicated interface includes an error injection system address register and an error injection mask register coupled to the error injection system address register. If the error injection system address register includes a system address that matches an incoming write address, the error injection mask register outputs an error to the memory.
摘要:
A semiconductor device and method of manufacturing the device is disclosed. In one aspect, the device includes a semiconductor substrate and a GaN-type layer stack on top of the semiconductor substrate. The GaN-type layer stack has at least one buffer layer, a first active layer and a second active layer. Active device regions are definable at an interface of the first and second active layer. The semiconductor substrate is present on an insulating layer and is patterned to define trenches according to a predefined pattern, which includes at least one trench underlying the active device region. The trenches extend from the insulating layer into at least one buffer layer of the GaN-type layer stack and are overgrown within the at least one buffer layer, so as to obtain that the first and the second active layer are continuous at least within the active device regions.
摘要:
Embodiments of the present disclosure describe methods, apparatus, and system configurations for providing rank-specific cyclic redundancy checks in memory systems.
摘要:
A system (100) is described for characterizing and/or manipulating molecules. The system may especially be suitable for biological molecules, although the invention is not limited thereto. The system (100) comprises a substrate (110) comprising a nanostructure (120) being suitable for translocation of molecules through the nanostructure (120). It furthermore comprises a means (210) for translocating molecules through the nanostructure (120) and a plasmonic force field generating means (130) adapted for influencing the translocation speed of the particle by applying a plasmonic force field at the nanostructure (120). A corresponding method also is described.