摘要:
Methods and apparatus provide for: receiving tokens from a plurality of resources, where the tokens are exchange media for permitting a plurality of requesters to access an associated one of the resources for a period of time; allocating the tokens to respective requester groups into which the plurality of requesters have been assigned; receiving requests for one or more of the tokens from one or more of the requesters and granting, from the allocated tokens, one or more of the tokens to the requesters based on token granting criteria; and re-distributing unused tokens from one or more requester groups to one or more other requester groups based on priorities assigned to the requester groups; and dynamically altering the priorities assigned to the requester groups when a number of unused tokens of one or more of the requester groups has fallen below a threshold.
摘要:
Methods and apparatus provide for: assigning each of a plurality of requesters to a respective one of a plurality of requester groups; receiving tokens from a plurality of resources, where each token is an exchange medium for permitting one of the requesters having the token to access an associated one of the resources for a period of time; receiving requests for the tokens from one or more of the requesters; allocating the tokens to at least one of the respective requester groups and the requesters thereof based on token allocation criteria; and dynamically re-assigning one or more of the requesters among the requester groups based on feedback information concerning at least some prior token allocations.
摘要:
Methods and apparatus provide for: receiving tokens from a plurality of resources, where the tokens are exchange media for permitting a plurality of requesters to access an associated one of the resources for a period of time; allocating the tokens to respective requester groups into which the plurality of requesters have been assigned; receiving requests for one or more of the tokens from one or more of the requesters and granting, from the allocated tokens, one or more of the tokens to the requesters based on token granting criteria; and re-distributing unused tokens from one or more requester groups to one or more other requester groups based on priorities assigned to the requester groups; and dynamically altering the priorities assigned to the requester groups when a number of unused tokens of one or more of the requester groups has fallen below a threshold.
摘要:
Processors arranged in a multi-processor configuration for substantially parallel operations receive their initialization data in order to start operations, such as graphics computations, real-time multimedia streaming, etc. Due to a change in the processing load, one or more processors might be deactivated. Subsequently, the load increases to such a level that requires all or some of the deactivated processors to be active again. In this case, the boot-up process of the entire system is not carried out as it would be time-consuming and wasteful; instead, responsive to a control signal only those processors that were previously in inactive mode are re-initialized by selecting a configuration data supplied by another processor, controller or any other intelligent programmable device. Alternatively, the configuration data may be accessed and retrieved from a local storage medium individually located in each processor, thereby re-booting only those inactive processors and without re-initializing the entire system.
摘要:
Methods and apparatus provide for dynamically determining unit-AC paths between a plurality of processing units and a plurality of address concentrators. The unit-AC paths may be determined by configuring a plurality of selector settings of a plurality of selector circuits. The plurality of selector circuits, the plurality of selector settings and the plurality of address concentrators combine to enable a plurality of possible AC connections. The selector settings may be configured in accordance with a prioritization of the units and/or the unit-AC paths in view of a ranking of the plurality of possible AC connections. The unit-AC paths may be adjusted by reconfiguring the selector settings, such as when a priority mismatch is identified between the prioritization of the units and the ranking of AC connections. The methods and apparatus may also include updating one or more of the address concentrators regarding the current prioritization and/or selector settings in the system.
摘要:
Processors arranged in a multi-processor configuration for substantially parallel operations receive their initialization data in order to start operations, such as graphics computations, real-time multimedia streaming, etc. Due to a change in the processing load, one or more processors might be deactivated. Subsequently, the load increases to such a level that requires all or some of the deactivated processors to be active again. In this case, the boot-up process of the entire system is not carried out as it would be time-consuming and wasteful; instead, responsive to a control signal only those processors that were previously in inactive mode are re-initialized by selecting a configuration data supplied by another processor, controller or any other intelligent programmable device. Alternatively, the configuration data may be accessed and retrieved from a local storage medium individually located in each processor, thereby re-booting only those inactive processors and without re-initializing the entire system.
摘要:
Methods and apparatus provide for: assigning each of a plurality of requesters to a respective one of a plurality of requester groups; receiving tokens from a plurality of resources, where each token is an exchange medium for permitting one of the requesters having the token to access an associated one of the resources for a period of time; receiving requests for the tokens from one or more of the requesters; allocating the tokens to at least one of the respective requester groups and the requesters thereof based on token allocation criteria; and dynamically re-assigning one or more of the requesters among the requester groups based on feedback information concerning at least some prior token allocations.