Global maintenance command protocol in a cache coherent system

摘要

A system may include a command queue controller coupled to a number of clusters of cores, where each cluster includes a cache shared amongst the cores. An originating core of one of the clusters may detect a global maintenance command and send the global maintenance command to the command queue controller. The command queue controller may broadcast the global maintenance command to the clusters including the originating core's cluster. Each of the cores of the clusters may execute the global maintenance command. Each cluster may send an acknowledgement to the command queue controller upon completed execution of the global maintenance command by each core of the cluster. The command queue controller may also send, upon receiving an acknowledgement from each cluster, a final acknowledgement to the originating core's cluster.

主权项

1. A system comprising:
a plurality of clusters, wherein each cluster includes a plurality of cores, and wherein each cluster includes a respective cache shared amongst the plurality of cores of the cluster; and a command queue controller, wherein the command queue controller is coupled to each cluster of the plurality of clusters, wherein each core of the plurality of cores of each of the plurality of clusters is configured to operate as an originating core by: detecting a global maintenance command; and sending, to the command queue controller, the global maintenance command; wherein the command queue controller is configured to broadcast the global maintenance command to one or more of the plurality of clusters including the originating core's cluster; wherein each core of the plurality of cores of each of the plurality of clusters receiving the broadcast is configured to execute the global maintenance command; wherein each cluster receiving the broadcast is configured to send an acknowledgement to the command queue controller upon completed execution of the global maintenance command by each core of each cluster receiving the broadcast; and wherein the command queue controller is further configured to send, upon receiving the acknowledgement from each cluster receiving the broadcast, a final acknowledgement to the originating core's cluster.