摘要:
To prevent a decrease in performance of controlling a snoop tag. A queue is stored with REPLACE target WAY information and an index as an entry associated with a REPLACE request received from a processor, the index stored in the queue is compared with an index of a subsequent READ request, and, as a result of the comparison, a process based on the index-coincident READ request is executed with respect to the snoop tag corresponding to a content of a cache memory of the processor. Further, the REPLACE target WAY information of the READ request is replaced with the WAY information in the index-coincident entry within the queue.
摘要:
A northbridge, when detecting a synchronization break of a redundant CPU, stops the operation of an abnormal CPU bus where an error has occurred and the firmware in an FWH instructs the northbridge to inhibit an external instruction. In addition, the firmware save the inside information of a normal CPU connected to a normal CPU bus and cache data on a memory and the northbridge issues reset to all CPUs in the home system board. The firmware then restores the inside information of the CPU save on the memory to the all CPUs and instructs the northbridge to cancel the inhibition of the external instruction.
摘要:
In a system controller including a CPU-issued request queue having a circuit that processes plural requests having identical addresses not to be inputted to the CPU-issued request queue, a latest request other than a cache replace request is retained by an input-request retaining section. Consequently, even if an address of an issued request for cache replace request matches an address of a request retained by the CPU-issued request queue, the issued request for the cache replace request is not retried but is queued in the CPU-issued request queue when the address of the issued request for the cache replace request does not match the entire address retained by the input-request retaining section.
摘要:
In a multiprocessor system, a system controller includes snoop tags which are copy information on cache tags retained by respective CPUs. If the same address is registered in S (Shared state) in the cache tag of each of the CPUs connected to the same CPU bus, the address is registered in S (Shared state) in only any one of the snoop tags corresponding to the CPUs in which the same address is registered.
摘要:
An information processing device having two processing units capable of operating in synchronization with each other, includes: a common unit capable of outputting an identical signal to the two processing units; detection units that are respectively provided for the processing units and each detects errors occurred in corresponding processing unit respectively; a comparison unit that compares outputs from the two processing units; and a control unit that controls signals from the processing units to the common unit, based on a detection result of the detection units and a comparison result of the comparison unit, and determines, if errors of an identical type are simultaneously detected by the detection units, that the errors are due to an error of the common unit.
摘要:
A multiprocessor system includes a judging unit judging whether a read command inputted to a global address crossbar is a read command to a memory on an own system board, an executing unit speculatively executing, when the judging unit judges that the read command is a read command to the memory on the own system board, the read command before global access based on an address notified from the global address crossbar, a setting unit setting for queuing data read from the memory in a data queue provided on a CPU without queuing the data in a data queue provided on the memory, and an instructing unit instructing, based on notification from the global address crossbar, the data queue provided on the CPU to discard the data or transmit the data to the CPU.
摘要:
A multiprocessor system includes a judging unit judging whether a read command inputted to a global address crossbar is a read command to a memory on an own system board, an executing unit speculatively executing, when the judging unit judges that the read command is a read command to the memory on the own system board, the read command before global access based on an address notified from the global address crossbar, a setting unit setting for queuing data read from the memory in a data queue provided on a CPU without queuing the data in a data queue provided on the memory, and an instructing unit instructing, based on notification from the global address crossbar, the data queue provided on the CPU to discard the data or transmit the data to the CPU.
摘要:
A control device managing a plurality of nodes transmitting and receiving data containing an error correcting code, comprises means accepting, when any one of the nodes detects an uncorrectable error from the data containing the error correcting code, a signal transmitted by the node detecting the error, means judging from a record of the detection of a first node, when accepting the signal from a second node receiving data transmitted by the first node, whether or not the first node has detected the uncorrectable error from the data transmitted to the second node, and means stopping, when the first node has detected the uncorrectable error from the data transmitted to the second node, a process attributed to the acceptance of the signal from the second node.
摘要:
In a system controller including a CPU-issued request queue having a circuit that processes plural requests having identical addresses not to be inputted to the CPU-issued request queue, a latest request other than a cache replace request is retained by an input-request retaining section. Consequently, even if an address of an issued request for cache replace request matches an address of a request retained by the CPU-issued request queue, the issued request for the cache replace request is not retried but is queued in the CPU-issued request queue when the address of the issued request for the cache replace request does not match the entire address retained by the input-request retaining section.
摘要:
A request issued by the CPU is output from the local arbiter by way of the CPU bus and the CPU-issued request queue. The cache replacement request loop-back circuit determines at the loop-back determination circuit whether the outputted request is a cache replacement request or not. A request other than a cache replacement request is output onto the local bus. A cache replacement request is output to the selector and sent to the request handling section when there is no valid request on the global bus.