Abstract: A stream processing computer architecture includes creating a stream computer processing (SCP) system by forming a super node cluster of processors representing physical computation nodes (“nodes”), communicatively coupling the processors via a local interconnection means (“interconnect”), and communicatively coupling the cluster to an optical circuit switch (OCS), via optical external links (“links”). The OCS is communicatively coupled to another cluster of processors via the links. The method also includes generating a stream computation graph including kernels and data streams, and mapping the graph to the SCP system, which includes assigning the kernels to the clusters and respective nodes, assigning data stream traffic between the kernels to the interconnection when the data stream is between nodes in the same cluster, and assigning traffic between the kernels to the links when the data stream is between nodes in different clusters.

Abstract: A method for implementing a stream processing computer architecture includes creating a stream computer processing (SCP) system by forming a super node cluster of processors representing physical computation nodes (“nodes”), communicatively coupling the processors via a local interconnection means (“interconnect”), and communicatively coupling the cluster to an optical circuit switch (OCS), via optical external links (“links”). The OCS is communicatively coupled to another cluster of processors via the links. The method also includes generating a stream computation graph including kernels and data streams, and mapping the graph to the SCP system, which includes assigning the kernels to the clusters and respective nodes, assigning data stream traffic between the kernels to the interconnection when the data stream is between nodes in the same cluster, and assigning traffic between the kernels to the links when the data stream is between nodes in different clusters.

Abstract: A method for implementing a stream processing computer architecture includes creating a stream computer processing (SCP) system by forming a super node cluster of processors representing physical computation nodes (“nodes”), communicatively coupling the processors via a local interconnection means (“interconnect”), and communicatively coupling the cluster to an optical circuit switch (OCS), via optical external links (“links”). The OCS is communicatively coupled to another cluster of processors via the links. The method also includes generating a stream computation graph including kernels and data streams, and mapping the graph to the SCP system, which includes assigning the kernels to the clusters and respective nodes, assigning data stream traffic between the kernels to the interconnection when the data stream is between nodes in the same cluster, and assigning traffic between the kernels to the links when the data stream is between nodes in different clusters.

Abstract: A cascaded interconnect system including a memory controller, one or more memory modules, an upstream memory bus and a downstream memory bus. The one or more memory modules include a first memory module with cache data. The memory modules and the memory controller are interconnected by a packetized multi-transfer interface via the downstream memory bus and the upstream memory bus. The first memory module and the memory controller are in direct communication via the upstream memory bus and the downstream memory bus.

Abstract: A cascaded interconnect system including a memory controller, one or more memory modules, an upstream memory bus and a downstream memory bus. The one or more memory modules include a first memory module with cache data. The memory modules and the memory controller are interconnected by a packetized multi-transfer interface via the downstream memory bus and the upstream memory bus. The first memory module and the memory controller are in direct communication via the upstream memory bus and the downstream memory bus.

Abstract: A cascaded interconnect system including a memory controller, one or more memory modules, an upstream memory bus and a downstream memory bus. The one or more memory modules include a first memory module with cache data. The memory modules and the memory controller are interconnected by a packetized multi-transfer interface via the downstream memory bus and the upstream memory bus. The first memory module and the memory controller are in direct communication via the upstream memory bus and the downstream memory bus.

Abstract: A large, fault tolerant, highly reliable semiconductor data storage system (memory) is designed to have the memory function striped across multiple symbol planes which comprise individual fault containment regions. Each fault containment region includes such a symbol plane which, in turn, stores at least one bit of any given memory word accessed in the system. The system further includes a processing core module, including at least symbol plane addressing controls, and a channel adapter is provided for selectively connecting the memory to high speed communications channels for, in turn, communicating with client processors or other functional entities attached to the data store system. The processing core contains an error correction/detection mechanism for the error checking and correction of all data fetched from the memory and for generating error correction and detection code bits for all data to be stored in memory.

Abstract: A quad oscillator fault-tolerant clock system for a computer complex comprises two clock sources at each of two computer locations, which are coupled by two duplex links. Each clock source supplies its own clock signal to the other clock source at the same location as well as to the clock sources at the other location over one of the duplex links coupling the two locations. Each clock source continually measures the phase difference between its clock signal and each of the other three clock signals. Periodically, the propagation delay for each link is calculated by taking the average of the phase differences measured by the clock sources driving the two ends of that link. These calculated propagation delays are supplied to each individual clock source, which corrects the phase differences measured by it for the propagation delays.

Abstract: A system for providing fault-tolerant synchronized operation of the Time of Day (TOD) clocks of the respective data processors in a multiprocessor complex. Basically, the system is comprised of a duplex implementation having redundant TOD clock sources, and a plurality of TOD slaves which provide the TOD clocks in the associated processors. A register/counter in each TOD clock source is incremented by a high frequency signal to achieve the required TOD value resolution, and the latter signal is divided down to provide a lower reference frequency signal for synchronization of the clock sources. Each TOD slave includes terminals for receiving a pair of reference frequency signals and for trouble-free switching between the signals, as required. Alternatively, a quad implementation of clock sources which is substantially free of single points of failure of the synchronization mechanism is described. Frequency steering of the clock sources provides increased accuracy and conformity to real time when desired.