摘要:
This invention provides dynamic balance of the traffic among data processing devices interconnecting networks and thereby improve the networking performance. For network traffic flowing between a first network and a second network, the traffic is distributed among the data processing devices that act as routers according to the traffic amount. An algorithm for balancing the traffic is used to select appropriate data processing devices as routers.
摘要:
A processor (1) for a multiprocessor system, such as a parallel processor system, connected to a network (2) has a sending unit (11) and a receiving unit (12) for transferring and receiving data to and from the network (2) as well as receive cache (22) and main cache. When data are received from the network (2), it is determined whether a hit or miss occurs to the main cache (21) and receive cache (22), respectively. If a hit to the receive cache (22) occurs, then the receive cache controller (14) stores the data directly in the receive cache (22) as it is received. When a hit to the main cache (21) occurs, an intercache transfer is executed for transferring the hit block in the main cache (21) to the receive cache (22) so that the data can be stored in the receive cache (22). When an instruction processor (10) requests access to data held in the receive cache (22), the data are retrieved to the instruction processor (10) and at the same time transferred to the main cache (21).
摘要:
This invention provides dynamic balance of the traffic among data processing devices interconnecting networks and thereby improve the networking performance. For network traffic flowing between a first network and a second network, the traffic is distributed among the data processing devices that act as routers according to the traffic amount. An algorithm for balancing the traffic is used to select appropriate data processing devices as routers.
摘要:
To reduce the overhead of the interrupt on a processor (103) associated with packet send and receive control in a network (102), packet send command chaining means is provided. Based on the control field in each packet send command, a send node controls an interrupt request to the processor (103) in the packet level and sends a packet set with the control information to a receive node. Based on the control field in the received data packet, the receive node controls a receive circuit interrupt request, thereby reducing the number of times the interrupt on the instruction processor (103) is caused for each packet send and receive operation.