COMPUTING SYSTEM HAVING FAULT CONTAINMENT

摘要

A multi-processor, shared memory computing system is disclosed. The system comprises a plurality of processing element modules, input/output processor modu les and shared memory modules interconnected with the processing elements and input/outp ut processors. The modules are interconnected by point to multi-point communication links. Each module has one transmit port and a plurality of receive ports interconnecte d through the links with transmit ports of other modules. Shared memory is updated and rea d by exchanging frames forming memory access transactions over these links. Between f rames, idle codes are broadcast across the links. In response to locally sensing a fail ure at a module, a failed module modifies its idle code to reflect the failure. Interconn ected modules lock-out a failed module in response to sensing an inhibit-on idle code at the receive ports. This facilitates the isolation and containment of faulty modules.