Abstract: The object of the present invention is to improve the execution of instructions using speculative operations in Superscalar or Very Long Instruction Word (VLIW) processors having multiple Arithmetic Logic Units (ALUs). More particularly, the invention relates to a system and method for using standard registers as shadow registers. The addresses of all standard registers are translated using a Relocation Table (RT) array. The addresses of registers used as shadow registers are translated another time using a Speculative Registers Table (SRT) array. At branch completion time, for the speculative operations that have previously been executed and correctly predicted, the Relocation Table (RT) is updated with the Speculative Registers Table (SRT) content. For the speculative operations that have previously been executed and incorrectly predicted, the Relocation Table (RT) remains unchanged.

Abstract: A data transmission network congestion control mechanism requires knowledge of the sequence of occurrence of two dates d1 and d2, respectively defined by times t1 and t2 provided by a wraparound timer as respective numbers A and B coded in a 2's-complement form. Relative date discrimination is implemented by dividing the wraparound timer period into four consecutive intervals, each defined by the two most significant bits of the timer count. The value of the most significant bits and the sign of A-B, are used to derive a one-bit "X" indicator, the binary value of which indicates which of the two dates d1 and d2 was first to occur.

Abstract: The present invention relates to an efficient point-to-point and multi-points routing system and method for programmable data communication adapters in packet switching nodes of high speed networks. The general principles of this efficiency are the following:First, data packets are never copied, only packet pointers are copied for each destination: Space in Buffer Memory is saved, the number of instructions is significantly reduced improving the packet throughput (number of packets per seconds that the adapter is able to transmit). and the routing is independant of the packets length.Second, no overhead is generated by the multi-points mechanism in the real time procedures: the underrun/overrun problems on the ouputs are reduced and the efficiency of the adapter in term data throughput (bits per second) is significantly improved.Third, each output is processed independently by means of interrupts: lines are managed in real time and lines of different speed or protocol can be supported in parallel.

Abstract: A high performance data packet buffering method and a programmable data communication adapter for high speed packet transmission networks are disclosed. The line adapter includes programmable processing means, for receiving and transmitting data packets of fixed or variable length. This system is characterized in that it comprisesmeans for buffering (132) said data packets,means for identifying said buffering means and said data packets in said buffering means,means for queueing (FIG. 15) in storing means (131) said identifying means in a single instruction,means for dequeueing (FIG. 16) from said storing (131) means said identifying means in another single instruction,means for releasing said buffering means,Each instruction comprises up to three operations executed in parallel by said processing means:an arithmetical and logical (ALU) operation on said identifying means,memory operation on said storing means, anda sequence operation.

Abstract: An elastic buffer is provided between two busses working with independent clocking. The buffer is implemented by a piece of RAM memory (37) partitioned into sectors (41), each of which contains successive memory addresses. Each sector (41), can be alternatively written and read, so that at a given moment, a sector in write mode and a sector in read mode may coexist. Each sector is controlled by a mark flag (MF), a set flag corresponding to a fully written sector, and a reset flag corresponding to a sector that has been read onto the destination bus. The mark flag of each sector is set, respectively reset, upon the event of a move in pointer, respectively move out pointer, reaching the next adjacent sector. For a given elastic buffer size, the size of the sectors (41) and the number of mark flags are adaptable to the specifications of the data flow between the origin and destination busses.

Abstract: A device which switches data processing from an active processor, about to fail, to a back-up processor includes a memory change detector which captures memory changes in the memory of the active processor and a mirroring control circuit which causes the memory changes when committed by establishing recovery point signals generated by the active processor to be dumped into the memory of the back-up processor so that the back-up processor resumes operation of the active processor from the last established recovery point.

Abstract: In a shared memory system, wherein several memory users MU wish access to a plurality of memory banks, a set of high level commands (CREATE, PUT, GET, RELEASE) is provided, to transfer data between a given memory user and the memory banks or another memory user. The high level commands sent by the memory users are built up by memory interfaces MI connected to the memory users, and transmitted through an interconnection network to Packet Memory Command Executors PMCE integrated into each memory bank. The high level commands work with data records identified by Logical Record Addresses (LRA) known by the memory users. During execution of the high level commands by the PMCE, the LRA are translated into physical addresses corresponding to physical address space in the memory banks. The physical address space is created dynamically and released upon need, through the Create or Release Commands.

Abstract: A non-locking queueing mechanism is described for transferring information from a sending unit to a receiving unit through a queue in which there is no interference between the independent units (sender and receiver) during enqueueing or dequeueing. The invention thus avoids any form of interlock or serialism. The mechanism includes a first pointer (D), identifying the element area in the queueing device where the last dequeued information element, if any, was located, and a second pointer register for logging a second pointer (E) identifying the element area in the queueing device where the last enqueued information element, if any, was located, a first control block activated by the sending unit to enqueue the information element into the queueing device and for updating the second pointer, and a second control block activated by the recieving unit to dequeue the information element from the queueing device and for updating the first pointer.

Abstract: A device for loading data in and reading data out of latch strings located in field replaceable units containing the circuitry of a data processing system realized in accordance with the Level-Scan Sensitive Design (LSSD) technique. Each field replaceable unit includes an addressing circuit. The addressing circuits are interconnected by a monitoring loop over which a configuration of address bits is serially transmitted by a control circuit. The data to be loaded and read out propagate in a data loop and are entered in a latch string under control of the addressing circuit.