Abstract: The invention comprises a write buffer system which provides residence time information that increases the merge potential and enhances the write collapse feature of a "smart" buffer. The write buffer has multiple buffer locations for storing data received from a central processor, a data merger for merging data designated by the central processor for subsequent storage in contiguous memory locations, and a write controller for selectively writing the data from the plurality of write buffer locations to the memory. The system further includes time stamp registers in communication with the write controller for storing and updating a time signal representative of the write buffer location having most recently received data. The controller is responsive in part to the time signal, and inhibits the writing to memory of the data in the write buffer location having most recently received data since this is the location most likely to be eligible for a data merge.

Abstract: A technique for providing skew compensation particularly in association with a pipelined processor. The skew occurs between first and second clock signals. The skew compensation technique of the invention provides for the proper transfer of information between stages even though the clock signals may have a skew greater than the inter-stage delay. A holding or latching means is provided between stages so as to hold the previous stage data for clocking into the subsequent stage register.

Abstract: In a computer system in which addressable components are physically organized on separately-replaceable printed circuit boards each containing an array of separately addressable components, diagnostic apparatus operates in the event of a component failure to assist a technician in physically locating the circuit board which contains the failed component. Each array includes a selection circuit which responds to component addresses located in the component array on that board. In the case of a component failure, diagnostic circuitry detects the address of the faulty component and places the address on the system address bus. The diagnostic circuitry controls each array to forward the output signal from the selection circuit on the associated printed circuit board to a register which has a position associated with each printed circuit board.

Abstract: A data processing system, wherein the central processing unit has an arithmetic element for processing data in response to machine program instructions and a control store for microcode program storage responsive to the machine instructions for implementing the instruction, has an improved arithmetic unit for enabling higher throughput without substantially increasing hardware cost. The arithmetic unit has a reconfigurable arithmetic logic unit which is controlled in response to both hardware generated data signals and microcode generated data signals. A data string manipulation circuitry provides for aligning data strings for processing by the arithmetic logic unit. Circuitry is provided, responsive to a decoded machine instruction, for generating control signals for configuring the arithmetic unit and for controlling the data string manipulation circuitry.

Abstract: A data processing system for processing a sequence of program instructions has a pipeline structure including an instruction pipeline and an execution pipeline. Each pipeline has a plurality of serially operating stages. The stages read from and modify memory at various stages of instruction processing. Collisions between data read from a register in the instruction pipeline phase of operation in response to a first instruction and write data written into the register during the execution phase of operation in response to an earlier instruction can be detected and predicted. In response thereto, the new data can be substituted directly for the modified data in the pipeline itself to provide continued valid operation. In addition, the apparatus and method provide for altering the flow of the instructions through the pipeline in order to accommodate newly generated data and to avoid invalid operation.

Abstract: A data processing system for processing a sequence of program instructions has two independent pipelines, an instruction pipeline and an execution pipeline. Each pipeline has a plurality of serially operating stages. The instruction stages read instructions from storage and form therefrom address data to be employed by the execution pipeline. The execution pipeline receives the address data and uses it for referencing stored data to be employed for execution of the program instructions. Both pipelines operate synchronously under the control of a pipeline control unit which initiates operation of at least one stage of the execution pipeline prior to completion of the instruction pipeline for a particular instruction. Thereby operation of at least one instruction stage and one execution stage of the respective pipelines overlap for each program instruction. The instruction and execution pipelines share high speed memory.

Abstract: A multiprocessor data processing system in which a number of independent processors can concurrently operate on a shared memory even when one processor is performing a read-modify-write (RMW) operation, the system having a locking, content-associative write buffer and a controller for identifying RMW requests, for addressing the buffer and, for issuing directives to lock the buffer, to validate particular data blocks in the buffer and to transfer data back and forth between the processors, the memory and the buffer.