Abstract: A method of correcting an error in an ECC protected mechanism of a computer system, such as a cache or system bus, by applying data with a number of bits N to an error correction code (ECC) matrix to yield an error detection syndrome, wherein the ECC matrix has a plurality of rows and columns with a given column corresponding to a respective one of the data bits, and selected bits are set in the ECC matrix along each column and each row such that encoding for the ECC matrix allows N-bit error correction and (N?1)-bit error detection. In the illustrative embodiment, the ECC matrix has an odd number of bits set in each row thereof. In the case of an ECC protected mechanism such as a memory device, these properties facilitate the use of an inversion bit for correcting hard faults in the stored data. When an error is detected and after it is corrected, the corrected data is inverted and then rewritten to the cache array.

Abstract: A method of reducing errors in a cache memory of a computer system (e.g., an L2 cache) by periodically issuing a series of purge commands to the L2 cache, sequentially flushing cache lines from the L2 cache to an L3 cache in response to the purge commands, and correcting errors (single-bit) in the cache lines as they are flushed to the L3 cache. Purge commands are issued only when the processor cores associated with the L2 cache have an idle cycle available in a store pipe to the cache. The flush rate of the purge commands can be programmably set, and the purge mechanism can be implemented either in software running on the computer system, or in hardware integrated with the L2 cache. In the case of the software, the purge mechanism can be incorporated into the operating system. In the case of hardware, a purge engine can be provided which advantageously utilizes the store pipe that is provided between the L1 and L2 caches.

Abstract: A method of reducing errors in a cache memory of a computer system (e.g., an L2 cache) by periodically issuing a series of purge commands to the L2 cache, sequentially flushing cache lines from the L2 cache to an L3 cache in response to the purge commands, and correcting errors (single-bit) in the cache lines as they are flushed to the L3 cache. Purge commands are issued only when the processor cores associated with the L2 cache have an idle cycle available in a store pipe to the cache. The flush rate of the purge commands can be programmably set, and the purge mechanism can be implemented either in software running on the computer system, or in hardware integrated with the L2 cache. In the case of the software, the purge mechanism can be incorporated into the operating system. In the case of hardware, a purge engine can be provided which advantageously utilizes the store pipe that is provided between the L1 and L2 caches.

Abstract: Methods, and arrangements to enhance speed and reduce power consumption in a scanable latch circuit are disclosed. Embodiments include a wired-or circuit to facilitate independent paths for scan data and normal input data through the scanable latch circuit. In particular, to reduce delays related to gates between the input pin for the system clock and a normal input gate, dual, substantially independent paths are implemented: a scan path and a normal input path. Embodiments coordinate transmission of data from a normal input gate and a scan input gate to an output latch, a scan out pin, and/or combinational logic by incorporating buffers that isolate a wired-or node from either the scan input gate, the normal input gate, or both with a high impedance.

Abstract: This invention describes a method to verify non-looping properties of programs implemented as rule-based expert systems. Our method detects conditions under which the expert system enters erroneous infinite program loops, which lead to non-terminating or oscillating computations, or otherwise proves the absence of such conditions. Our automatic procedure also gives advice on how to correct these errors. The expert systems considered consist of condition-action rules (IF-THEN-statements), where the conditions are logical expressions (formulas of a propositional finite domain logic), and the actions modify the value of a single variable which in turn can be part of other logical expressions. There may be additional (external) variables not controlled by the expert system, and each rule may have an associated evaluation priority.

Abstract: A method of determining a circuit response (such as delay or slew) from a ramp input of an RC circuit calculates two circuit response parameters using a given circuit response metric based on a step input for the RC circuit, and extends the circuit response metric to a ramp input of the RC circuit by combining the first and second circuit response parameters to yield an estimated ramp response. The novel technique is based on the use of probability distribution functions and cumulative distribution functions to characterize the impulse response of the RC circuit, and the calculating steps derive the first and second circuit response parameters from such statistical distribution functions. In particular, the calculating steps may use a standard deviation or a mean of a probability distribution function corresponding to the circuit response parameter. In one application, the invention is used to estimate delay response for the ramp input of the RC circuit.

Abstract: In one form, a method for communicating among subsystems coupled to a bus of a computer system on an integrated circuitry chip includes operating subsystems at independent clock frequencies when the subsystems are not communicating with one another on the bus. Selected pairs of the subsystems are operated at a shared clock frequency by selectively varying frequencies of clock signals to the subsystems, so that communication can occur at the shared clock frequency on the bus between the selected subsystems, but at different clock frequencies for respective different pairings of the subsystems, and so that the subsystems can operate at independent clock frequencies when not communicating with other ones of the subsystems. Communication among the subsystems is by a bus-based protocol, according to which when a subsystem is granted access to the bus the subsystem has exclusive use of the bus.

Abstract: Methods and arrangements for a low power, phase-locked loop (PLL) are disclosed. Embodiments include a multi-phase oscillator like a voltage-controlled oscillator (VCO) to generate multiple phases of a clock signal. The multiple phases are then combined to generate a single clock signal having a frequency substantially equivalent to the number of phases multiplied by the frequency of the clock signal generated by the multi-phase VCO. Advantageously, embodiments can generate clock signals having frequencies that are multiples of the frequency generated by the VCO, reducing the power consumed by the VCO to produce a clock signal having the same frequency as a clock signal generated by a single phase VCO. Further, the achievable frequency for the VCO is increased. In many embodiments, a high speed, n-bit frequency divider that implements a pulse latch facilitates the use of the multi-phase VCO to generate the very high frequency clock signals.

Abstract: A data processing system that includes a mode/reserve bit utilized to dynamically change a processor's operating mode between a virtual addressing mode and a real addressing mode. Each address block includes a reserve bit that indicates whether real or virtual addressing is desired, and the reserve bit is assigned a value by the software application executing on the processor. The value of the reserve bit is dynamically set and signals the processor which operating mode is required for the particular address block. The selection of virtual or real addressing mode is determined by the particular application that is being executed by the processor. When the particular application process seeks increased performance rather than protection, the virtual operating mode is selected, allowing the application process to send the effective addresses directly to the OS and hypervisor. This is accomplished by setting the reserve bit to the value for virtual addressing mode.

Abstract: A method and system for reducing noise in a power grid of an integrated circuit, which optimizes the placement and sizing of decoupling capacitors in the power grid. Logic cells are located in a first layout of the integrated circuit with empty spaces between the adjacent cells, and the placement of the cells is changed to a second layout wherein the size of the empty spaces between the adjacent cells also change. The decoupling capacitors are placed in the empty spaces of the second layout. In the example of a row-oriented cell structure, the empty spaces may be uniformly distributed along each row for the initial layout. An adjoint sensitivity analysis is performed of the sensitivity of a noise function of the integrated circuit with respect to sizes of the empty spaces between adjacent cells, and an original noise waveform is convolved with an adjoint noise waveform. The convolution may use piecewise linear compressions of the original and adjoint noise waveforms.

Abstract: Methods and arrangements for enhancing domino logic are disclosed. Embodiments include a keeper circuit to pull up a domino node in response an output of an output circuit when the domino node is at a high voltage and to stop pulling up the domino node before the output changes to a first logical output. Further embodiments include an accelerator circuit to pull down the domino node when the keeper circuit stops pulling up the domino node. The domino node may couple with a pre-charge circuit and be pre-charged to a high voltage during a first portion of a clock cycle. The domino node may also couple with a logic input circuit to pull down the domino node during a second portion of the clock cycle, causing the output circuit to change the output from low to high in response to logic signals.

Abstract: A method of determining a circuit response (such as delay or slew) from a ramp input of an RC circuit calculates two circuit response parameters using a given circuit response metric based on a step input for the RC circuit, and extends the circuit response metric to a ramp input of the RC circuit by combining the first and second circuit response parameters to yield an estimated ramp response. The novel technique is based on the use of probability distribution functions and cumulative distribution functions to characterize the impulse response of the RC circuit, and the calculating steps derive the first and second circuit response parameters from such statistical distribution functions. In particular, the calculating steps may use a standard deviation or a mean of a probability distribution function corresponding to the circuit response parameter. In one application, the invention is used to estimate delay response for the ramp input of the RC circuit.

Abstract: A method and controller for managing power and performance of a multiprocessor (MP) system is described. The controller receives sensor data corresponding to physical parameters within the MP system. The controller also receives quality of service and policy parameters corresponding to the MP system. The quality of service parameters define commitments to customers for utilization of the MP system. The policy parameters correspond to operation limits on inputs and outputs of the MP system. The operation input limits relate to the cost and availability of power or individual processor availability. The operation output limits relate to the amount of heat, acoustic noise levels, EMC levels, etc. that the individual or group of processors in the MP system are allowed to generate in a particular environment. A controller receives the physical parameters, the quality of service parameters and policy parameters and determines performance goals for the MP system and processors within the MP system.

Abstract: An apparatus and method for swapping out real memory by inhibiting input/output (I/O) operations to a memory region are provided. The apparatus and method provide a mechanism in which a quiesce indicator is provided in a field containing the current outstanding I/O count associated with the memory region whose real memory is to be swapped out. The current I/O field and the quiesce indicator are used as a means for communicating between a shared resource arbitrator and a guest consumer. When the quiesce indicator is set, the guest consumer is informed that it should not send any further I/O operations to that memory region. When the number of pending I/O operations against the memory region is zero, a valid bit in a protection table is set to invalid, and the real memory associated with the memory region may be swapped out.

Abstract: Within a data processing system, a pool of facilities are allocated to an operating system, where each facility within the pool of facilities has an associated real address. The operating system allocates from the pool at least one bypass facility to a first process that the first process is permitted to directly access by its associated real address without first obtaining translation of a non-real address. The operating system also allocates from the pool at least one protected facility to a second process that the second process accesses only by translation of a non-real address to obtain the real address associated with the protected facility. Accesses to the facilities are either protected or unprotected based upon the state of a bypass field within a request address.

Abstract: A system comprising a cluster of diskless servers employing a distributed shared memory abstraction that presents an area of shared memory for two or more processes executing on different servers in the cluster. The invention provides the appearance of a shared memory space between two or more processes thereby potentially reducing disk latency or eliminating redundant computation associated with conventional server clusters. The DSM abstraction may be dynamically alterable such that selectable groups of processes executing on the cluster share a common address space temporarily. The shared memory spaces may be determined empirically or servers may subscribe to a group reactively in response to client requests. Multiple groups may exist simultaneously and a single server may belong to more than one group. The types of objects to which the abstraction is applied may be restricted. Shared memory may be restricted, for example, to read-only objects to alleviate consistency considerations.

Abstract: A data processing system includes a global promotion facility and a plurality of processors coupled by an interconnect. In response to execution of an acquisition instruction by a first processor among the plurality of processors, the first processor transmits an address-only operation on the interconnect to acquire a promotion bit field within the global promotion facility exclusive of at least a second processor among the plurality of processors. In response to receipt of a combined response for the address-only operation representing a collective response of others of the plurality of processors to the address-only operation, the first processor determines whether or not acquisition of the promotion bit field was successful by reference to the combined response.

Abstract: A system and method for capturing a point-in-time image of a file system utilized within a data storage system employing logical volumes mapped across physical volumes. The point-in-time backup process begins with receipt of a file system backup request. In response to receiving the file system backup request, buffered file system data is written or flushed to the logical volumes. Specifically, in response to the file system backup request, buffered user data is written to the logical volumes, buffered meta data associated with the buffered user data is written to a file system backup log within the logical volumes, and the buffered meta data is written to the logical volumes. A disk copy of the designated physical volumes is initiated in response to the writing of the buffered file system data to the logical volumes.

Abstract: A method, processor architecture, computer program product, and data processing system for determining when an instruction in a pipelined processor should be completed is provided. As each instruction is issued to an execution unit, an entry for that instruction is placed within a “finish pipe,” which consists of a series of consecutively numbered stages. Each clock cycle, the entries in the finish pipe advance one stage. When an entry has reached the stage corresponding to the latency of its associated execution unit, it becomes mature. Each clock cycle, the finish pipe is scanned to find the entry having the highest-numbered stage of any entry in the finish pipe. If that entry is mature, it is removed from the finish pipe and the instructions associated with that entry is allowed to complete. If not, the entry simply advances along with the other entries and the pipe rescanned in the next cycle.

Abstract: A symmetric multiprocessor data processing system having a highly scalable shared cache memory hierarchy is disclosed. The symmetric multiprocessor data processing system includes multiple processing units. Each of the processing units includes a level one cache memory. All the level one cache memories are associated with a level two cache memory. The level two cache memory is non-inclusive of all the level one cache memories. An invalidation bus is connected to all of the level one cache memories. In response to a write access to a specific cache line within one of the level one cache memories, the invalidation bus invalidates other cache lines that shared identical information with the specific cache line within the rest of the level one cache memories.