Abstract: Embodiments provide for a computer implemented method comprising sampling one or more power and performance metrics of a processor; determining an energy cost per instruction based on the one or more power and performance metrics; determining an efficiency metric based on the energy cost per instruction; computing an efficiency control error based on a difference between a current efficiency metric and a target efficiency metric; setting an efficiency control effort based on the efficiency control error; determining a performance control effort, the performance control effort determined by a performance controller for the processor; and adjusting the performance control effort based on the efficiency control effort, wherein adjusting the performance control effort reduces power consumption of the processor.

Abstract: Embodiments relate to modulating a power supply voltage for varying a propagation delay of data paths within an integrated circuit. The power supply voltage is modulated to increase the delay of shorter data paths for reducing an incidence of hold time violations without substantially affecting the delay of longer data paths. For example, the power supply voltage is reduced from a nominal value in the first half clock cycle to increase delay of both the shorter data paths and the longer data paths. The power supply voltage is increased from the nominal value in the second half clock cycle to decrease delay of the longer data paths within the second half clock cycle such that the overall delay of the longer data paths is virtually same as when the power supply voltage is fixed at the nominal value for the entire clock cycle.

Abstract: Embodiments relate to modulating a power supply voltage for varying a propagation delay of data paths within an integrated circuit. The power supply voltage is modulated to increase the delay of shorter data paths for reducing an incidence of hold time violations without substantially affecting the delay of longer data paths. For example, the power supply voltage is reduced from a nominal value in the first half clock cycle to increase delay of both the shorter data paths and the longer data paths. The power supply voltage is increased from the nominal value in the second half clock cycle to decrease delay of the longer data paths within the second half clock cycle such that the overall delay of the longer data paths is virtually same as when the power supply voltage is fixed at the nominal value for the entire clock cycle.

Abstract: In an embodiment, the amount of supply voltage guardband to prevent incorrect operation due to aging effects may be modeled using an IC-specific age model generated early in the product life cycle of the IC. For example, high temperature operating life (HTOL) testing may be performed at multiple temperatures and/or voltages to develop the IC-specific age model. The IC-specific age model may be more accurate then the calculations used to develop guardband voltage as discussed previously, which rely on the aging of a single transistor. The IC-specific age model may be used along with monitoring of the aging effects during operation of the IC to predict an amount of increased guardband voltage that is currently desirable to apply to the IC. The predicted amount may vary from about zero when the IC is new to the full amount of guardband voltage when the IC is nearing end of life.

Abstract: Embodiments of an apparatus are disclosed that may allow for the isolation of power domains. The apparatus may include a first power switch, a second power switch, and a boundary switch. The first power switch may be coupled between a global power supply and a first local power supply, and the second power switch may be coupled between the global power supply and a second local power supply. The first and second power switches may open in response to first and second power down signals respectively. The boundary switch may be coupled between the first local power supply and the second local power supply and may be configured to open in response to an isolation signal.

Abstract: A method, system, and computer program for using an array of networked 3D voltage regulation modules (VRMs) to optimize power usage by components on a voltage island in real time is presented. The networked VRM devices work in parallel to supply adequate power to connected voltage islands, and to supplement other VRMs in the system that may require additional power in the case of a critical event.

Abstract: Disclose are embodiments of an integrated circuit design method based on a combination of manufacturability, test coverage and, optionally, diagnostic coverage. Design-for manufacturability (DFM) modifications to the layout of an integrated circuit can be made in light of test coverage. Alternatively, test coverage of an integrated circuit can be established in light of DFM modifications. Alternatively, an iterative process can be performed, where DFM modifications to the layout of an integrated circuit are made in light of test coverage and then test coverage is altered in light of the DFM modifications. Alternatively, DFM modifications to the layout of an integrated circuit can be made in light of test coverage and also diagnostic coverage. In any case, after making DFM modifications and establishing test coverage, any unmodified and untested nodes (and, optionally, any unmodified and undiagnosable tested nodes) in the integrated circuit can be identified and tagged for subsequent in-line inspection.

Abstract: A method, system, and computer program for using an array of networked 3D voltage regulation modules (VRMs) to optimize power usage by components on a voltage island in real time is presented. The networked VRM devices work in parallel to supply adequate power to connected voltage islands, and to supplement other VRMs in the system that may require additional power in the case of a critical event.

Abstract: An error-correction code is generated on a line-by-line basis of the physical logic register and latch contents that store encoded words within a processor just before the processor is put into sleep mode, and later-generated syndrome bits are checked for any soft errors when the processor wakes back up, e.g., as part of the power-up sequence.

Abstract: A method, system, and computer program for using an array of networked 3D voltage regulation modules (VRMs) to optimize power usage by components on a voltage island in real time is presented. The networked VRM devices work in parallel to supply adequate power to connected voltage islands, and to supplement other VRMs in the system that may require additional power in the case of a critical event.

Abstract: Disclose are embodiments of an integrated circuit design method based on a combination of manufacturability, test coverage and, optionally, diagnostic coverage. Design-for manufacturability (DFM) modifications to the layout of an integrated circuit can be made in light of test coverage. Alternatively, test coverage of an integrated circuit can be established in light of DFM modifications. Alternatively, an iterative process can be performed, where DFM modifications to the layout of an integrated circuit are made in light of test coverage and then test coverage is altered in light of the DFM modifications. Alternatively, DFM modifications to the layout of an integrated circuit can be made in light of test coverage and also diagnostic coverage. In any case, after making DFM modifications and establishing test coverage, any unmodified and untested nodes (and, optionally, any unmodified and undiagnosable tested nodes) in the integrated circuit can be identified and tagged for subsequent in-line inspection.

Abstract: An integrated circuit includes a plurality of processor cores and a readable non-volatile memory that stores information expressive of at least one operating characteristic for each of the plurality of processor cores. Also disclosed is a method to operate a data processing system, where the method includes providing a multicore processor that contains a plurality of processor cores and a readable non-volatile memory that stores information, determined during a testing operation, that is indicative of at least a maximum operating frequency for each of the plurality of processor cores. The method further includes operating a scheduler coupled to an operating system and to the multicore processor, where the scheduler is operated to be responsive at least in part to information read from the memory to schedule the execution of threads to individual ones of the processor cores for a more optimal usage of energy.

Abstract: A computer system for designing a low leakage monotonic CMOS logic circuit. The system performing the computer implements steps of: (a) specifying a reference PFET having its threshold voltage and its gate dielectric thickness and a reference NFET having its threshold voltage and its gate dielectric thickness; (b) synthesizing a schematic circuit design with standard design elements, the standard design elements including one or more reference PFETS and one or more reference NFETs; (c) analyzing one or more circuits for logic stages having predominantly high input logic states or predominantly low input logic states; (d) selecting one or more logic stages determined to have predominantly high input logic states or predominantly low input logic states; and (e) replacing the standard design elements of the selected logic stages with reduced current leakage elements.

Abstract: The invention includes a design structure embodied in a computer readable medium for performing a method for inserting a scan chain into a VLSI circuit design. The scan chain structure, or structures, are included in the design structure for the VLSI circuit design. The scan chain structure includes a first flip-flop (L1) and a second flip-flop (L2) configured to operate the first flip-flop (L1) in normal mode operation, in scan mode operation, in initialization mode and in low leakage power mode operation. A buffer circuit is electrically connected between the scan-out output of the second flip-flop (L2) and the scan-in input of the first flip-flop (L1) for the next latch in the scan chain. Buffer circuit control elements control the first flip-flop (L1) to switch between scan mode or low power leakage mode. The switching occurs in only one clock cycle.

Abstract: The invention includes a novel scan chain structure for LSSD or GSD IC operation. The scan chain structure includes a first flip-flop (L1) and a second flip-flop (L2) configured to operate the first flip-flop (L1) in normal mode operation, in scan mode operation, in initialization mode and in low leakage power mode operation, wherein each flip-flop within a long scan chain of latches includes a data input, data output, a clock input, a scan-in input and a scan-out output, arranged for normal mode operation. A buffer circuit is electrically connected between the scan-out output of the second flip-flop (L2) and the scan-in input of the first flip-flop (L1) for the next latch in the scan chain, the buffer circuit including a control element that controls the operation the first flip-flop (L1) to scan mode or low power leakage mode. The first flip-flop (L1) is set to a data output value upon exit from low power leakage mode that is the same value that it is set to at initialization during normal mode operation.

Abstract: A static random access memory (SRAM) circuit includes first SRAM cell and a second SRAM cell that are configured to operate in a shared mode and/or an independent mode. In one example, a shared mode includes the sharing of a memory node of a first SRAM cell. In another example, an independent mode includes isolating a first SRAM cell from a second SRAM cell such that they operate independently.

Abstract: The present invention relates to a laser fuse structure for high power applications. Specifically, the laser fuse structure of the present invention comprises first and second conductive supporting elements (12a, 12b), at least one conductive fusible link (14), first and second connection elements (20a, 20b), and first and second metal lines (22a, 22b). The conductive supporting elements (12a, 12b), the conductive fusible link (14), and the metal lines (22a, 22b) are located at a first metal level (3), while the connect elements (20a, 20b) are located at a second, different metal level (4) and are connected to the conductive supporting elements (12a, 12b) and the metal lines (22a, 22b) by conductive via stacks (18a, 18b, 23a, 23b) that extend between the first and second metal levels (3, 4).

Abstract: A method, system and computer program product for estimating a static power consumption of an integrated circuit are disclosed. The static power consumption of a cell of the integrated circuit is characterized based on contributions of an input node(s) and an output node(s) of the cell. A contribution considers a leakage weight and a leakage probability of a node. A logic template of the cell may be created to better represent a contribution of an internal node to the static power consumption of the cell.

Abstract: An error-correction code is generated on a line-by-line basis of the physical logic register and latch contents that store encoded words within a processor just before the processor is put into sleep mode, and later-generated syndrome bits are checked for any soft errors when the processor wakes back up, e.g., as part of the power-up sequence.

Abstract: The invention includes an error correcting logic system that allows critical circuits to be hardened with only one redundant unit and without loss of circuit performance. The system provides an interconnecting gate that suppresses a fault in one of at least two redundant dynamic logic gates that feed to the interconnecting gate. The system is applicable to dynamic or static logic systems. The system prevents propagation of a fault, and addresses not only soft errors, but noise-induced errors. Also, there is provided a design structure embodied in a machine readable medium used in a design process, and which includes such error correcting logic system.