Abstract: A critical data path of an integrated circuit includes a flip flop configured to receive a data input and provide a latched data output. A monitoring circuit includes a delay generator configured to receive the data input and provide a plurality of delayed data outputs corresponding to delayed versions of the data input with increasing amounts of delay, a selector circuit configured to select one of the plurality of delayed outputs based on a programmable control value, and a shadow latch coupled to an output of the selector circuit and configured to latch a value at its input to provide as a latched shadow output. A comparator circuit provides a match error indicator based on a comparison between the first latched data output and the latched shadow output, and an error indicator is provided which indicates whether or not an impending failure of the critical data path is detected.

Abstract: A circuit includes a P-channel transistor formed in a P-well and an N-channel transistor formed in an N-well. The first P-channel transistor has a control electrode connected to the P-well. The N-channel transistor is coupled in series with the P-channel transistor and has a control electrode connected to the N-well. Connecting the control electrodes of the P-channel and N-channel transistors to respective P-well and N-well effectively reduces crowbar current in the circuit.

Abstract: An integrated circuit includes a pulse generator having at least one delay circuit with an input that receives a clock signal and an output that provides a delayed clock pulse. The delayed clock pulse has a width proportional to an amount of time required to maintain a magnitude of the clock signal. A pulse latch circuit includes a clock input coupled to receive the delayed clock pulse, a data input coupled to receive a data value, and a data output, wherein the pulse latch circuit outputs and holds the data value at the data output each time the delayed clock pulse is provided at the clock input, and the pulse latch circuit operates on a continuous voltage source that supplies power during power on and power off modes.

Abstract: A semiconductor apparatus includes a first device cell and a second device cell. The first device cell includes a first active region including a first set of device fins, an insulator layer disposed over the first set of device fins, a first gate fin over the first set of fins, and a first edge fin disposed over a first edge of the first active region. The second device cell is adjacent the first device cell and includes a second active region including a second set of device fins, the insulator layer disposed over the second set of device fins, a second gate fin over the second set of device fins, and a second edge fin disposed over a second edge of the second active region. The first edge fin and the second edge fin are connected to a power rail, a ground rail, or to each other to define a capacitor between the first device cell and the second device cell.

Abstract: A method and apparatus for reducing dynamic switching current in high speed logic. The apparatus may include a CMOS logic circuit, which in turn includes an NMOS FinFET, a first PMOS FinFET, and a second PMOS FinFET. A gate of the NMOS FinFET is connected to a gate of the first PMOS FinFET, a drain of the NMOS FinFET is connected to a drain of the first PMOS FinFET, and the second PMOS FinFET is connected to the first PMOS FinFET to create a capacitor between a source and the drain of the first PMOS FinFET. In one embodiment, the second PMOS FinFET is contained in and positioned at an edge of a cell that also contains the first PMOS FinFET and the NMOS FinFET.

Abstract: An LVT-RVT cell includes an LVT PMOS transistor adjacent to an RVT NMOS transistor, whereby the LVT and RVT transistors are placed inside a common p-well and are biased using the same voltage potential. The cell thus employs a flipped well for the PMOS transistor and a conventional (unflipped) well for the NMOS transistor. By arranging the LVT-RVT cell in this way, the cell can function at lower voltages, thereby conserving power, while also improving the performance of the composite function. Furthermore, the LVT-RVT cell can be placed adjacent to RVT cells to further reduce power consumption and improve performance of the RVT cells within the block.

Abstract: A method and apparatus for reducing dynamic switching current in high speed logic. The apparatus may include a CMOS logic circuit, which in turn includes an NMOS FinFET, a first PMOS FinFET, and a second PMOS FinFET. A gate of the NMOS FinFET is connected to a gate of the first PMOS FinFET, a drain of the NMOS FinFET is connected to a drain of the first PMOS FinFET, and the second PMOS FinFET is connected to the first PMOS FinFET to create a capacitor between a source and the drain of the first PMOS FinFET. In one embodiment, the second PMOS FinFET is contained in and positioned at an edge of a cell that also contains the first PMOS FinFET and the NMOS FinFET.

Abstract: A circuit includes a P-channel transistor formed in a P-well and an N-channel transistor formed in an N-well. The first P-channel transistor has a control electrode connected to the P-well. The N-channel transistor is coupled in series with the P-channel transistor and has a control electrode connected to the N-well. Connecting the control electrodes of the P-channel and N-channel transistors to respective P-well and N-well effectively reduces crowbar current in the circuit.

Abstract: A package to die connection system and method are provided. The system includes a semiconductor device having a substrate with a top surface. A gasket is affixed to the top surface of the substrate and has at least one cavity with a portion of the cavity open to a sidewall of the gasket. A semiconductor die is attached to the top surface of the substrate. A sidewall of the semiconductor die is abutted with the sidewall of the gasket. A portion of a metal layer is exposed to the open portion of the cavity. A pillar located in the cavity is electrically connected to the exposed portion of the metal layer.

Abstract: A FinFet capacitor structure includes a first, second, third, and fourth FinFet fin, a contiguous gate layer over the fins, first and second source/drain contacts in direct physical contact with the first FinFet fin on either side of the gate layer, a first gate contact in direct physical contact with a portion of the contiguous gate layer directly over the second FinFet fin, third and fourth source/drain contacts in direct physical contact with the third FinFet fin on either side of the gate layer, and a second gate contact in direct physical contact with a portion of the contiguous gate layer directly over the fourth FinFet fin. The first, second, third, and fourth source/drain contacts are all connected to a first power supply rail, and the first and second gate contacts are all connected to a second power supply rail.

Abstract: An integrated circuit includes a pulse generator having at least one delay circuit with an input that receives a clock signal and an output that provides a delayed clock pulse. The delayed clock pulse has a width proportional to an amount of time required to maintain a magnitude of the clock signal. A pulse latch circuit includes a clock input coupled to receive the delayed clock pulse, a data input coupled to receive a data value, and a data output, wherein the pulse latch circuit outputs and holds the data value at the data output each time the delayed clock pulse is provided at the clock input, and the pulse latch circuit operates on a continuous voltage source that supplies power during power on and power off modes.

Abstract: A multi-bit flip-flop includes at least two storage stages. Each of the storage stages includes redundant latches to suppress state corruptions resulting from soft error upset at the storage stage. In addition, the multi-bit flip-flop includes a split clock path that routes different shared clock signals that control the timing of the latches. The shared split clock path reduces or eliminates the impact of soft errors on the clock signals, thereby further limiting the impact of such errors on data stored at the flip-flop. In particular, the split clock path can be distributed over disparate cells in a layout of multi-bit flip-flop, thereby reducing the likelihood that a transient charge will cause a soft error in all paths of the split clock path.

Abstract: A multi-bit flip-flop includes at least two storage stages. Each of the storage stages includes redundant latches to suppress state corruptions resulting from soft error upset at the storage stage. In addition, the multi-bit flip-flop includes a split clock path that routes different shared clock signals that control the timing of the latches. The shared split clock path reduces or eliminates the impact of soft errors on the clock signals, thereby further limiting the impact of such errors on data stored at the flip-flop. In particular, the split clock path can be distributed over disparate cells in a layout of multi-bit flip-flop, thereby reducing the likelihood that a transient charge will cause a soft error in all paths of the split clock path.

Abstract: A method and system for performing memory repair via redundant rows of memory uses memory elements (208 and 210) for redundant row selection instead of conventional fuses. An on-chip test controller (110) is capable of testing memory rows (106) either at wafer probe, at final testing after manufacturing, or after memory chip packaging and/or final sale to end users. If this testing identifies faulty memory rows in the memory array at any time, the electrically programmable memory elements (208 and 210) can be internally re-programmed to create a new memory configuration which includes redundant memory rows (108). This new memory configuration is enabled in order to remove the newly-detected and previously-detected faulty memory rows from active memory in the memory array.