Abstract: Signals sent from one system-on-chip core become switched to a reconfigurable logic core (RLC) for observation and, perhaps, replacement with another signal. A first signal line couples together a plurality of cores. A switch, disposed between the first signal line and an input signal line of the RLC, selectively controls whether the signal gets sent to the RLC. A multiplexer, having the first signal line and an output signal line of the RLC as inputs, selectively controls whether the signal or a replacement signal becomes conveyed to another core of the system-on-chip. Observation and control configuration memory bits act as inputs in the selective control of the switch and the multiplexer. Other embodiments teach shared RLC input signal lines amongst multiple cores. The RLC may contain an inverter, a test circuit, a logic analyzer or other. Methods of observing and replacing signals are also taught.

Abstract: A multiport memory in one embodiment of the invention includes a memory cell array, where each column in the array has two exterior complementary bitline pairs and zero, one, or more interior complementary bitline pairs. Across each pair of adjacent columns in the array, the adjacent exterior bitline pairs are associated with the same port in the multiport memory. In addition, within each column, the two exterior bitline pairs have the same, non-zero number of crossovers, and, across each pair of adjacent columns, the exterior bitline pairs have different numbers of crossovers. Furthermore, each column has at least one reference signal line located between the two exterior bitline pairs.

Abstract: A clock generator has a reset circuit and (at least) two dividers, where each divider divides a reference clock signal by a divisor value to generate an output clock signal. The reset circuit generates reset signals for the dividers, where the reset signals are delayed relative to one another by a selected number of reference clock cycles, such that the dividers generate output clock signals having a desired phase offset between them.

Abstract: In one embodiment of the invention, a block of configuration memory has rows of memory cells, at least one row having a set of one or more dual-port memory cells adapted to selectively store either configuration data or local data. The configuration address line for that row is segmented such that the address line is connected to the configuration address ports of the dual-port memory cells via access control circuitry that can be programmably configured to prevent access to those memory cells via the configuration address line. The access control circuitry enables local data to be efficiently and accurately stored in the dual-port memory cells without interference from configuration readback operations during normal operation or from partial reconfiguration of the configuration memory.

Abstract: Fractional cycle stealing units are introduced in the routing of a programmable device such as an FPGA or FPSC to increase system performance resulting from the particular clock routing. The disclosed fractional cycle stealing units enable given amounts of clock skew between individual distribution sinks, and/or between logic paths that are in series. Each of the delay elements ‘steals’ a portion of a clock cycle (and perhaps one or more full clock cycles) from subsequent circuits to provide a more reliable logical function, and to avoid the need for overall additional clock cycles. These fractional cycle stealing elements offer a signal skew adjustment at the sinks of the distribution with no additional routing congestion expense. The disclosed cycle stealing delay elements are programmable, and enable clock skew between individual distribution sinks.

Abstract: A multi-functional programmable I/O buffer in a Field Programmable Gate Array (FPGA) device. The I/O buffer is programmably configurable to meet any of a wide range of I/O standards, be it single ended or differential, 5V, 3.3V, 2.5V or 1.5V logic, without the need for implementing multiple I/O buffers to properly handle each different iteration of I/O requirements. An embedded, internal programmable resistor (e.g., a programmable 100 ohm resistor) is programmably selected for use in differential I/O applications, thus eliminating the conventional requirement for the use of an external resistor connected to each differential receiver I/O pin. The present invention also separates I/O pads into groups in each of a plurality of banks in a programmable device (e.g., PLD, FPGA, etc.), with each group being separately powered by the user. The disclosed multi-functional I/O buffer may be programmably configured by the user to be, e.g.

Abstract: Fractional cycle stealing units are introduced in the routing of a programmable device such as an FPGA or FPSC to increase system performance resulting from the particular clock routing. The disclosed fractional cycle stealing units enable given amounts of clock skew between individual distribution sinks, and/or between logic paths that are in series. Each of the delay elements ‘steals’ a portion of a clock cycle (and perhaps one or more full clock cycles) from subsequent circuits to provide a more reliable logical function, and to avoid the need for overall additional clock cycles. These fractional cycle stealing elements offer a signal skew adjustment at the sinks of the distribution with no additional routing congestion expense. The disclosed cycle stealing delay elements are programmable, and enable clock skew between individual distribution sinks.