Abstract: Memory performance in programmable logic is significantly increased by adjusting a timing of control signals sent to a memory to compensate for variations in process, voltage, or temperature. A calibration circuit can adjust the control signal timing, dynamically and automatically, to provide accurate and high performance memory operations. For example, timing settings for the control signals can be determined such that data written/read from the memory are accurate. The timing setting can also be changed to provide faster memory operations while still providing accuracy. A feedback system using a control block and a dummy mimicking concept are also provided.

Abstract: Providing ASIC equivalents of FPGAs is facilitated and made more efficient and economical by using an ASIC architecture including a plurality of so-called hybrid logic elements (“HLEs”), each of which can provide a portion of the full functionality of an FPGA logic element (“LE”). The functionality of each FPGA LE implementing a user's logic design can be mapped to one or more HLEs without re-synthesis of the user's logic. Only as many HLEs as are necessary are used to perform the functions of each LE. The one-for-one equivalence between each LE and either (1) one HLE or (2) a group of HLEs facilitates mapping (without re-synthesis) in either direction between FPGA and ASIC designs.

Abstract: Methods and apparatus for programmably powering down a structured application-specific integrated circuit are provided. At least one of the programmable layers of the structured ASIC that frequently provides some programmability as between or among a small number of alternative functions is used to provide this programmability.

Abstract: A delay circuit that includes a first delay cell oriented in a first orientation and a second delay cell oriented in a second orientation is described. In one embodiment, the first orientation is perpendicular to the second orientation. More specifically, in one embodiment, the first orientation is vertical and the second orientation is horizontal.

Abstract: A logic module for a structured ASIC is mask-programmable to perform any of a plurality of logic functions or to alternatively function as two static random access memory (“SRAM”) cells. Most or all of the circuitry needed to enable the logic module to function as SRAM cells is circuitry that is available for use in logic mode configuration of the module. Similarly, most or all of the logic circuitry of the logic module is put to use to provide the SRAM cells when the module is configured in SRAM mode. The circuitry of the logic module is therefore used very efficiently, whether the module is configured for logic mode or SRAM mode.

Abstract: Clock distribution circuitry for a structured ASIC device includes a deterministic portion and configurable portions. The deterministic portion employs a predetermined arrangement of conductor segments and buffers for distributing a clock signal to a plurality of predetermined locations on the device. From each predetermined location, an associated configurable portion of the clock distribution circuitry distributes the clock signal to any clock utilization circuitry needing that clock signal in a predetermined area of the structured ASIC that is served from that predetermined location.

Abstract: Clock distribution circuitry for a structured ASIC device includes a deterministic portion and configurable portions. The deterministic portion employs a predetermined arrangement of conductor segments and buffers for distributing a clock signal to a plurality of predetermined locations on the device. From each predetermined location, an associated configurable portion of the clock distribution circuitry distributes the clock signal to any clock utilization circuitry needing that clock signal in a predetermined area of the structured ASIC that is served from that predetermined location.

Abstract: A method is provided for creating and using a library of known logic elements for facilitating the design of equivalent FPGA, structured ASIC, or other integrated circuits. Each cell in the library corresponds to a circuit element and contains equivalent circuit models of the element for implementation in different integrated circuit technologies. The cells are named and indexed using a library cell key that contains a fixed set of cell properties which uniquely characterize the function of the cell. The cell key can be used to locate cells in the library or in circuit designs, and to verify that the circuit models contained in the cell implement the correct logic function.

Abstract: Methods and apparatus for programmably powering down a structured application-specific integrated circuit are provided. At least one of the programmable layers of the structured ASIC that frequently provides some programmability as between or among a small number of alternative functions is used to provide this programmability.

Abstract: Structured ASICs that are equivalent to FPGA logic designs are produced by making use of a library of known structured ASIC equivalents to FPGA logic functions. Such a library is expanded by a process that searches new FPGA logic designs for logic functions that either do not already have structured ASIC equivalents in the library or for which possibly improved structured ASIC equivalents can now be devised. The new and/or improved structured ASIC equivalents are added to the library, preferably with version information in the case of FPGA logic functions for which more than one structured ASIC equivalent is known.

Abstract: Providing ASIC equivalents of FPGAs is facilitated and made more efficient and economical by using an ASIC architecture including a plurality of so-called hybrid logic elements (“HLEs”), each of which can provide a portion of the full functionality of an FPGA logic element (“LE”). The functionality of each FPGA LE implementing a user's logic design can be mapped to one or more HLEs without re-synthesis of the user's logic. Only as many HLEs as are necessary are used to perform the functions of each LE. The one-for-one equivalence between each LE and either (1) one HLE or (2) a group of HLEs facilitates mapping (without re-synthesis) in either direction between FPGA and ASIC designs.

Abstract: As part of a process for producing a structured ASIC that is functionally equivalent to an FPGA that has been programmed to perform a user's logic design, a compilation of that design that has been prepared for ASIC implementation is converted to a physical layout of the structured ASIC. The production of this physical layout honors timing constraints supplied by the user, and also preserves functional equivalence to the reference programmed FPGA. The structured ASIC can be manufactured from the physical layout produced.

Abstract: Memory performance in programmable logic is significantly increased by adjusting circuitry operation to adjust for variations in process, voltage, or temperature. A calibration circuit adjusts control signal timing, dynamically and automatically, to compensate real time to process, voltage, and temperature variation. A feedback system using a control block and a dummy mimicking concept are provided.

Abstract: Memory performance of an integrated circuit, such as a programmable logic integrated circuit, is increased by pipelining. In a single clock cycle, more than one operation may be performed on the memory, which improves bandwidth. In an implementation, the memory architecture having one port supports pipelining, so reading from and writing to the memory can be accomplished in a one clock cycle and both Read and Write operation can occupy the full clock cycle at the same time on the same port. The pipelining architecture has relatively minimal circuit changes compared to a standard memory architecture which not supporting simultaneous-clock-cycle reads and writes, without requiring two or more ports.

Abstract: Circuitry and methods are provided for controlling memory operation by comparing bit line voltages to preset reference voltages. By relying on bit line voltage levels to determine when to start and end each stage of a read or write operation, reliance on precisely tuned delay chains is removed. Parasitic effects are automatically accounted for, as well as process, voltage, and temperature variations. This precise matching of operation timing to memory bit line conditions results in improved system performance.

Abstract: A test platform is configured to test a mult-die package having at a first die and a second die. The test platform includes a first lead that is connected to the VCC input on the first die. The test platform also includes a second lead that is connected to VCCIO input on the second die. The VCC input on the second die is connected to ground. The I/O pin of the second die can then be tri-stated using a control circuit disposed between the pre-driver and the driver of the I/O buffer.