Abstract: A system and method are disclosed for providing highly parallel, FFT calculations in a circuit including a plurality of RADIX-2 elements. Partitioned RAM resources allow RADIXes at all stages to have optimal bandwidth memory access. Preferably more memory is made available for early RADIX stages and a “critical” stage. RADIXes within stages beyond the critical stage preferably each need only a single RAM partition, and can therefore simultaneously operate without fighting for memory resources. In a preferred configuration having P RAM partitions and P RADIX stages, the critical stage is stage number log2 P, and until the critical stage, only P/2 RADIX elements can simultaneously operate within each stage. After the critical stage, all RADIXes within each stage can simultaneously operate.

Abstract: A method and system for distributing by a server data processing system computing tasks for execution amongst a plurality of client data processing systems having different resource characteristics. Each task includes one or more subtasks, and each subtask has one or more resource requirements. Each of the clients requests from the server a subtask upon occurrence of a first event, for example, an idle machine. The server in response to a request from a client, selects a subtask for execution by the client as a function of the resource requirements of the subtask and the particular resource characteristics of the client.

Abstract: An improved CLB architecture, wherein the use of dedicated AND gates to generate a carry chain input signal facilitates low latency multiplication and makes efficient use of four-input function generators. In one embodiment of the invention, when multiplication using a binary addition tree algorithm is used, AND gates to implement single-bit multiplication are provided within the available function generators and duplicated in a dedicated AND gate accessible outside the corresponding function generator as a carry-chain input signal. In another embodiment, carry chain multiplexers can be selectively configured as AND or OR gates to facilitate certain arithmetic or comparison functions for the outputs of a plurality of function generators.

Abstract: A one-shot circuit provides a pulse on receipt of a first edge, and removes the pulse after a delay generated by a delay chain. However, a second, opposite edge resets the circuit without an intervening delay chain delay. The delay chain can be implemented using a chain of AND circuits (one-shot high) or OR circuits (one-shot low), each driven by the preceding circuit in the chain and by the input signal. In some embodiments, an output circuit includes a pass gate coupled between the one-shot input and output terminals and a pulldown (one-shot high) or pullup (one-shot low) that provides an inactive value when the pulse is not being applied. The pass gate and pullup or pulldown are controlled by the output of the daisy chain. Other embodiments offer programmable capabilities, such as the ability to correct for process shift by altering the effective delay of the delay chain.

Abstract: A low-temperature semiconductor device test apparatus that includes a device tester having a purge box mounted thereon, a low-temperature handler system, and a load board having a IC test socket. The purge box is located between the load board and a support plate of the device tester, and between groups of compressible test pins used to pass test signals to the test socket through conductive traces formed in the load board. The purge box includes rigid outer walls defining a chamber that is located opposite to the test sockets. During low-temperature testing, dry air is pumped into the chamber through conduits formed in the walls of the purge box to prevent the condensation of moisture on conductors formed on the load board and exposed in the chamber. In addition, the purge box resists bending of the load board when semiconductor devices are pressed against the test sockets.

Abstract: Efficient register circuits allow the loading of data values into a memory element using set and reset terminals in addition to loading via the data input terminal. A register circuit includes a memory element and a logical AND gate. A load command input terminal enables the load, and a load value input terminal provides the new value to be loaded. The memory element has set and reset terminals. In one embodiment, the reset function overrides the set function when both terminals provide active signals. The set terminal is coupled to the load command input terminal. The logical AND gate has input terminals coupled to the load command and load value input terminals, and an output terminal coupled to the reset terminal of the memory element. In another embodiment, the set function overrides the reset function, and the signals driving the set and reset terminals are reversed.

Abstract: A scalable physical coding sublayer (PCS) can be adjusted to provide different combinations of communication channels and data widths. The PCS can use 8B/10B encoders having a disparity input connection and at least one disparity output connection. In one embodiment, the encoder has both a synchronous and an asynchronous disparity output connection. The encoder can be coupled with additional encoders to provide an expanded width channel of 16B/20B encoding. Additional configurations are possible. In expanded operation, only one of the encoders needs to output special codes. The encoders, therefore, include a slave input connection to place the encoder in a slave mode so that a special code is replaced with an inert special code. All but one encoder in an expanded system are slave encoders. An idle input connection is also provided in the encoders to place the encoder in an idle mode where pre-defined data is output from the encoder.

Abstract: A configurable logic element (CLE) for a field programmable gate array (FPGA) includes “expanders”, i.e., connectors that allow fast signal communication between logic blocks. Expanders allow the configurable interconnection of a plurality of logic blocks, or portions thereof, to form a single logical entity that can implement large user circuits such as PALs, lookup tables, multiplexers, tristate buffers, and memories. One embodiment includes a configurable logic block. In a first mode, the logic block provides two N-input LUTs having N shared inputs and two separate outputs. The outputs are then combined using an expander to generate an (N+1)-input function. In a second mode, the logic block provides two N-input LUTs having M unshared inputs. An optional third mode provides a plurality of product term output signals based on the values of the N input signals.

Abstract: Interconnecting logic provides connectivity of an embedded fixed logic circuit, or circuits, with programmable logic fabric of a programmable gate array such that the fixed logic circuit functions as an extension of the programmable logic fabric. The interconnecting logic includes interconnecting tiles and may further include interconnecting logic. The interconnecting tiles provide selective connectivity between inputs and/or outputs of the fixed logic circuit and the interconnects of the programmable logic fabric. The interconnecting logic, when included, provides logic circuitry that conditions data transfers between the fixed logic circuit and the programmable logic fabric. The invention is directed towards the various needs and requirements of the layout and floor planning of a device having both fixed logic circuitry and programmable logic circuitry.

Abstract: Methods of optimizing the use of routing resources in programmable logic devices (PLDs) to minimize test time. A set of routing resources is identified that are not used in most designs, and a device model is provided to the user that prevents the use of these resources. Because the routing resources will never be used, they need not be tested by the PLD manufacturer, significantly reducing the test time. For example, each PLD within a PLD family is typically designed using a different number of similar tiles. Thus, smaller PLDs in the family include an unnecessarily large number of routing resources. These excessive routing resources can be disabled during implementation of a design. In another example, each tile along the edges of an array includes routing resources designed primarily to provide access to tiles that are not present. These redundant routing resources can be disabled during implementation of a design.

Abstract: In some communications circuits a phenomenon called duty-cycle distortion—that is, a distortion of the apparent duration of the pulses in clock signals—causes the circuits to read clock signals as having a different duration than intended. Accordingly, the inventors devised unique circuitry for correcting or preventing this distortion. One exemplary circuit uses a voltage divider, comprising a pair of transistors, to set the DC or average voltage of the clock signals input to the digital circuit at a level approximating the logic threshold voltage of the digital circuit. In another example, a feedback circuit drives the DC or average voltage of signals input to the digital circuit to match a reference voltage that is substantially equal to the logic threshold voltage. In both examples, equating the DC or average voltage of the clock signals to the logic threshold voltage of the digital circuit reduces or prevents duty-cycle distortion.

Abstract: Method and apparatus for data sampling is described. More particularly, a data sampling circuit having a delay line and a plurality of tap circuits is used to sample data and provide a vector indicative of a transition region of a sampled input signal. Additionally, a hybrid sampling circuit is described with a fine grain delay line and coarse grain delay lines. Furthermore, a controller is described for using such a vector to control which data samples are used.

Abstract: Described is a method of converting one representation of a circuit into another. For example, a first network representation adapted for use with an FPGA can be converted into a second network representation adapted for use in a mask-programmable gate array. The method begins with accessing the first network representation, such as a netlist, and identifying signal paths that might be sensitive to race conditions. Representations of delay elements are then inserted into each sensitive signal path. The timing of the modified network representation is then modeled by calculating the delays associated with each signal path. Any differences in the modeled delay values are minimized by modifying one or more of the inserted delay-element representations. In one embodiment, the inserted delay-element representations include stopper cells that maintain the nets to and/or from the delay-element representations.

Abstract: Structures and methods for selectively applying a well bias to only those portions of a PLD where such a bias is necessary or desirable, e.g., applying a positive well bias to transistors on critical paths within a user's design. A substrate for an integrated circuit includes a plurality of wells, each of which can be independently and programmably biased with the same or a different well bias voltage. In one embodiment, FPGA implementation software automatically determines the critical paths and generates a configuration bitstream that enables positive well biasing only for the transistors participating in the critical paths, or only for programmable logic elements (e.g., CLBs or lookup tables) containing those transistors. In another embodiment, negative well biasing is selectively applied to reduce leakage current.

Abstract: A programmable capacitor in an integrated circuit (IC) comprises a conductive line located parallel to an interconnect. When a bias voltage is applied to the conductive line, a parasitic capacitance is created between the interconnect and the conductive line. By properly sizing and locating the conductive line, a desired capacitance can be coupled to the interconnect. A bias control circuit can apply or remove the bias voltage from the conductive line, thereby enabling the capacitance to be coupled or decoupled, respectively, from the interconnect. Because of its simple construction, multiple capacitive structures can be formed around a single interconnect to provide capacitive adjustment capability. By changing the number of conductive lines to which the bias voltage is applied, the total capacitance provided by the multiple capacitive structures can be varied. A feedback loop can be incorporated to provide adjustment during IC operation.

Abstract: Disclosed are circuits and methods of identifying defective memory cells among rows and columns of memory cells. In one embodiment, all the memory cells in an array are programmed to conduct with a conventional read voltage applied and not to conduct with a conventional read-inhibit voltage applied. Any rows that conduct with the read-inhibit voltage applied are termed “leaky,” and are defective. Another read-inhibit voltage lower than the conventional level is selected to cause even leaky cells not to conduct. This test read-inhibit voltage is consecutively applied to each row under test. If one of the rows includes a leaky bit, that bit will conduct with the conventional read-inhibit voltage applied but will not conduct with the test read-inhibit voltage applied. The test flow therefore identifies a row as including a leaky bit when a leak is suppressed by application of the test read-inhibit voltage. A redundant row can be provided to replace a row having a leaky bit.

Abstract: A method of generating a test bit pattern for a memory device is provided. The method includes, for example, the steps of loading a data register with an initial test bit pattern and storing the initial test bit pattern in the memory device. The method also includes the steps of generating a additional test bit patterns by shifting the initial test bit pattern by a predetermined number of bits and storing the additional test bit pattern in the memory device. The step of shifting the initial test bit pattern includes, for example, the step of pushing a one or two-bit pattern into the initial test bit pattern. Subsequent successive test bit patterns are similarly generated by pushing a one or two-bit pattern into the previously generated test bit patterns. Hence, the number of bits loaded into the data register is greatly reduced and the required test bit pattern still generated.

Abstract: Methods of detecting shorts affecting nets of a specified design in a partially defective PLD. The nets participating in the design are identified, along with the interconnect lines used to implement each net. The nets are then divided into two or more groups, where no two nets in a single group can be shorted together by the inadvertent enablement of a single programmable interconnect point between two interconnect lines. The groups are then tested for inadvertent shorts. According to a first aspect of the invention, each group is tested sequentially against all interconnect lines not in the group, or against all nets in other groups. According to another aspect, the groups are tested simultaneously by applying a different stimulus pattern to each group. By comparing a detected value pattern to the stimulus patterns applied to other groups, it can be determined which two groups are participating in the short.

Abstract: Described are mask-alignment detection structures that measure both the direction and extent of misalignment between layers of an integrated circuit using resistive elements for which resistance varies with misalignment in one dimension. Measurements in accordance with the invention are relatively insensitive to process variations, and the structures using to take these measurements can be formed along with other features on an integrated circuit using standard processes. One embodiment of the invention may be used to measure misalignment between two conductive layers. Other embodiments measure misalignment between diffusion regions and conductors and between diffusion regions and windows through which other diffusion regions are to be formed. A circuit in accordance with one embodiment includes row and column decoders for independently selecting mask-alignment detection structures to reduce the number of test terminals required to implement the detection structures.

Abstract: An interface structure includes first and second portions. The first portion has physical dimensions that are compatible with the docking area of an associated device tester, and includes a first socket configured to receive a first BGA package. The second portion, which is adjacent to and contiguous with the first portion, extends laterally beyond the docking area of the device tester to provide additional testing area that may include one or more additional sockets. In one embodiment, the second portion includes a second socket configured to receive a second BGA package, wherein the second size and configuration of second BGA package are different from the size and configuration of the first BGA package.