Abstract: Methods and systems for producing random numbers include a random number generator with a first port and a second port. The first port is configured to receive a first type of random data request, and the random number generator is configured to generate first random data while the first type of request is asserted on the first port. The second port is configured to receive a second type of random data request, and the random number generator is configured to generate only a specified length of second random data in response to receiving the second type of request on the second port. An embodiment of a system also includes a data structure configured to store multiple random values, which are derived from the first random data generated by the random number generator in response to the first type of random data request.

Abstract: Embodiments include bitstring generators and methods of their operation. A sampling parameter of the bitstring generator is set to a current value, and values of one or more bits are then repeatedly sampled based on the current value of the sampling parameter. The repeated sampling results in a set of test bits, which is analyzed to determine a randomness measurement associated with the set of test bits. A determination is made whether the randomness measurement meets a criterion. If not, the current value of the sampling parameter is changed to a different value that corresponds to a lower probability of being able to correctly predict the values of the one or more bits produced by the bitstring generator. The steps of repeatedly sampling, analyzing the set of test bits, and determining whether the randomness measurement meets the criteria are then repeated.

Abstract: Embodiments of methods and systems for producing random values include a first module that provides a random data request (e.g., a request for an unspecified length of random data) to a random number generator. The random number generator generates random data in response to the random data request, and multiple random values derived from the random data are stored in a buffer. In response to receiving a request for a random value (e.g., an initialization vector), the first module produces the random value based on the multiple random values stored in the buffer. The system also may be configured to receive requests for other types of random values, and to fulfill those requests using random data that is not buffered (e.g., random data that is received directly from the random number generator in response to a request for a specified length of random data).

Abstract: Embodiments include bitstring generators and methods of their operation. A sampling parameter of the bitstring generator is set to a current value, and values of one or more bits are then repeatedly sampled based on the current value of the sampling parameter. The repeated sampling results in a set of test bits, which is analyzed to determine a randomness measurement associated with the set of test bits. A determination is made whether the randomness measurement meets a criterion. If not, the current value of the sampling parameter is changed to a different value that corresponds to a lower probability of being able to correctly predict the values of the one or more bits produced by the bitstring generator. The steps of repeatedly sampling, analyzing the set of test bits, and determining whether the randomness measurement meets the criteria are then repeated.

Abstract: Embodiments of methods and systems for producing random values include a first module that provides a random data request (e.g., a request for an unspecified length of random data) to a random number generator. The random number generator generates random data in response to the random data request, and multiple random values derived from the random data are stored in a buffer. In response to receiving a request for a random value (e.g., an initialization vector), the first module produces the random value based on the multiple random values stored in the buffer. The system also may be configured to receive requests for other types of random values, and to fulfill those requests using random data that is not buffered (e.g., random data that is received directly from the random number generator in response to a request for a specified length of random data).

Abstract: Methods and systems for producing random numbers include a random number generator with a first port and a second port. The first port is configured to receive a first type of random data request, and the random number generator is configured to generate first random data while the first type of request is asserted on the first port. The second port is configured to receive a second type of random data request, and the random number generator is configured to generate only a specified length of second random data in response to receiving the second type of request on the second port. An embodiment of a system also includes a data structure configured to store multiple random values, which are derived from the first random data generated by the random number generator in response to the first type of random data request.

Abstract: A digital bus is driven to the logic state of a data input signal upon activating a data enable signal. A bus keeper enable signal activates a buffer having its input and output connected to the digital bus. The data is thus buffered and driven back onto the digital bus during the active state of the bus keeper enable signal in order to retain the logic state after the data enable signal deactivates. The bus keeper enable signal remains active until a subsequent data enable signal becomes active thereby retaining the data on the bus potentially indefinitely.

Abstract: A computer system includes an error detection and correction system for detecting and correcting single-bit errors, two-adjacent-bit errors, and four-adjacent-bit errors. Two identical error detection and correction (EDC) circuits are connected to the system memory array, and each EDC circuit is connected to half of the data bits in alternating pairs. Each EDC circuit detects single-bit errors and two-adjacent-bit errors. The EDC circuits are connected to alternating pairs of data bits so that errors of up to four adjacent bits are actually detected and corrected, two bits by the first EDC circuit and two bits by the second. Thus, if one of the x4 DRAMs in a memory array fails, each erroneous data bit from the DRAM is corrected to its original value, and the failure of the DRAM is registered.

Abstract: A clock buffer circuit for a computer system, and a computer system incorporating the same, are disclosed. The clock buffer circuit includes a differential input buffer for receiving the input clock signal, with its output coupled to the input of a phase locked loop (PLL). The switching level of the differential input buffer is adjustable, either by adjusting the DC bias applied to the input clock signal, or by adjusting the reference signal, which changes the point in the cycle of the input clock signal at which the differential buffer switches. The PLL synchronizes its output to an edge of the output of the differential buffer, but maintains the same duty cycle (e.g., 50%). Accordingly, the clock buffer circuit may have its delay adjusted, by modifying a voltage divider, applying a variable voltage, or programmably via a digital-to-analog converter, to match the delays of other clock buffer circuits in the computer system, reducing the clock skew in the system.

Abstract: A formatter for combining timing signals with data from an algorithmic pattern generator (APG). In the disclosed embodiment, the formatter receives address signals from an APG and timing signals from a timing unit. Each timing signal from the timing unit corresponds to an address signal received from the APG. The address and timing signals are communicated to a signal select unit wherein each timing signal selects its corresponding APG address signal. The selected APG address signal then is communicated to a signal combining unit which combines the APG address with its corresponding timing signal and generates leading edge and trailing edge marker pulses having a user-specified pulse width.

Abstract: An adder and a comparator form portions of a modular memory address block determination circuit. The starting address of the first block and the enable signal of the first block are added to produce the starting address of the second block. This procedure is repeated for each block. The determined starting address for each block is compared with the requested memory address and, unless the block is utilized or disabled, if equal a signal indicates the match. The circuit is used on a circuit board which emulates three conventionally separate memory circuit boards. The registers for each emulated board are provided and appropriate bus signals are developed.

Abstract: A timing generator for generating timing signals representing the leading and trailing edges of test pulses. In one embodiment of the invention, a period circuit repetitively measures time intervals, or periods, based on signals from a clock circuit, and a marker circuit generates timing signals representing leading edge and trailing edge markers precisely within each period. The period circuit comprises a period-end memory having a plurality of storage locations which are addressed by a modulo(n) counter. To support multiple timing sets, or timing cycles, one or more of the most significant bits of the address field for the period-end memory may be reserved for designating each timing signal. The time interval measured by the period-end memory may be selectively extended by delaying the clocking signal used for incrementing the modulo(n) counter.