Abstract: An automated method for periodically evaluating media streams on a network of computers. The invention is used to determine the availability of various media streams and identify various characteristics of each stream. By repeatedly obtaining the addresses, attempting to establish communication with each media stream, and then reporting the results the system can verify that each media stream is accessible and report various characteristics of the stream.

Abstract: A memory architecture including a memory cache which uses a single level of write buffering in combination with page mode writes to attain zero wait state operation for most memory accesses by a microprocessor. By the use of such a memory architecture, the speed advantages of more expensive buffering schemes, such as FIFO buffering, are obtained using less complex designs. The memory architecture utilizes same page detection logic and latching circuitry and takes advantage of a feature built into industry standard dynamic RAMs, namely page mode writes, to perform writes to memory which allow the processor to be freed before the write is completed for the most frequently occurring type of write operations.

Abstract: A memory system where a cache miss is fielded with a retry access to main memory, but instead of waiting for the microprocessor to resynchronize and re-initiate the memory cycle, the memory cycle is started immediately. The speed of the tag array is specified so that the status of the cache, hit or miss, is known at the same time that the microprocessor's memory cycle start signal is known to be valid. The addresses are then latched and the memory cycle is started in anticipation of the retried cycle. The access time of memory is then overlapped with microprocessor resynchronization and memory cycle reinitialization. By using this technique, clock cycles are needed for the initial cycle, additional clock cycles are needed to perform the resynchronization, and additional clock cycles are needed for the retried cycle since the data is already waiting from memory. The above-described improvement is implemented by decoupling the direct connection of the memory array from the address bus.

Abstract: A circuit for allowing a clock of any specified duty cycle to be created from a clock of the same frequency using standard digital delay lines. In particular, an EXOR function is implemented to generate a clock signal having a frequency which is twice the frequency of its input signals by using standard logic components such that the active branch for each input edge has an independent path to the output signal. In this manner, if a time delay is introduced into the active branch and only the active branch, the corresponding output edge and only that edge will be delayed by a like amount. Over a complete cycle of the input waveforms, four output edges are produced (two clock cycles). By varying the delay on the input branches, these output edges can be placed independently and arbitrarily within the period. As such, an output waveform having any desired duty cycle can be created independent of the phase relationship between the two input waveforms.

Abstract: A clock circuit for generating two clock signals, one (CLK) having stretched clock phases on a cycle by cycle basis, and the second (2X CLK) being a clock signal having a frequency twice the frequency of the first clock signal which is phase and edge coherent with the first clock signal, including the stretched clock phases. The circuit inputs a signal generated by an oscillator which is twice the frequency of the CLK signal which is then used to generate the CLK signal for use by a microprocessor, either phase of which can be stretched on demand, while the second 2X CLK signal remains phase coherent with the microprocessor CLK signal.