Abstract: An execution unit (2) interface protocol allowing flow-through of data, where a function is specified once and the execution unit performs the function for multiple sets of input data. Function execution is pipelined through the execution unit, where an input unit (6) stores information, while a function logic unit (4) processes data and an output unit (8) holds results to be output. The execution unit (2) allows for data rate distortion, in applications such as data compression, where the amount of data received is different from the amount of data generated as output.

Abstract: A multiply and accumulate (MAC) unit (52) having a multiplier (54) and an adder (56) for providing a calculated result of a MAC operation, an accumulator (58) for storing the calculated result, and a scaler unit (60) for selecting a subset of the calculated result stored in the accumulator. The scaler unit receives information defining the subset as a predetermined number of contiguous bits and defining the location of the bits with respect to a radix point.

Abstract: Method for maintaining an execution interval for a task requestor to a DMA. A timer is provided with two counters, one (34) to maintain the execution interval and the second (32) to track the execution time of a task in the DMA. Each task has a predetermined execution time allowance. A task acknowledge (TACK) signal enables the tracking. A task request signal (TREQ) is generated during each execution interval until the execution time allowance is completed. The length of the second counter is less than the first counter. In one embodiment, if the first counter expires before the execution time allowance is completed, a task error signal (TERR) is illustrated.

Abstract: Direct memory access controller (DMA) (2) adapted to directly execute C language style FOR tasks, where the FOR task includes a movement of a data element from a first location to a second location in memory, and the movement is controlled by a master DMA engine (MDE) (6). A master DMA engine (MDE) (6) includes a top level state machine (52) to coordinate a context save state machine (54), a parse state machine (56), and a running state machine (58). An loop control descriptor (LCD) queue (74) and a data routing descriptor (DRD) cache store information. The LCD queue allows pipelining of descriptor parsing, while the DRD cache avoids refetching of DRDs on reentry of loops.