How does the eTPU achieve such fine resolution?
The key feature of the eTPU is its proactive approach to handling events and the resulting fine resolution achieved for waveform generation and measurement. The eTPU consists of a single microsequencer and dedicated channel hardware (there are 16 channels in the TPU and 64 total channels in the eTPU). The microsequencer prepares the channel-specific hardware to react to possible future events. When a particular event actually occurs, the hardware autonomously performs some action and requests microsequencer servicing. Because the hardware is autonomous, the usual latency penalties associated with an undedicated microsequencer are not incurred. The microsequencer responds to the service request (after servicing any existing requests based on an optimized scheduling mechanism) and prepares the channel hardware for the next event. A helpful example of this mechanism is found in the off-the-shelf frequency-to-digital conversion routine. For a particular eTPU channel the microsequencer arms
