SPRUI30H November 2015 – May 2024 DRA745 , DRA746 , DRA750 , DRA756
DPS and SLM are similar concepts: they consist of switching the system between high- and low-power consumption modes. However, their operating timescales differ, principally in the latency allowed for mode transitions.
DPS is generally used in an applicative context (tasks are started). Therefore, mode transitions are related to system performance requirements or the processor load. DPS transition latencies must be small (typically between 10 µs and 100 µs) compared to the time constraints or deadlines of the application so that they do not degrade application performance. DPS requires performance prediction to ensure that transition latencies do not deteriorate device performance to the point that real-time application deadlines are missed or the user experience degrades too much for an interactive application.
SLM is not used in an applicative context (no task started). Mode transitions are related more to system responsiveness, and the transition latencies must be small compared to user sensitivity so that they do not degrade the user experience. For SLM, transition latencies are typically 1–10 ms or more.
DPS and SLM also differ in the type of wake-up event used to exit low-power idle mode. For DPS, wake-up events are application-related (timer, DMA request, peripheral interrupt, etc.); for SLM, wake-up events are user-related (touch screen, key pressed, peripheral connections, etc.).