Cycle-Resolution Timing Enhances Embedded-System Performance

The right real-time operating system (RTOS) is crucial for optimizing embedded system performance, reliability, and power efficiency.

- By Bianca Schmidt, Technical Writer at SEGGER Microcontroller

Cycle-Resolution Timing is an advanced scheduling technique used in SEGGER’s RTOS embOS-Ultra that offers significant benefits in terms of precision, efficiency, and power consumption.This article shows how Cycle-Resolution Timing works, how it compares to RTOSs using traditional methods, and the advantages it offers for embedded-system development.

KNOW YOUR EMBEDDED-SYSTEM ARCHITECTURE

Before diving into the specifics of Cycle-Resolution Timing, it is important to understand how traditional RTOSs work.

Bare-metal and super-loop architecture: In many simple embedded systems, developers use a bare-metal and/or super-loop architecture, where a system continuously runs a in a loop, checking for events and executing tasks. This approach is simple, but it has limitations:

Task management is carried out manually, which can quickly become complex as the system grows.

Power consumption is high because the system runs continuously, even when no tasks are scheduled.

Debugging and scaling are difficult when tasks become more complex or time sensitive.

While this approach is sufficient for small systems, it doesn’t provide the level of precision or energy efficiency needed in more complex or resource-constrained embedded systems.

Traditional RTOS approaches: A typical RTOS improves on the super-loop model by managing tasks with system-tick interrupts. These interrupts, usually set at regular intervals (e.g., every millisecond), serve as a time base for scheduling tasks. While this approach allows for multitasking, it still has limitations:

• System ticks occur regardless of whether a task is pending, which can waste processor time and increase power consumption.

• Time-based tasks must be scheduled based on the system tick intervals, which can reduce precision in task execution.