Here are latest results of RTOS benchmarks. Chart below displays relative performance of benchmarked operating systems. RTOS which scored highest score in any given test is set as reference for that given test. Score of other systems is then expressed as fraction of highest score.
The chart above displays relative score of various real-time operating systems in several benchmark tests. The RTOS in batch with highest score achieved in given test is chosen as a reference for each benchmark test and was assigned relative performance of 100. Each other RTOS score was then compared to this highest score in that particular test. So if the score is 92, it means that particular system reached 92% of the score of fastest RTOS in that test. Comparing of scores across tests is meaningless as different tests do test different behavior.
Table below summarizes actual numeric score of real-time operating systems achieved in the benchmark. These numbers are verbatim. Score in these benchmarks records number of repetitions each of tests achieved in 30 seconds. Thus higher number is better.
| Test Name | FreeRTOS 10.3.1 | CMRX RTOS 0.1.2 | Zephyr 4.1.0 stack protected | Zephyr 4.1.0 |
|---|---|---|---|---|
| Calibration | 265819464 | 264299737 | 266669300 | 266669379 |
| Message Passing | 4689994 | 1473332 | 11538364 | 11538368 |
| Synchronization | 7183628 | 4925995 | 19199961 | 19199974 |
| Cooperative Scheduling | 17044259 | 2316773 | 7272652 | 13043448 |
| Preemptive Scheduling | 4357890 | 1265385 | 3007385 | 5419805 |
The benchmark used to obtain the data is “thread metric” benchmark. It tests various RTOS subsystems. The goal of these benchmarks is to find out what is the baseline overhead of subsystems rather than test kernels in real-world scenarios.
You can read more on what this benchmark measures and what hardware was produced on this page.