Additionally, measuring the longest-running thread won't accurately reflect total time consumed if multiple threads are time-sharing the same CPUs, especially when the number of threads exceeds the number of available CPUs. This is why I believe wtime() remains the most appropriate function for performance measurement in our case.

From my latest observations, it seems plausible that Setonix may be employing time-sharing instead of running threads on multiple CPU cores.

If you test it on your own Setonix account, you might get different results.