# Top runs: * **Highest FPS**: `BORE` delivers the highest average FPS, outperforming others by approximately 3% in terms of mean FPS. * **Smoothest FPS**: `fair-k` exhibits the smoothest FPS with the lowest standard deviation, resulting in 20% less variation compared to the next smoothest performer. * **Best overall**: `lavd` is the best overall, balancing high FPS with stability, showing only slight variability in FPS while maintaining strong performance. # Issues: * There is a significant FPS drop in the non-slice adjusted `bpfland` runs, with FPS 15% lower on average compared to the high-performing configurations like `BORE`. * `fair` shows notably higher FPS variance, over 30% more than the smoothest run, which could affect user experience. # Summary The benchmark focuses on evaluating different BPF scheduler configurations under stress conditions with osu!stable. The `BORE` configuration emerges as the top performer in terms of FPS, exploiting higher clock speeds and consistent power usage to achieve better frame rates. Meanwhile, both `lavd` and `fair-k` schedulers achieve a good balance of FPS and stability, though `fair-k` excels in smoothness with minimal FPS variation. Conversely, `bpfland` falters in this setup, likely due to higher default slice settings compared to the modified configurations, which also suggests optimization opportunities with slice size adjustments.