# Top runs: * **Highest FPS**: `bore` with a mean FPS approximately 6% higher than the second-best run. This run demonstrates the highest performance. * **Smoothest FPS**: `lavd-no-core-compact` with a standard deviation in FPS about 4% lower compared to the other smooth runs. This provides the most consistent frame rates. * **Best overall**: `bore` as it has the highest FPS and a very competitive standard deviation. # Issues: * `bpfland` and `rusty` have FPS variances roughly double compared to all other runs, leading to substantial instability. * `rusty` has about 10% higher CPU load variance, causing inconsistent performance. * The GPU clock variance in `bpfland` is roughly 15% higher, indicating potential performance spikes and drops. # Summary This benchmark compares different scheduler optimizations using `CachyOS`, an AMD Radeon GPU, and an AMD Ryzen CPU running the `osu!stable` game and `ffmpeg`. The `bore` scheduler stands out for delivering the highest and most stable FPS overall. In contrast, the `bpfland` and `rusty` schedulers show significant performance inconsistencies, primarily due to higher variance in FPS and CPU load. The `lavd-no-core-compact` scheduler provides the smoothest overall gameplay with the lowest standard deviation in FPS. These insights can help in selecting the appropriate scheduler when prioritizing either peak performance or frame stability.