# Top runs: * **Highest FPS**: `defaults` - This run achieved the highest average FPS. * **Smoothest FPS**: `t 0` - This run has the lowest standard deviation in FPS, indicating the smoothest experience. * **Best overall**: `defaults` - A combination of the highest FPS and relatively low FPS variance makes this the best overall run. # Issues: * The run with the `t 1` argument experienced significantly lower average FPS (10-15% lower) and higher frame times (up to 15% higher) compared to the best overall run, indicating performance issues. * The run with the `t 50` argument showed lower average FPS (almost 15% lower) and higher FPS variance (almost double) compared to the best overall run, suggesting instability. * The run with the `t 2000` argument also had lower average FPS (slightly more than 5% lower) and markedly higher standard deviation in frame times (20% higher), pointing to less consistent performance. # Summary This benchmark focused on testing various `bpfland -t` argument values on a Steam Deck using `Cyberpunk 2077` to identify optimal performance configurations. All configurations used `CachyOS`, the same AMD custom GPU and APU, and the `schedutil` Linux scheduler. The findings indicate that the default settings offered the highest and most consistent FPS, while certain `bpfland -t` values, notably `t 1`, `t 50`, and `t 2000`, resulted in significant performance degradation in terms of both FPS and frame time variance. The highest and smoothest performance measurements emerged from runs without these `-t` modifications.