# Top runs: * **Highest FPS**: `BPFLAND Next --slice-us 1000` achieves the highest average FPS, providing a robust overall performance considering both stability and FPS. * **Smoothest FPS**: `BPFLAND --slice-us 1000` has the lowest FPS standard deviation and variance, indicating the most stable frame output. * **Best overall**: `BPFLAND Next --slice-us 1000` balances high FPS with respectable smoothness, making it the best overall choice. # Issues: * The standard `BPFLAND` configuration shows a significantly higher standard deviation and variance in FPS compared to other runs, which might result in less stable performance under similar conditions, despite using the same hardware and software configurations. * The standard `BPFLAND Next` configuration resulted in higher GPU temperature than others and slightly increased frame time variance, indicating potential pushes towards thermal limits without performance benefits. # Summary This benchmark compares the `sched_ext` schedulers `BPFLAND` and its successor `BPFLAND Next` under identical hardware and software conditions, focused on a CPU-limited scene. Across multiple configurations, `BPFLAND Next --slice-us 1000` outperforms others in both mean FPS and consistency, demonstrating its ability to leverage the sliced scheduling strategy. In contrast, the basic `BPFLAND` configuration, although similar in average FPS, suffers from noticeable frame rate volatility. Overall, the findings suggest that `BPFLAND Next --slice-us 1000` is the superior option in delivering high and steady FPS in CPU-intensive gaming scenarios.