### Top runs - **Highest FPS:** `scx_bpfland-bpfland-next-f80123c`. This run achieved the highest average FPS among all runs. - **Smoothest FPS:** `scx_bpfland-bpfland-next-9955777`. This run had the lowest standard deviation and variance, indicating the most consistent performance. - **Best overall:** `scx_bpfland-bpfland-next-f80123c`. Despite its higher standard deviation compared to `scx_bpfland-bpfland-next-9955777`, its significant increase in average FPS justifies it being the best overall. ### Issues - `scx_bpfland-main-a72c905` had a significantly lower average FPS compared to other runs under the same benchmark conditions, despite using the same hardware and software. The lowest FPS and low1percent FPS values were much worse than others, indicating instability or inefficiency in this specific commit. ### Summary Among the four runs benchmarked for Mirror's Edge under stress, `scx_bpfland-bpfland-next-f80123c` stood out with the highest FPS and was deemed the best overall, although its stability was slightly compromised compared to `scx_bpfland-bpfland-next-9955777`, which had the smoothest performance. In contrast, `scx_bpfland-main-a72c905` exhibited significant issues with FPS, resulting in lower performance compared to others. The other two runs, `scx_bpfland-bpfland-next-9955777` and `scx_bpfland-bpfland-next-d1064a8`, showed similar results in terms of performance, although the former was slightly more consistent. Overall, the `bpfland-next` branch appears to be better optimized under load compared to the `main` branch.