Benchmark #470

Download
scx 1.0.3-git.20240904.7dede2d vs 1.0.4-git.20240908.92b887c

Submitted 2 months ago by sacredsin

Specifications
Label OS GPU CPU RAM OS specific
Warhammer 40000 Space Marine 2 - BORE Steam Runtime 3 (sniper) AMD Radeon RX 6600 XT (RADV NAVI23) AMD Ryzen 7 3700X 8-Core Processor 34 GB 6.10.7-200.fsync.fc40.x86_64 powersave
Warhammer 40000 Space Marine 2 - BPFLAND_1.0.3 Steam Runtime 3 (sniper) AMD Radeon RX 6600 XT (RADV NAVI23) AMD Ryzen 7 3700X 8-Core Processor 34 GB 6.10.7-200.fsync.fc40.x86_64 powersave
Warhammer 40000 Space Marine 2 - BPFLAND_1.0.4 Steam Runtime 3 (sniper) AMD Radeon RX 6600 XT (RADV NAVI23) AMD Ryzen 7 3700X 8-Core Processor 34 GB 6.10.7-200.fsync.fc40.x86_64 powersave
Warhammer 40000 Space Marine 2 - LAVD_1.0.3 Steam Runtime 3 (sniper) AMD Radeon RX 6600 XT (RADV NAVI23) AMD Ryzen 7 3700X 8-Core Processor 34 GB 6.10.7-200.fsync.fc40.x86_64 powersave
Warhammer 40000 Space Marine 2 - LAVD_1.0.4 Steam Runtime 3 (sniper) AMD Radeon RX 6600 XT (RADV NAVI23) AMD Ryzen 7 3700X 8-Core Processor 34 GB 6.10.7-200.fsync.fc40.x86_64 powersave
# Top runs: * **Highest FPS**: `Warhammer 40000 Space Marine 2 - BORE` - This run showed the highest average FPS, outperforming the other runs by a significant margin. This suggests better resource utilization and overall performance. * **Smoothest FPS**: `Warhammer 40000 Space Marine 2 - BPFLAND_1.0.3` - This run had the lowest standard deviation in FPS and frametime, indicating consistent and smooth performance throughout the benchmark. * **Best overall**: `Warhammer 40000 Space Marine 2 - BORE` - This run not only achieved the highest FPS but also maintained a competitive standard deviation, providing the best overall combination of performance and stability. # Issues: * The second `BPFLAND` run on version `1.0.4` shows a significantly higher variance and standard deviation in FPS compared to other runs. Additionally, its average FPS is considerably lower, suggesting possible inconsistencies or inefficiencies in the newer version of the scheduler. * The second `LAVD` run on version `1.0.4` also displayed reduced performance and higher FPS variability compared to its `1.0.3` version, indicating a potential regression with the new update. # Summary The benchmark compares different versions of the `BPFLAND` and `LAVD` schedulers, alongside a control run (`BORE`). The data indicates that the `1.0.4` versions of both `BPFLAND` and `LAVD` schedulers generally underperform compared to their respective `1.0.3` versions, showing higher variability and lower average FPS. The `BORE` run outperformed all other configurations, achieving the highest average FPS and maintaining a reasonable level of stability, making it the best choice among the tested scenarios. This suggests that while newer scheduler versions may introduce changes, they do not necessarily guarantee performance improvements.