Benchmark #1801

Download
Nine Sols: BORE vs SCX

GPU Limited static scene + 1080p60 Discord Stream with Software Encoding | 30 second frame recording with 0ms interval | Performance Governor/EPP | Proton GE 10.17 Wayland Native | SCX Commit: 4089bc8 | kernel-cachyos-lto-6.17.0-cachyos1.lto.fc44.x86_64 with Staging SCX Patches.

Submitted 1 month ago by summedkibbles36

Specifications
Label OS GPU CPU RAM OS specific
BORE Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_beerland Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_bpfland -m all Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_cosmos -m all -c 0 -p 0 Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_flash -m all -C 0 -c 0 Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_lavd --performance Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
scx_p2dq --task-slice true -f Steam Runtime 3 (sniper) AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 7 7800X3D 8-Core Processor 33 GB 6.17.0-cachyos1.lto.fc44.x86_64 performance
# Top runs: * **Highest FPS**: `scx_cosmos -m all -c 0 -p 0` with approximately 40% higher mean FPS than the lowest performing run, showcasing exceptional frame rate performance. * **Smoothest FPS**: `scx_lavd --performance`, which exhibits the lowest standard deviation of FPS among all runs, indicating superior stability with around 70% less variability compared to the most variable run. * **Best overall**: `scx_cosmos -m all -c 0 -p 0`, offering a balanced performance of very high FPS and relatively smooth frame time, though not the lowest variation, it was accompanied by strong FPS improvements. # Issues: * The BORE configuration experiences roughly 30% lower mean FPS compared to the highest FPS run, likely indicating a performance disadvantage within the specific setup or configuration compared to others in the GPU-limited benchmark. # Summary This benchmark was designed to evaluate the performance differences between the traditional `BORE` configuration and various `sched_ext (scx)` schedulers, using a GPU-limited scenario at 1080p with software encoding. All runs shared the same core hardware setup and Linux environment, but varied in the scheduler used. It was evident that the different `scx` implementations generally outperformed `BORE`, with `scx_cosmos` emerging as a top choice due to its high FPS and acceptable stability. On the other hand, `BORE` noticeably lagged behind in FPS performance, highlighting its limitations in this context. The results suggest that certain `scx` configurations can offer superior performance for gaming under similar conditions.