Benchmark #48

Download
Forza Horizon 5: noesync/nofsync vs defaults vs winesync

Did 3 benchmarks, where esync/fsync is disabled, then enabled (defaults) and lastly used PROTON_USE_WINESYNC=1 with esync/fsync enabled (defaults)

Submitted 4 months ago by erikas2

Specifications
Label OS GPU CPU RAM OS specific
defaults Arch Linux AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 9 3900X 12-Core Processor 34 GB 6.9.8-2-cachyos-lto performance
noescync⁄nofsync Arch Linux AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 9 3900X 12-Core Processor 34 GB 6.9.8-2-cachyos-lto performance
winesync Arch Linux AMD Radeon RX 7900 GRE (RADV NAVI31) AMD Ryzen 9 3900X 12-Core Processor 34 GB 6.9.8-2-cachyos-lto performance
## Top runs - **Highest FPS**: `defaults` has the highest average FPS, approximately 30% higher than `noescync⁄nofsync`. - **Smoothest FPS**: `noescync⁄nofsync` has the smoothest performance with the lowest std.dev. and variance of FPS, roughly 35% more stable than `defaults`. - **Best overall**: `defaults` performs the best overall, combining the highest FPS with relatively low variability, around 60% lesser variance compared to `winesync`. ## Issues - **Noesync/nofsync** displays significant performance degradation, with FPS being approximately 25% lower than the other two runs. All hardware/software settings are consistent across the benchmarks. ## Summary The benchmarks show that enabling `esync/fsync` (`defaults`) offers the highest FPS, approximately 20% better than `winesync` and 30% better than `noescync⁄nofsync`. `noescync⁄nofsync` has the smoothest performance with the lowest fps variability but suffers from a significant FPS downside. The `winesync` run provides better FPS than `noescync⁄nofsync` but isn't as stable as `defaults`. Overall, `defaults` emerges as the top performer both in FPS and stability, but the `noescync⁄nofsync` setting clearly struggles, indicating an issue when `esync/fsync` is not utilized.