CPU: AMD Ryzen 5 5500
GPU: Nvidia RTX 3070
RAM: 16 GB DDR4 3200MT/s
By @_kod
(Note: These benchmark graphs are only meant to be compared within themselves. Do not compare them with other graphs on this page)
The above graphs show how much performance is affected by different presets and also GPU/CPU utilization
High Quality was unlocked by Spoofing the GPU to 4070 Super. This does not affect the performance.
FPS Limit was unlocked by Spoofing the GPU and using Frame generation. This way the fps is unlocked but Frame gen is not actually active because its a 3070.
Flight Challenges was used for the benchmarks because they provide a dynamic world loading and need no additional user inputs.
Presets affect how far the objects are rendered. I did some calculations based on how far a house is properly rendered and my current co-ordinates
High Quality: 130%
Quality: 100%
Balanced: 80%
Performance: 65%
Ultra Performance: 45%
CPU: AMD Ryzen 7 9800x3D
GPU: Nvidia RTX 3070 Ti
RAM: 32GB DDR5
Thanks to @setsugekka.infinitia (Discord)
(Note: These benchmark graphs are only meant to be compared within themselves. Do not compare them with other graphs on this page)
The above graphs show how much performance is affected by different Ray Tracing Quality and also GPU/CPU utilization
Flight Challenges was used for the benchmarks because they provide a dynamic world loading and need no additional user inputs.
We ran the benchmarks multiple times and by multiple people and found that RT Low vs Medium vs High had minimal impact on performance. In some of our runs, RT High beats out Low and Medium. We don't know why this happens.
Benchmarks were done by me (@_kod), @setsugekka.infinitia and @argon1ut and all had the similar results in RT with three different devices where RT Low vs Medium vs High has no difference or were within the margin of error.
Thanks to @setsugekka.infinitia and @argon1ut for providing additonal benchmark data.