Project: Vulcanus Configuration, Testing & Benchmarks
02-02-25
Verified All components installed correctly and displayed in BIOS
Windows 11 Installation and Configuration
Drivers Installed
BIOS updated
Set EXPO profile for DDR5 RAM to 6000MHz
03-02-25
3D Mark Benchmarking/Stresstesting Tool and Cinebench deployed -
Note: Vulcanus is operating with AMD Expo enabled but all results compared to similar systems worldwide with CPU Ryzen 7 9800X3D and GPU Nvidia 4070 Ti at stock and overclocked.
CPU Profile
Max Threads vs
Worldwide Highest score: 11056
Worldwide Average score: 10059
Tritonis scored within the top 31% of similar systems worldwide
1 Threads vs
Highest score: 1357
Average score: 1211
Tritonis scored within the top 42% of similar systems worldwide
Additional Threads
2 Threads: Tritonis scored within the bottom 25% of similar systems worldwide
4 Threads: Tritonis scored within the bottom 12% of similar systems worldwide
8 Threads: Tritonis scored within the top 47% of similar systems worldwide
16 Threads: Tritonis scored within the top 35% of similar systems worldwide
Maximum CPU Temperature recorded: 70C
Time Spy Extreme
Time Spy Extreme is a 4K version of the Time Spy DirectX 12 benchmark
Vulcanus ranks in the top 48% of benchmarks worldwide
Max temp: 65C
This is a -2.8% decrese from Tritonis' last benchmark. This is due to Time Spy not accounting for the Ryzen 3D cache correctly.
Port Royal
Port Royal is a Real-time ray tracing benchmark @ 2560 x 1440
Vulcanus ranks in the top 55% of tests worldwide
Max Temp recorded: 40C
This is a 0.2% performance increase over Tritonis Last Benchmark
Time Spy Extreme Stress Test
Time Spy Stress Test runs Time Spy Extreme continuously for 20 loops without pausing or reloading between loops. A full test lasts around 20 minutes.
To pass this test a system must achieve frame rate stability of 97%
Vulcanus achieved a frame rate stability of 98.5% with a 67 FPS average and GPU temp of approx 72C
04-02-25
Owing to the fact that Vulcanus will use Tritonis' GPU and the majority of 3D-Mark static benchmarks do not take the 3d cache of the Ryzen CPU into account, it was deemed better to examine a number of real world benchmarks to guage any performance changes between the platforms.
Tritonis' benchmarks presented here were performed on 27th - 29th January before its refit.
Bethesda's Indiana Jones requires hardware raytracing by default. My normal settings are within the margin of error.
Increasing to Medium path tracing in balanced upscaling demonstrates the power of Ryzen's 3D cache.
At full path tracing with frame generation and Ultra performance upscaling the results are with the margin of error.
Bethesda's Starfield shows a modest uplift in average and 1% lows.
Black Myth Wukong static benchmark shows a significan decreate in 0.1% lows when raytracing is on.
Ryzen 3D is of significant benefit to Sony's Horizon: Zero Dawn performance smoothness
With raytracing off there is no difference between the processors.
Nacon's Robocop: Rogue City seems to favour Intel
Ubisoft's Far Cry 6 favours AMD so this is expected
Ubisort's AC Valhalla also favours AMD but not as much as Far Cry 6
Cyberpunk 2077 is a mixed bag and a difficult to benchmark but at the highest playable settings 1% lows have taken a significant hit.
Larian's BG3 gets a significant average FPS uplift (I play capped at 90FPS for stability, but Ryzen can push it further on these settings)
Significant improvement to EA's Dragon Age:The Veilguard smoothness
Sony's Spider-Man Remastered showed one of the largest average FPS increases but as the cost of possible increased stuttering.
Sony's GOW shows a 0.1% low increase increasing the smoothness overall.
and Sony's TLOU shows a 0.1% low increase also.
The Callisto Protocol benchmark ran far better on the Ryzen 7 which is not surprising as an AMD sponsored title.
Square Enix's FFXIV has traditionally favored Intel but the increase in 1% lows is significant.
Note that the 1% lows are likely to be the benchmark loading different scenarios during a 3 minute run which is not the way other benchmarks are done (and not what the benchmark is actually designed to measure).
15 Game Average
10% increase in Average Frames Per Second
6.5% increase in 1% lows
12.5% increase on 0.1% Lows
Result: Success
05-02-25
Testing complete - System Commissioned 05-02-25 @ 21:00