How can the RTX 3080 achieve so much different scores in the 1080 Preset test?

[pcservis]

Here are examples from https://gpuscore.top/furmark/

1.P1080    15916    NVIDIA GeForce RTX 3080/PCIe/SSE2    2022-11-13
2.P1080    20106    NVIDIA GeForce RTX 3080/PCIe/SSE2    2022-11-13
3.P1080    17212    NVIDIA GeForce RTX 3080/PCIe/SSE2    2022-11-13
4.P1080    19622    NVIDIA GeForce RTX 3080 Ti/PCIe/SSE2    2022-11-13

How can there be nearly 4100 poens difference in scores (1. and 2. examples) for same model RTX 3080 graphics cards ? Second and fourth examples are close and it is understandable, but differences of this two compared with 2. and 3. scores is unbeliavble.

Any explanation ?

[Stefan]

15542 https://gpuscore.top/furmark/show.php?id=801739
17636 https://gpuscore.top/furmark/show.php?id=801804

• low scores because of power saving activated ?
• CPU or  driver version should make very little difference
• Issue in latest Windows version - i have no idea how much impact that has https://learn.microsoft.com/en-us/windows/release-health/status-windows-11-22h2#lower-than-expected-performance-in-some-games
• high score because of xtreme overclocking ?

Last but not least you can cheat on benchmarks:
force wireframe mode
manipulate Windows timers
I would not expect that with FurMark though.