Question Strange benchmarks

[Axial Tilt]

I recently bought an MSI SUPRIM X GeForce RTX 4090, and while using MSI afterburner, I was able to get a stable overclock of around 2985mhz (+190) on the core clock. here is a link to my userbenchmark results:
https://www.userbenchmark.com/UserRun/58178563
The weird thing is that when I try to benchmark in 3d mark using time spy extreme, (my main monitor is a 1440p LG ultragear 2019 so that might have something to do with it but I am not sure), I am at the bottom 21%. I have already stress tested in kombustor for 6 hours so it's not a stability issue, but the same thing happened for a previous Asus TUF 4080 I bought and returned because it was in the bottom 0% percentile so I just have no idea what is going on.
here is my time spy results:
https://www.3dmark.com/spy/34454913
Could it be a software issue or do I just have the worst luck when it comes to the Silicon lottery?

 

[Aeacus]

I am at the bottom 21%.

You aren't actually that far off, considering that you compare your rig against another rig, which too has 11700K CPU.

Comparison: https://www.3dmark.com/compare/spy/34454913/spy/33302725#

As seen from comparison, you vs #1 with same CPU and GPU, you're not that far off. #1 spot has a bit higher base clock and memory clock on GPU, than you do. And they also have newer GPU drivers. Oh, 64 GB of RAM as well.

So, all-in-all, i'd say your system performs good.

Overall #1 spot, in TimeSpy Extreme, is using Threadripper and LN2 cooled 4090,
comparison: https://www.3dmark.com/compare/spy/34454913/spy/33302725/spy/32329815#

Just look the insane base/memory clocks the Threadripper guy got, while GPU temps are -35C. <- This is cryogenic fluid cooling.

 

[Axial Tilt]

So you're saying I'm just a bit below average when comparing my rig to identical ones?

 

[Aeacus]

I'd say average, since the metric goes from 15000-13000, and you have 14100.

but the same thing happened for a previous Asus TUF 4080 I bought and returned because it was in the bottom 0% percentile

I think you even didn't need to return your 4080, since for your CPU/GPU combo, the metric is from 12200-10600.

Though, it seems that you compare against TimeSpy, while you're benching TimeSpy Extreme. TimeSpy scores are ~double of what TimeSpy Extreme scores are. With this, it's easy to see being part of 0%, if you bench TimeSpy Extreme but compare it against TimeSpy normal scores, with 13900K CPU.

 

[Axial Tilt]

Well, that's a relief. I was so worried I keep getting unlucky with my GPUs. Thanks!

 

[sizzling]

So you're saying I'm just a bit below average when comparing my rig to identical ones?

Quite normal. The user’s overclocking their hardware hit these benchmark tools very hard with multiple runs. It skews the numbers upwards.