FaceFoiled

Yeah definitely try just pointing the light right at you to start with, that should help some.

Beyond that you can definitely upgrade your camera, maybe move the light to the right so its on the middle/right side of your right monitor to get a more across the face light that will give you some shadow and depth in your face for the camera

-> Moved to Photography and Videography
 
As in people who understand lights.

Yeah. This is part of the reason why I tried undervolting it a bit, as I figured that might help keep temps down. 
 
I might look into "proper" water cooling at some point, but not for now. Not too happy with the AIO I put in there for my CPU. Cooling is fine, but software for it does not work and would like to monitor things a bit more. So an overall "real" water cooling solution might be something I'd look into down the line. Happy with where it is at now though, will see if I get an urge to boost it more down the line. 😄 

That is a pretty good graphics score, ~1800 points higher than average for those cards. 
 
Undervolting is only really a good idea with those cards when thermally or power limited. If you aren't thermal or power limited, those cards do scale to about 1.25-1.3V on air. That said, you are in the high 60s for temps according to the 3DMark readout, that's probably high enough to make more voltage cause instability (my card personally needs to stay in the mid 50s in order to be stable at those frequencies). The Gaming Z you've got is a good card, it is one of the best 6950 XTs you can get, it's just not the best cooler for a 6950 XT you can get. If you want to try pushing it further, you need to pull the temps down, either by maxing out the fans, putting the card on water, or just pointing an AC unit at it. That said, that's a good score, and if you want to stop there I wouldn't blame you. 

Keep me posted, just kinda curious how far you can end up pushing that card. 

1000 points ain't bad. Restricting voltage likely won't get you anywhere with that card though, they scale with more voltage, not less. If you stick below 1.25V on VCore, you should be pretty OK, you're just running benchmarks so there isn't really much risk doing so. I just wouldn't daily anything above 1.25V

I've used MPT to give it more power. I tried the manual way first, but could not get it to "stick". Re-reading:
I don't think I did this when doing it the manual way first. Likely why.. 😛 Oh well.
 
 
So far I've been getting this on average: https://www.3dmark.com/3dm/78634306?
This is with the power set to 800 (does not really go above 480 when doing my 3dmark tests), and Min/Max Frequency of the GPU set to 2619 - 2719MHz and undervolting to 1150mV. I've messed around with VRAM Tuning as well, but not been getting much in regards to better scores.
 
I've been able to mess a bit boosting frequencies and such, but with this setup I get easily over 23500. The best score I had was 23751 I think. Happy with that. 🙂
 

Have you re-enabled PBO? If you've done a Clear CMOS, the auto overclock from ASUS should be completely gone, but PBO (the auto overclock from AMD that actually works) is disabled as well. It might also be a good idea to setup custom PBO limits, pretty sure PPT 300, TDC 177, EDC 177 is more or less optimal from what I remember, but I also don't have a 5950X to play with so can't tell you for certain. 
 
Also, I assume that you're referring to Time Spy Extreme, not Time Spy, since the CPU benchmarks for them are pretty different (Time Spy vanilla maxes out at 16 threads, so the solution to getting more performance from it is to disable hyper threading, Time Spy Extreme has no limits). 10-11K is what I'd expect for that CPU in Time Spy Extreme, in normal Time Spy I'd expect closer to 14K. 

Holy crap that's low. In Extreme (what I mainly run) my 5900X does 8500 with just PBO (and admittedly very good memory settings, Time Spy really scales with high speed RAM). Do you think you could run HWiNFO64 in the background and take a log when running Time Spy, that should record all your sensors and hopefully give some insight into what's going on (I'd expect the chip to be either hitting limits really hard, thermal throttling for some reason, or clock stretching a ton for whatever reason, and that should hopefully show exactly what's going on). 
 
It would also be a decent idea to download ZenTimings and send a screenshot. Time Spy scales a lot with memory speeds (the difference between 3200MHz CL16-18-18-38 and 3800MHz CL14-13-13-21 is 1000 points in Time Spy Extreme in my testing) so if there's some memory settings that the board happened to misconfigure that could be contributing to the overall poor performance, and that program reads every relevant memory setting. That won't be the main factor for the bad score, but it might be a factor in it. 
 
What board are you running, and what settings can't you find? Those settings may or may not be relevant for why performance is so bad. 
 
As for the limits and stuff, PBO boosts adjusts stuff within certain limits. It's a bit weird though since it sweet spots hard a lot lower than where a lot of boards set the limits manually. I'm guessing if you've got a higher end board (like I'd kinda assume you do if you've got a 5950X and 6950XT) the limits will be by default something like 450 PPT, 300 TDC, 400 EDC, much higher than what's optimal, and counterintuitively limiting the CPU more makes performance better. Those limits should be closer to within the ballpark of optimal than by default, hence why I said run those. 

Yeah, that required disabling Hyperthreading and manually locking the CPU to 4.8GHz with really good memory settings, it's not super easy to get. That said, you should be closer to 14000 points than you should be to 10000. 
 
That link is private. 

Wait a minute, I just realized you're running a 5800X3D. That makes a bit more sense now, I could've sworn you wrote 5950X in there at some point, I'm guessing I just went full dyslexia and mixed up 6950 XT with 5950X. that score makes a lot more sense now. 
 
Ok, so it's still likely a misconfigured PBO setup (the X3D is pretty much immune to memory settings as long as the Infinity fabric is clocked well enough). Unfortunately, I don't know where the limits are optimal on a 5800X3D, so I will need to do a bit of searching to find that out first. I'll go through the data though when I get home and see for certain what's going on. 

OK, finally had a chance to look at it, the CPU is just not boosting in the slightest. If you look, it's running at a constant 3.6GHz, with no cores going above that at all. With PBO enabled it should be averaging closer to 4.5GHz or higher, 3.6GHz is base clock. Temps look fine, so it should be running a lot higher than it currently is. That explains the score.
 
The question now is why is it running at just 3.6GHz? First thing I'd check is in the BIOS to make sure that the multiplier is set to auto, as well as a couple other settings like turbo boost (or whatever it's called in AMD boards, I don't touch that setting often enough to know by heart). I'll be looking through the log some more to try and see a reasoning behind that, but start looking through some of the BIOS settings. It might not be a bad idea to just do a Clear CMOS and start all over again. 

Nice!
 
If you want to go over a bit more, 6900 XTs/6950 XTs overclock really far if you do it right. Unlocking the power limits and overvolting them (they are one of the few GPUs that actually scale with voltage at ambient) can be done entirely within the OS. On my 6900 XT (admittedly one with pretty bad silicon quality in comparison, whether that will affect this up or down) I'm able to get over 1000 points higher in Time Spy Extreme GPU score by doing that, though since the 6950 XT has a higher out of the box power limit I wouldn't expect as much from that as I would from a 6900 XT (it still should be a fair bit more though, Time Spy in particular gets these cards to pull ridiculous power figures). That's mostly just for bragging rights though, going for 21000 points or more isn't really that meaningful. Still, if you want to do it (if you find this type of thing fun, for example), I can find the couple of tutorials for how to do those things to get a fairly decent performance improvement. 

OK, for the power limits there are two options: manually or MorePowerTool. If you want to do it manually, this is the tutorial I used
If you want to do it with MorePowerTool (this is how you overvolt the card, so if you want to overvolt the card you might just want to do this). You start off by going to TechPowerUp's website and downloading a 6950 XT vBIOS (it doesn't really matter which one, you can get one for the MSI card you've got, but you can use an XFX or PowerColor BIOS as well just fine, all it changes is what a couple of the initial power limits are). Then download MorePowerTool (MPT), and run it. In the dropdown along the top, select the 6950 XT you've got, then click "Load" and select the vBIOS you downloaded earlier. Go to "Power" and change the Power Limit and TDC limits so something kinda high (800 for both works just fine) and should let the cards draw as much power as they want. Hit Write SPPT and you're done. 
 
To overvolt the card, do roughly the same thing with MPT at first, then follow this tutorial and write the new SPPT
 
Also, you will need to reboot every time you write a SPPT for it to take effect (technically you only need to restart the AMD driver, but it's usually easier to just do a system restart)

I have some work ahead of me, thanks for the resources! 🙂 

Thanks for checking in, that's one of the few things I did remember to do after the bios default/update :D. That would be DOCP for my CPU/Motherboard.

Got my new GPU aswell!
Very happy with it.
Benchmark: TIME SPY
CPU: AMD Ryzen 7 5800X 12-Core Processor
GPU: AMD Radeon RX 6950 XT
3DMark Link: https://www.3dmark.com/3dm/75830493?locale=en_GB
Score: 21564 Total , 23869 Graphics, 13939 CPU
 
 
 

Aye it's running nicely so far, though playing games that don't really need it .:P That's beside the point though! 😄
 
Will need to look into cooling though, it's getting mighty toasty.

After waiting for a very long time, I finally purchased a 3080ti at good value!
 
Figured my current CPU ( AMD 5600x) would be bottlenecked, so splashed for the new AMD7 5800X3D as well (should be delivered today!).
 
It seems it was all for naught, as the GPU turned out to be a bust. Despite ordering it from Amazon directly (not a retailer on amazon), I was delivered a 2080TI.. or so I thought. Thought it was an honest mistake at first, rang up to have it sorted out. Replacement not available, sold out - return to refund. So I get to packing it up to refund - huh that's odd, the sticker on the side does say 3080ti.. let's open it up just in case!
 
Hmm, this is not a GPU at all.. at least I have a new CPU now?
 
Slightly gutted.
(wearing shorts in the picture, apologies for my hairy knee :D)

Thank you so much for the time investigated so far, and the information shared. 
For the NVME's, I initially had 2 different "options" as I was not sure which would be better. Is there a reason why you decided to go with 2 different options here?

Good to hear it was just a loose connection. It'd really suck if it was a break along the wire.
 
To utilize STP the shielding needs to be in contact with the shielded RJ-45 connector which you would usually crimp down on the metal jacket then both the PC, Switch, or Router needs a metal interface with proper grounding. That's how STP works.
 
Without actually looking it up to verify I believe Cat6 will do 10Gbit up to 50m. Shielding may or may not help that since crosstalk is also a factor in that.

No, I don't have another computer or router unfortunately. I have added a switch in between though. PC to Switch = 1000mbps. But actual speeds are limited to 100mbps.
 
I'm currently crimping the ends of the leftover of that Cat6 I have, and will see if I can run that in-house between the rooms to test. Sadly not really in a position to move PC around (WFH, I need the PC online :P). I'm guessing I'll have to replace the Cat6 cable, but hope to find other solutions as removing the current and re-running it again will cost my quite a bit of time and money.
 
T568B. I tried A as well to see if that somehow made a difference (did not). Both ends are the same yes. I can confirm this with the tester I use as well, as it sends a signal through each wire in the cable. So it shows "Sending signal 1" and at the other end "receiving signal 1". That is repeated through all 8 wires (so easy to spot if a wire a connector is not crimped properly, or if the ends are mismatched (in which case it could be something like, "sending signal 1" and at the other end "receiving signal 4".

Id say try anouther cable, or get a premade cable, or just run again. There isn't much else you can do without a proper test(and those aren't cheap)
 
What pattern are you crimping with? you sure its exctly the same on both ends?

Have you tried connecting a different computer on on end and either a different switch/router or computer to the other?
 
It's hard to say what the problem is if we know all 8 wires are good and the network interfaces are known working at gigabit.

If you do router -> switch -> Cat6 to PC what happens?
 
I have to assume the signal is getting lost somewhere. If you opted for STP are both ends grounded? If not it defeats the purpose (not that this does anything for gigabit unless you routed it along lengths of powerline.
 
Improper terminations can cause problems but usually at gigabit you don't see them. Problems tend to crop up when going 10gig so your situation is a unique one. We've seen people have problems with 100mbit/1gbit negotiation before but they never verify it isn't a interface issue before posting so here I'm at a bit of a loss outside trying different hardware.