ExalyThor

"what games you play?, what's your system?, yada, yada, yada" in the latest and greatest games my newly purchased 5700xt can only do 60fps at 1080p high settings (Dying Light 2, A Plague Tale Requiem etc). I like playing older games (2015-2018) and in those I either get to choose a 120+ fps experience at 1080p or a 80-100 fps experience at 1440p. I would like the 1440p route for older games more but I sacrifice on the new games front so 1080p feels like the wiser (and cheaper) choice. I need your subjective opinions on this.

the 13700k is basically a 12900k

I looked at the peerless assassin and stumbled upon the deepcool ak620. That thing is within 2c of the dark rock pro 4 and the nh d15 at 260w power draw. I like the way it looks too. Might want to edit your comment to include the deepcool cooler for other people though, as I made it the solution.

Why do GPUs still hold their high prices? Mining got unprofitable since summer. The Merge happened, still no hundreds of GPUs flooding the used market. Back in 2018 as soon as mining wasn't profitable the GPUs were sold for pennies. This year they've been on borrowed time for 3 months now.

They also cheat overclocking by having the ring bus unlinked to the core!

I'd say don't mess with pcie power. Replace the resistors from the power cables if you want a higher power limit. I'd go for 2.5mOhm resistors. Doubling the power limit should be enough for pretty much everything, and it's easy to read what gpuz reports. You don't need a calculator for multiplying values by 2.

Prices in different regions differ but where I live GTX 960s got extremely cheap. It's a great GPU if you can find it for $30-35.

A laptop is a smarter idea (not because of the bad conditions, but because you can relocate it instead of relocating you and your bike to watch something on a fixed screen). Computers can live in much harsher conditions than people give them credit, but for this type of workload a 4th gen i3 should be plenty and cheap enough to not care if the system goes bust in 5-10 years.

That benchmark is so light on the gpu it is considered a cpu test as well. What you see there is nvidia's driver overhead limit cpu performance. If you want a real test, use unigine superposition. You'll see the difference there.

I'm a mechanic, you work at a computer. Who's the bigger nerd?

Did you try nuking all driver remains with ddu? Do that for both amd and nvidia drivers. If that doesn't work, try flashing to the latest bios available for your gpu. If that doesn't work, rma.

I'm a mechanic, you work at a computer. Who's the bigger nerd?

You already purchased it so asking whether it's worth using or not is irrelevant. Follow this link 
to Gamers Nexus' video about the contact frame, it starts directly at the guide on how to install it.

Sitting here, wanting to upgrade to a 1080.

I ran an MSI 980ti 6g from 2019 to 2022 without problems, and I had the thing bios modded with a 350w power limit. Ran it overclocked 24/7 and it never gave me a headache. Custom 980tis are fine.

Some boards have external clock generators which enable you to change the base clock of CPUs.

There is no point in disabling the shared memory pool. The GPU will never use system memory unless the 2GB available on board is not enough. The shared memory pool is there because when a game needs more than 2GB of VRAM, it overflows into RAM so you may still run the game (albeit at a reduced framerate) rather than having missing textures or straight up having the game crash on you.

All I can say is back up your data and do a fresh install. Even if you get rid of it, the registry would still remain broken.

I reached platinum 4 in lol =))

Read about skylake bclk oc. It seems you know very little about the subject.

New GPUs have come back in stock at (albeit inflated) MSRP so why used GPUs haven't come down in price? I sold my 980ti in March for $250 and in June I can only buy a 1080 for that money. I expected 5700s and 2070s to be selling for $250 by now. In 2018 used GPUs plummeted in price following the crash in mining profitability of that time.

If you reduce die size by 35% and power consumption by only 21% you end up with a higher heat density and therefore a harder chip to cool.

I've googled a bit and it seems like multi gpu compatible boards use a chip which splits the lanes. This chip wasn't designed with pcie 3.0 in mind, so using a 3770k won't enable 3.0 speeds. Funnily enough, motherboards which have a straight connection from the cpu to the pcie slot do work at 3.0 speeds if a compatible cpu is installed.
 
I don't worry about the high end gpus, my 2600k would bottleneck something like a 3080 even at 4k. I'm worried about mid range gpus like the 6600xt, which the old sandy chip is still able to handle, but pcie 2.0 x8 is a bottleneck.
 
Also ik the 3770k is a bad joke. Whatever minuscule ipc gains it holds over sandy, it loses in raw clock speeds when overclocked, not to mention one needs delidding to clock beyond 4ghz while the other is ice cold even at 5ghz. Pcie 3.0 was the only appealing thing to ivy.

I only need my question answered so I can make my own decisions based on that. But to answer YOUR question, I currently have no GPU.

Other apps were within 0.005v off the real voltage. Muh HWinfo was overshooting by over 0.01v, and it's fluctuation range was much higher as well. Which one would you deem more reliable based on those findings?