KarathKasun

Dead GPU.  RMA if still in warranty.  If its out of warranty, you are out of luck.
 
The only other possible problem is a faulty power supply, but that generally does not cause this kind of artifacting.
 
The "device removed" ding is because the GPU is dropping off the PCIe bus.  The GPU itself is locking up and no longer responding even though windows is semi-gracefully recovering.  If you had a USB video adapter you could probably plug it in and keep using the computer after the GPU is failing, and I would bet money that the GPU is no longer in device manager.

Going to say provider if everything else works.  You can always try setting your DNS server on your phone/pc/laptop to 1.1.1.1 manually.

Dead GPU.  RMA if still in warranty.  If its out of warranty, you are out of luck.
 
The only other possible problem is a faulty power supply, but that generally does not cause this kind of artifacting.
 
The "device removed" ding is because the GPU is dropping off the PCIe bus.  The GPU itself is locking up and no longer responding even though windows is semi-gracefully recovering.  If you had a USB video adapter you could probably plug it in and keep using the computer after the GPU is failing, and I would bet money that the GPU is no longer in device manager.

On intel systems, the PCH has a self contained microcontroller that has ring -2 or -3 access (direct memory access without the CPU being aware of it).  And it is hackable, and cant be wiped once infected.

It sounds like a blow off answer, but that is litterally all you have to do.
 
Generally motherboards will tell you what timings are being used, so you start by manually setting all of the timings to those.  You then bump memory voltage up to 1.45v-1.5v, and up the memory clock by one step.
 
Do some very memory intensive tests, and if all is good, bump it up another step.
 
If you have issues increasing memory speed, try adding one step to a few timings.  Make sure to keep track of which ones you change.
 
Lather, rinse, and repeat.

Manually setting the clock and timings of your memory.

Memory issues.  Is your RAM on the QVL?

Just going to say, install a fresh OS first.

@ConsoledotlogIf you are upgrading your system, keep your GTX 970 for a while to save some cash for the next upgrade. Forza will perform a lot better with a the upgraded CPU and with a GTX 970 you should be able to play the game at 1080p on high detail. 
 
In games like the Forza series, running over 60 FPS is not necessary most of the time. In fact the game may run better when locked/v-synced to a lower frame rate.
 
I run Forza Horizon 4 just fine with high details with my i5-7400 and GTX 970.
Your GTX 970 has more potential than you think, it's just being choked.?

FX anything is going to have a very hard time maintaining 60+ FPS in many games.  Its just too weak per core, even at ~5ghz.
 
In several games I can run into CPU limitations with an FX-8300 @ 4.7ghz core and 2.6ghz IMC, 32gb of DDR3 2133, and an RX 560.  Fallout 4 for example, at 1080p max details, can end up at 75% usage on the RX 560

Ive seen it on a 9900k and it was still choppy.

BL3 doesnt really require much... its the Denuvo VM running that kills performance, and a faster CPU cant get around that.

Slots 1 and 2 connect directly to the CPU.
 
Slots 4 and 5 connect to the chipset.
 
You want to use the CPU connections first for GPUs, and the first slot in this layout is generally the direct to CPU PCIe lanes.  If you use 1 and 2 it gets split to x8/x8.  Likewise, the x16 from the chipset is slot 4 and it gets split to x8/x8 when 4 and 5 are in use.
 
Slot 1 is where the manual says to put the GPU, so that is absolutely where the direct CPU connection is.
 

Yes

Soundblaster Live for anything from Windows 98 to Windows XP.  You can even stretch it into Windows 10 if you are not afraid of community written drivers.
 
For DOS, I would have to say Gravis UltraSound is the gold standard.

The only CPU that makes sense for a general build is the R5 2600/3600.  Pretty much anything else is overpriced for what you are getting unless you absolutely need the extra cores.

The only CPU that makes sense for a general build is the R5 2600/3600.  Pretty much anything else is overpriced for what you are getting unless you absolutely need the extra cores.

Go for Ryzen as there are bound to be new Intel vulnerabilities found lol

Latest Prime95 supports AVX-512. If your CPU supports that, it is WAY hotter than AVX2.
 
Userland software should never be able to damage hardware unless the hardware was inadequate. Specifically the only risk is for those doing extreme overclocks and inadequate cooling.

AFAIK its the MB that usually goes belly up and takes the CPU with it.
 
The OP though, they likely just have a clogged block or dead pump.  Not the first dead H100i that I have seen.
 
This is why I refuse to use a CLC on any of my builds.  More parts to break and if the pump goes south its not easy to replace at all.  While you can just swap fans if you get a large air cooler with standard fan sizes.

Quad core without hyperthreading will present a problem for newer games, regardless of the GPU.

Onset depends on a number of factors.  Panel temp is one of them.
 
I would assume that in an iMac the panel temp is much higher.
 
I had installed quality IPS displays at a doctors office that were on for 8-12 hours on the same static image without problems, for 12 years.  By the same token, I have purchased mid range TN panels for home use that have randomly developed image retention.  The worst offenders are pre-2000 TN panels though, and they usually ran pretty hot.  Had a Dell laptop from 97 with burn in so bad it looked like a 70s arcade cabinets CRT.

Wat?

On LCD's it gets worse over time, its material fatigue.  I have some very old high usage monitors that will burn in from as little as 5 minutes of displaying a static image now.  You can clear it fairly easily, but it is annoying.

Can all you base are belong to us?