Jurrunio

If it starts then there's no problem. I dont know if your card is a Gigabyte or an Asus but for my Asus card, that LED just means the PCIe cable is charged if it stays lit. It will flash when there's a problem instead.

Use HWinfo64, it has its own start menu, choose to start in sensor mode.

But you're using Afterburner for temperatures right? If there's only 1 CPU temperature value it's not the actual core temperature reported by the CPU. The latter being what the thermal throttling mechanisms follow. What you're reading is typically an averaged out value, but temperature difference between cores can be big especially when the thermal paste is not in good condition. Being a laptop doesnt help either, not all behaviours follow factory settings. Alao note that most GPUs boost a lot less at 90C than say, 80C. It's not thermal throttling since it's still above base clock, but you lose a lot of performance nonetheless.

It's just thermal throttling heavy enough to be noticed. Not fry.

There's not much reason to overclock Zen 4 because they are pretty maxed out from the factory, undervolt instead. Or if you do want more performance, focus on tuning up your memory sticks.

I think you could also bypass vbios certification with commands? Sorry not familiar with bios modding

i.e. he sold you a prototype. This is most likely one of the older Gaming 7 design that was then deemed not good enough. As for the boot loop, are you using a newer LGA2066 CPU? Since it's a prototype, it likely only supports the first bunch of X299 CPUs released like the i9-7900X. Q2'17 ones are more likely to work than Q3'17 ones, pretty sure not newer X299 CPUs than these

Model ID is 10DE 1C82, you have to match that with any BIOS you want to put in https://www.techpowerup.com/vgabios/198509/198509 and yes it's marked as a fake and is apprently Fermi https://forums.developer.nvidia.com/t/gtx-1050-ti-pci-id-10de-1c82-unsupported-by-430-34/77773/2 Ofc the easiest and most reliable way to tell if it's a fake is to remove the cooler and read the etchings on the GPU die.

DLSS does give extra performance at low resolutions. My guess is your resolution is already so low that even at native resolution the 3080 is getting bottlenecked by other parts.

the power supply won't hold it, it should be 750w minimum. Also refer to the PSU tier list for at least tier B units

https://www.techpowerup.com/review/samsung-980-pro-1-tb-ssd/6.html But 4TB SSDs cost so much on their own, RAID 0 of multiple smaller SSDs can become better options. However they also demand systems with more PCIe lanes so it's a tradeoff.

probably just residue of manufacturing processes, like how new cars smell too. Just keep your room ventilated.

This does not explain why 3pin fans on 4pin header runs max speed regardless and 4pin fans on 4pin header can appear to twitch at very low fan speed %. PWM is actually a "if there is power on the 4th pin, cut power" design, not lower voltage. You're describing fan speed control on 3pin headers.

Your direction feels wrong, when a CPU overheats hard it usually means the cooler is failing its job completely. Say the pump of an AIO died, or the thermal paste has dried out. Doubling fan speeds of already not-slow fans would not help enough to turn a "really freaking hot" CPU into a cool CPU, at best it will last a bit longer before crashing. PWM controls fans based on %, regardless of the actual max RPM of the fans. 3pin fans on a 4pin connector will just run 100% all the time. If you don't care about noise, honestly I don't see that as a reason to not try max fan speeds all the time first and see if the problem goes away. 3000RPM fans are just as loud at similar RPMs to slower fans anyways.

Maybe DOCP isnt turned on, so memory runs slowly? But really, don't try troubleshooting remotely when the one on-site isn't knowledgeable. Something obvious like an OS full of bloatware (if not malware) could also cause problems like this for example.

Depends on the display you will use and games you will play.

It's a minor issue, I think that guy makes it sound worse than it really is. But then of course, why accept small issues with the design when it's no cheaper than its rivals?

Here you go

Cuz in Hwinfo, I think it shows GPU core power separately from the rest of the components. AMD cards don't use the shunt resistor on power connector method that Nvidia uses.

that's not HWinfo?

number 3 (mostly), they compress user's own overclocking headroom.

Better start new, but if you're good with worse graphics and 30fps I dont get why this system wouldnt' do?

Stop using it then At least stop it from using background data would reduce it to some degree

More likely some software or bloatware drawing excess data. You could choose to block data from apps when they are not actively in use

Pulse seems to have a smaller cooler. Not so sure about pure and nitro+, particularly the thickness.