What are those red arrows supposed to be pointing at? At the interp bars? That means you've got interp.

Oh no, now we need a null reloading script. It's turtles all the way down.

Because it would break 20 year old configs.

You're right about rate and ratio being forced but cl_interp_ratio was supposed to replace cl_interp so it doesn't scale cl_interp, lerp is just max(cl_interp, cl_interp_ratio / cl_updaterate). Updaterate 66 and ratio 1.2 results in 0.01818... like it should but it's kind of pointless since interp is already set anyway and can't be forced to a lower/higher value via server vars.

If it's on a soundcard and it's actually the physical socket that's damaged it might be fixable, on a mobo or if it's the chips then I'd just buy a new soundcard.

Have you considered that the socket or soundcard is broken and not the headset? That should be easy to check with literally any headphones/earbuds.

#33
Great idea, the FX-8370 is identical to the FX-8350 except for the 0.1 GHz higher half-load turbo so you'll get 0 to 2.4% higher fps with no downside.

I don't know why I have to spell this out for your but guess what, it depends on which R9 200 series GPU you've got.
If you've got a 290X then big surprise, a 280X will not be worth it. If you've got a 270 it might be. If you've got a 280X, well...

The PCIe slot size hasn't changed since 2004 so any GPU will fit.

Streaming and TF2 are all CPU. Your CPU is ok-ish for streaming but not for TF2.

#28
No, I really don't know what you're trying to say.

#29
https://www.teamfortress.tv/7598/tf2-benchmarks
Download link is broken but it's still somewhere in the thread and any demo will do as long as you both use the same one.

Uhm what?

#26
Have someone with a similar or the same CPU (maybe viper) run timedemo on the same demo with the same settings and compare the results. Adjust for clockrate if necessary.

It's highly unlikely that your SSD or HDD cause stuttering, but you can read out the S.M.A.R.T data and check if they are broken.
The mobo is also unlikely.
RAM can be checked with memtest86.

Before buying a new mobo try reinstalling windows.

That's also wrong though. Ryzen is not better than Intel for x264.
More threads are better for x264 and there are some gaps in the Coffee Lake lineup where Ryzen is significantly cheaper for the same number of threads (e.g. 6/12 starting at 199$ vs 303$) due to most Intel SKUs not having SMT but Ryzen is not strictly better. 9900K vs 2700 I know which one is going to win. Of course one of them costs half as much as the other.

No one is forcing you to pay for a K-CPU, Z-mobo and cooler for overclocking. If you don't want to just don't. If you want to then yes, Intel is much more expensive, but don't act like you'd do a balls-to-the-wall OC on a 2700X with the stock cooler and 60$ B350 mobo.

#544/#547
You don't even know how electromigration works.
Also the IMC officially supports 1.35V, just like the CPUs before that supported 1.5V officially and the CPUs after only support 1.2V officially. I have seen exactly 0 CPUs die from running officially only 1.5V standard DDR3 controllers at 1.65V and I have also seen exactly 0 CPUs die from running 1.2V DDR4 controllers at 1.35V.

This is in no way "astonishingly dangerous". It's overclocking and voids the warranty like all overclocking, nothing more.

You are also confusing VCore and the IMC apparently. Changing the IMC voltage doesn't change VCore and doesn't lead to throttling.
If it was throttling due to the power limit why would the mobo "scream for help"? It's still the exact same amount of power as before.

I did not say that. Saying Intel is generally faster than AMD for TF2 is like saying Ford is generally faster than Chevy in curves, you can't generalize that over so many years and dozens of models because it's not even clear which Intel SKU you're comparing with which AMD SKU. It's mostly about single threaded performance. So clockrates and microarchitectures.
At 2 GHz it's going to run like shit no matter what. On Bulldozer (FX-xxxx) it'll run like shit even on 5 GHz. On an old Core 2 it'll also run like shit. On a single core it'll run like shit but you can't buy those anymore anyway and if you buy an old one then just like Bulldozer (2011) or Core 2 (even older) you deserve what you get for buying an ancient CPU. On 2 cores it'll run ok, beyond 2 cores + SMT or 3 cores you gain nothing.

End result, just don't buy old CPUs, neither Zen nor Skylake (current AMD/Intel architectures) suck, get >3 GHz (which already rules out 2 cores without SMT), ideally close to 4 GHz or if you can afford overclocking you can go beyond that but either way you'll get a decent CPU.

Yes.

That would take a bit too long and some of it is outdated.

Any recent CPU would require DDR4 so CPU + mobo + RAM is the only option.

Honestly if you don't need a CPU upgrade just buy a used LGA1150 mobo.

If you want to upgrade then we need to figure out how much of an upgrade you need and I'll need to look at Russian prices. There's a massive difference between mobo + RAM costing 150$ and having enough money left for a 200$ CPU and everything being 20-30% more expensive and only being able to buy a 100$ CPU.

More or less expected in TF2.
You should do it properly though and use a bit lower timings as well. TF2 cares about those too.
Also tREFI is not a performance setting. Do not mess with it unless you're actually doing proper long tests afterwards to verify that you're not losing data.