Maybe a damp cotton tip would work? But you might make things worse, since you won't be able to see what you are doing. You could also take the DVD drive apart and clean the lens that way. Obviously, don't look at (or in the direction of) the laser when the drive is plugged into anything, they can cause serious eye damage. So if you do take it apart, make sure it is fully unplugged from everything.
Maybe test other discs and if those don't work, get new drive, they are quite cheap and die at times.
But if you are at the point where you are going to buy a new one, at least give cleaning it a try. Its a very common issue, and is cheap and easy to fix.
Best you can get is an A10/A12 if you're on a FM2/FM2+ motherboard. If you have an AM4 you better get a Ryzen instead.
Even if you were to upgrade it, you may find it might not work due to power/thermal constraints, especially if its in a laptop.
In which case no one can tell you until you find out what motherboard it is. If its a prebuilt, the model number might be enough.
This much of CPU load on Intel is enough to drop Turbo clocks from 5Ghz to 4.3Ghz, in practice lowest turbo clock is which 8th gen CPU runs at, while 6th gen keeps higher turbo clocks far longer.
Not sure which CPU you have, but I think Intel uses different clock speeds for different cores on their more recent processors. So you might have one or two cores that will keep a higher clock speed while the rest drop down. AFAIK they also lower clock speeds at low loads on modern CPU's in order to be more efficient. So maybe something to do with that also? I know that my CPU supports speed stepping, but BSOD's (or just locks up completely) with it enabled, so it spends it's days at a consistent 4.2ghz.
i7-8086K and i7-6700, but it really reliably drops clocks to 4300 with any sight of CPU load No BSODs here, just needs overclock to get more than 4300 out from it in practice. I can have for example 1 core load 5ghz, 2 core load 4.8Ghz and all core load 4.6Ghz at 1.136V stable, 4.3Ghz this CPU does with around 1.05V without any stability issues. At low load CPU runs higher clocks and yeah it is funny how cores get constantly bit different speeds. It just is so funny how with stock clocks performance is there in theory, looks nice in benchmarks, but really it just is 4.3Ghz only or even 4.0Ghz only with default settings that motherboard applies as CPU hits power limits constantly with default settings, needs a lot of voltage lowering. Sure, with overclocking performance is there in reality and even with stock setup BeamNG runs mostly fine. There are video recording issues though, so for those need bit of overclock. I have to use manual voltage though for overclock, offset and adaptive just don't stay in sensible range. Weak VRM limits how high I can go though, while it does handle 100W, there is not going much beyond that. Theoretical maximum according to Buildzoid is 125W that this VRM can handle and that needs fans for VRM, almost all Z370 motherboards are like that or worse, only highest tier can be better, so don't get one used. 5Ghz on all cores is possible, but heat gets too much as well as VRM load, which is bit shame, would need a delid and motherboard that costs over 400 euros. Gamergull's traffic scenarios run nicely at 60fps though, just video ends up bit choppy and it seems overclocking helps with that, even very mild one seems to help a lot, like 300Mhz it is not even 10% increase, but seems to work.
It does of course look nice on the spec sheet. But realistically, with such a dynamic system, it really is impossible to give an exact clock speed. Even an average clock speed wouldn't work, since the speeds will be different for different load types. But ofc, the benchmarks don't care what is written on the spec sheet. The CPU will give results according to how it scales its cores for the given work load.
Kinda, issue is that for example single core benchmark stresses just that, single core, but in practice while benchmark runs at 5Ghz while gaming you get only 4.3Ghz, so it can give rather skewed picture of real performance. That is then where overclock is needed to get even close to that benchmark performance, but luckily CPU can do much higher clocks than what it has at stock. I would think that more realistic results could be seen when clock speeds would be limited to what they are in practical loads and that might actually narrow differences between CPU brands quite a bit or then AMD also has similar turbo clock dropoff. But all those high clocks are pretty much marketing and advertised performance is available only to those who overclock. 4 generations older hardware does pretty much the same for 95% of gamers. What I think is that far too long have been stuck to such small die surface area as well as TDP rating, I have been told that Zen2 addresses these aspects, so it will be interesting to see if it can hold higher clocks better, that would lead to higher practical performance, even if performance on benchmarks would not be quite as much higher. Something to keep in mind as Zen2 comes available.
But why would you look at single core benchmark results, if your intended use case was purely for multithreaded workloads? Moreover, there are CPU gaming benchmarks, if the purpose of a CPU was going to be playing games, aren't those the benchmarks that one should be looking at? * Not to mention that comparing clock speeds is mostly pointless anyway. Since different CPU architectures will take a different amount of ticks to process an instruction. If a 10Ghz CPU takes 100 ticks to do an operation, it will be half as fast as a 5Ghz CPU that takes 25 ticks. This becomes super apparent when comparing mobile phone processors to desktop x86 based processors. Then of course there are other factors such as speculative processing that can use idle CPU ticks before a branching statement has been decided to process all outcomes before one is selected. Traditionally Intel has had a more performant solution to this than AMD, however, that has of course bitten both of them quite hard recently, Intel more so than AMD. *For reference, my overclocked I5 2500k can beat many brand new processors in single core applications, especially >8 core CPU's. But that doesn't mean it is a faster CPU for gaming (unless you are playing retro games).
Because nothing else matters as much as single core performance in games, like BeamNG. If game uses more than 1 thread, it does not mean it would not be single core limited, this has been discussed to death many times over, with my 12 threads only 1 is limiting factor, most of the time it is enough though, but not always. ETS2, also drops below 60fps are because of single core performance. Gaming benchmarks are generally pretty much junk, they would need to be done whole lot different ways that they would be much use. 4th gen intel and 8th gen intel, there is not much improvement in terms of IPC, sure there is slight improvement, but still not too much. Now someone chimes in and says has not noticed such single core limits, well with gtx1080 and 1080p there really is not many games that would max it out, only reasons fps drops are because of CPU, when having something like 1060, one can't keep up 60fps at max settings anyway, GPU limits before single core performance limit becomes apparent, but with gtx1080 or faster it becomes quite apparent how especially at higher framerates modern CPUs just are not enough in single core performance to feed GPU. Oh, pretty much any game runs 60fps, but do video recording, go above 60fps and single thread limits start to manifest themselves. It is highly dependent from settings and GPU etc. Disabling two cores and pretty much no performance difference. It just works that way, you can't thread everything and even best games in their genre use DX9 or DX11, which are very much relying in single core performance, despite using many threads.
I would very much disagree with you on gaming benchmarks. If your goal is to play games, reading/viewing an article that has benchmarked a bunch of games is the best way to find the gaming performance of a processor. They are real world tests, conducted using real world games, they perfectly illustrate what a CPU is capable of when combined with the best current hardware. They can also be made a bit more synthetic by dropping the resolution down to 1080p or 720p etc in order to really highlight where each different CPU tested becomes a bottleneck. Modern games are heavily multithreaded. However, the approaches that they take to multithreading are different to other types of software. Some games will target a certain number of threads (typically 8 these days due to consoles). Some games will fire off a new thread for basically anything and everything, others will take a more large scale approach. Many games have their own schedulers now to provide the best multi-threaded experience. I don't play much ETS2, but I do know that they are a small company, and afaik they are using their own engine, which I would assume has its roots over a decade ago, that is likely why it doesn't scale very well. BeamNG is also an outlier, since there are very few games that need a full core to simulate a single object at 2000 simulation ticks per second within the game world. All of that said, if you have a GTX1080, why not use ShadowPlay? The quality is quite adequate, and it doesn't cause the CPU issues.
Gaming benchmarks don't work. Sure, they give indication of that specific game for those specific settings and that is all they give, none of game I play seem to be on those tests so bunch of useless data, you can of course disagree, but you do so mostly because you are thinking games featured in benchmarks, when I don't even look at those games Shadowplay requires Nividia bloat to be installed and has lot of issues. OBS does not have CPU issue, it does consume small percentage of CPU power like any other software, using nvenc is same way using GPU like shadowplay and I think shadowplay uses CPU too, especially when doing audioprocessing and video overlay with Facerig etc. JC3 is perhaps most modern AAA game I sometimes play, perhaps only one of mine featured in benchmarks, above 60fps it becomes very quickly single thread limited, keep it at 60fps and all is nice and fine though, perfectly balanced 8 thread load. Many sites then are totally clueless about single thread CPU limits, I can read stuff like "it was unclear why XXX did so bad in this test". Quality sometimes leaves lot to be desired. I have seen GPU benchmarks where they run overclocked CPU and are heavily CPU limited still, but indicate in text how they think their overclocked CPU will not hold GPU back. GPU utilization not monitored at all. Sure there are better sites, with more understanding, but there are so much totally horrible bad quality reviews that it brings tears to eyes. Dropping resolution, that is bad practice, surprise surprise, going up in resolution does add to CPU load, so going low resolution skews things. It is not so simple and more one learns about performance testing, more difficult it becomes. Oh and BeamNG needs only half a core to simulate that single object at 2000Hz, they did implement that few versions back I think, but graphics are what is hard for CPU, there is only so little room in single thread and bit of everything likes to tap on that, but also it does allow things, but also it does limit things However, cool thing at this era is that we are finally having CPUs that pretty much all have decent multicore capability, that is not limiting many CPU models anymore and we are very near of time when single core power is becoming plenty for these challenging games too. Year ago, I would of not dreamed to have 8 cars (7 in traffic + own car) running in BeamNG, like I can now have in Gamergull's dynamic traffic, yet better I can still run all video junk and still keep up perfect 60fps, does require mild overclock of one of the fastest CPUs, not because of multicore, but single core, however it does run incredible well, I would of not believed such ever be possible year ago, but it is.
