TR Forums

Ok, so awhile back I posted a thread here: viewtopic.php?f=12&t=119946&p=1359298#p1359298

I was having some strange GW2 performance issues. Seemed to be JUST that game, as all the benches indicated hardware was performing at specc or better. The only conclusion we came to in that thread was something about the game disliked my particular 980, or my board or something.

So here's where the mystery gets strange. Since then I've completely rebuilt my machine. I'm now running an 8700K, a 1080, the works. Only thing that remains is PSU, Case, and drives. While GW2 now runs better than it did before, it's not by much, and still runs worse than friends with now much lesser hardware.

However, I have recently actually run into another game that inexplicably runs like crap: Elder Scrolls Online. On medium in towns I get framedips in the 20s, and the average FPS is around 45ish. Out in the world it runs closer to 60, but it still dips to the 40s quite commonly.

Same situation, too. Hardware benches at around where other similar hardware is. Cinebench 1221, I don't remember the Basemark or Tomb Raider benches but they check out.

At this point I have no clue what could be causing these issues. I'm not OC'ing anything. I've gone through countless windows builds. I've tried fresh installed OS's. Since both offending games are MMOs I thought I'd try different network infrastructure, but no luck there either. Nothing seems to help. My best guess now is that I'm literally cursed to have performance issues in certain MMOs. That's not really what's going on, as scientifically there has to be some cause, but I have no ideas. This has plagued me for years, and I'm at my wits end.

Any mods?

Sorry if I'm being redundant but I didn't have time to look over the whole original thread... what resolution are you playing at? Are you set on windowed, full screen, or windowed full screen?

Usacomp2k3 wrote:

Any mods?

For neither game. Though I did install a few about an hour ago for ESO, I have yet to boot the game up with them. Also, fresh windows install with a fresh game install has already been tried.

kvndoom wrote:

Sorry if I'm being redundant but I didn't have time to look over the whole original thread... what resolution are you playing at? Are you set on windowed, full screen, or windowed full screen?

I've tried all three, at a myriad of resolutions. I typically play at 1200p win windowed full screen.

Have you tried 1080 instead of 1200? Maybe something in GW2s render engine isn't optimized for 16:10.

Also, both games you mention are MMOs, which means a variable you can't control for is the number of other players in the zone. MMOs that otherwise have very modest performance requirements can get very CPU heavy if there are a large number of other players in the area. Is there any correlation you can see with how populated the instance is at the time of the performance drops?

What motherboards did you use in these builds? Are audio and networking motherboard built-ins, or are you doing anything more specialized for them? Are the drives you carried over being used as entirely generic drives, or are they using some more specialized drivers? Do any other peripherals need anything trickier than the basic USB HID driver?

If you run LatencyMon while the game is running, what's the tab where it gives a breakdown of the stats of each driver look like? IIRC there's some stuff about total runtime per-driver in there, not just maximum times.

K-L-Waster wrote:

Have you tried 1080 instead of 1200? Maybe something in GW2s render engine isn't optimized for 16:10.

Also, both games you mention are MMOs, which means a variable you can't control for is the number of other players in the zone. MMOs that otherwise have very modest performance requirements can get very CPU heavy if there are a large number of other players in the area. Is there any correlation you can see with how populated the instance is at the time of the performance drops?

I've tried 1080p, no change.

Your second point is a good one, however me and several different friends can stand in the same areas, crowded or no, and they will get significantly better performance. My machine struggles to maintain more than 30 or so FPS in any kind of field combat in GW2 no matter the player density. ESO isn't quite that bad, but the framedips definitely happen even when players are not around.

synthtel2 wrote:

What motherboards did you use in these builds? Are audio and networking motherboard built-ins, or are you doing anything more specialized for them? Are the drives you carried over being used as entirely generic drives, or are they using some more specialized drivers? Do any other peripherals need anything trickier than the basic USB HID driver?

If you run LatencyMon while the game is running, what's the tab where it gives a breakdown of the stats of each driver look like? IIRC there's some stuff about total runtime per-driver in there, not just maximum times.

1. Gigabyte Z370XP. Used this board in a few builds and have had no performance issues with it before. My old board was, uhh, I can't remember. Sandy Bridge era chipset.
2. Audio and networking are integrated, though I use my blue yeti for audio most of the time. The problem has been happening since before I got that piece of equipment though.
3. I'm not sure what you're asking here. The drives are standard formatted NTFS SATA drives. Nothing funky with the setup. One is an Intel 540s 480 GB SSD. The other is a HGST NAS 4TB winchester drive.
4. I don't believe so. Though I'm not sure a single peripheral I'm using hasn't been replaced in the last 2 years, and the problem is older than that.

I haven't heard of LatencyMon, but it sounds like just the kind of utility that might assist in getting to the bottom of this. I'll run some tests with this in a bit.

Vinceant wrote:

Only thing that remains is PSU, Case, and drives.

well, since you've already changed everything else, these are the next to check.

you can try switching to a different PSU and drive, while running the system outside of the case. this should test all three components.

Vinceant wrote:

1. Gigabyte Z370XP. Used this board in a few builds and have had no performance issues with it before. My old board was, uhh, I can't remember. Sandy Bridge era chipset.
2. Audio and networking are integrated, though I use my blue yeti for audio most of the time. The problem has been happening since before I got that piece of equipment though.
3. I'm not sure what you're asking here. The drives are standard formatted NTFS SATA drives. Nothing funky with the setup. One is an Intel 540s 480 GB SSD. The other is a HGST NAS 4TB winchester drive.
4. I don't believe so. Though I'm not sure a single peripheral I'm using hasn't been replaced in the last 2 years, and the problem is older than that.

Sounds like the only thing that might carry over would be Intel ethernet, and that should be solid unless LatencyMon says otherwise.

If LatencyMon doesn't say anything, moriz has got the right idea. What case and PSU are you using? It'd be weird for electrical or EM gremlins to have such a similar effect, but given all the things that haven't worked, it'd probably be worth messing with that environment as much as is convenient.

moriz wrote:

Vinceant wrote:

Only thing that remains is PSU, Case, and drives.

well, since you've already changed everything else, these are the next to check.

you can try switching to a different PSU and drive, while running the system outside of the case. this should test all three components.

When I tried the fresh install I used a different drive. I don't know that I own a PSU big enough to power this unit laying around, but I could probably acquire one. I want to say that I replaced mine a few years ago, but can't remember if that was before or after I got my 980.

Do you really think the case could be the culprit...? I mean, I've seen cases "go bad" as it were, but usually it's just bent in a weird way (or dirty, or has terrible standoffs) and shorts out the motherboard and the system exhibits much more extreme symptoms, if it even powers on at all.

synthtel2 wrote:

If LatencyMon doesn't say anything, moriz has got the right idea. What case and PSU are you using? It'd be weird for electrical or EM gremlins to have such a similar effect, but given all the things that haven't worked, it'd probably be worth messing with that environment as much as is convenient.

Case is an Antec 300 two (what a confusing name). I've been considering replacing it, but I really, REALLY like the side panels on it, and most new cases I've seen that are reasonable in price have **** mechanisms.

PSU is uhhh, seasonic something. a 550 I believe, might be 650. Can't remember the model offhand. I think my old one was a Corsair, but I've found the caps and fans in those to be less than reliable.

Buying another PSU is not a big deal. It's always good to have a spare around for the business. The question is, should I buy another Seasonic? Or try a different brand for science?

Sorry for the double post, but I tried LatencyMon... I don't see anything I don't expect to see topping this list.

The DX graphics kernel tops the list of total execution by far. Then the latency monitor itself, which is no surprise. The next few are stuff like the kernel mode driver framework runtime, the nvidia kernel mode driver, the HDA bus driver, and there's an afd.sys, ancillary function driver for winsock. The rest all have 200 or less MS of execution time in 5:30 minutes of the game running.

Other than the possibility of the case doing something weird with grounding, there's also EM. A problem I recently ran into is that my wifi went nuts when I tried to run my RAM at 3066 (~5/4ths of 2.4's actual), and that sorted itself out when I closed up some holes in the case. If your problem were something along those lines, it'd be far weirder than the wifi problem, but we're fast running out of options.

There shouldn't be any need to buy a new case to test, just putting the system together outside the case works.

Aside from Seasonic, the EVGA G2/G3 are excellent.

What are the actual execution times like for those top items?

So while I've quit cold turkey for the last four months I have thousands of hours of GW2 playtime under my belt. A few things to keep in mind:

1. Unless your GPU is a potato, GW2 is going to be CPU-bound.
2. GW2 has no multi-threading to speak of, so single thread performance is basically a direct determinant of game performance.
3. No PC on the planet is capable of running GW2 at 60 fps (or even 30 fps) all of the time. This is partly due to the age of the engine (DirectX 9) and partly due to typical MMO variables. I'm saying this as a player with an i7-7700K overclocked to 5GHz and a GTX 1070.
4. It's very difficult to compare different people's FPS experiences with GW2. Even if you're on the same map playing the same event, the position of the players being compared can have a huge impact on their relative experience.
5. An extra variable related to number 4 is the large number of settings in GW2 that impact performance. Some affect it a little, some quite a bit and some have a variable effect depending on where in the world you're standing. Unless you synchronize every setting with the person with whom you're comparing performance you should view any findings with a high level of skepticism, and that doesn't even take point 4 into account.

There are two types of performance issue I see with GW2. One is a momentary lapse in frame rate that can happen if I turn in a direction I haven't looked in before. I attribute this to texture loads that aren't very well-handled by the relatively ancient engine. The other issue is an extended and sometimes severe drop in frame rate that tends to show up in a particular area. I find this most directly related to the number of players in the area and it can range from FPS dropping into the ~35 fps range if I'm wandering around Lion's Arch to FPS dropping into the teens (or worse) if I'm fighting a world boss alongside a hundred or more other people.

My suggestion would be to not worry about comparing your FPS numbers to other people, even on the same map. There are just too many variables (and too much variance of frame rate) for it to matter, and no way to conduct a standardized repeatable benchmark. Concentrate instead on whether the game is playable (i.e. in the 30-60 FPS range). If it isn't in a lot of situations, you can either try messing around with settings yourself or google which settings make the most difference. If you're comfortable overclocking and haven't done so already, try bumping up your CPU clock speed, especially the max turbo on a single core. Your 1080 is beyond overkill for this game unless you're running at 4K or more (and perhaps not even then) so don't worry about that.

Finally, accept the fact that in certain situations your frame rate is just going to suck. If you go do the Maw or the Shadow Behemoth with a couple hundred of your closest friends (or a map limit of people if either of those is a Daily) or join the mob at Tequatl at reset then your frame rates may dip into single digits until the event is over and that's completely expected. Hope this helps.

While all of that is certainly true. Wouldn't you agree that I should at least be able to consistently hit 60 FPS while staring at a wall? Because I cannot even do that. Also, wiggling the mouse AT ALL will drop the FPS, even while staring at no players, and no NPC's, to 40s or lower. I don't mean rotating to look at something else, I mean wiggling by a degree or two back and forth.

I can tell you that the only graphics settings that seem to have any effect on performance for me are shadow quality, and model limit. Nothing else changes performance in any meaningful way.

What is your average FPS on your 1070? What's the highest you can achieve in a low populated area out of combat? Also, what does GW2 auto detect your settings at? Because on my machine it auto detects to the lowest settings possible, regardless of what hardware set I was running at the time.

I fully understand that it is impossible to get a locked 60 FPS in GW2, but other people report much, MUCH better performance than I get regardless of my settings. Also, watching videos of people running fractals or dungeons, it is possible to get close to solid 60 FPS (apparently even while recording) in a lot of situations. Yes the game chugs when looking at the bank in LA during prime time (ancient engine, indeed), but it shouldn't struggle to maintain 30 when soloing in Queensdale.

ESO is the same way. It shouldn't drop to the 30s while soloing out in the field.

synthtel2 wrote:

Other than the possibility of the case doing something weird with grounding, there's also EM. A problem I recently ran into is that my wifi went nuts when I tried to run my RAM at 3066 (~5/4ths of 2.4's actual), and that sorted itself out when I closed up some holes in the case. If your problem were something along those lines, it'd be far weirder than the wifi problem, but we're fast running out of options.

There shouldn't be any need to buy a new case to test, just putting the system together outside the case works.

Aside from Seasonic, the EVGA G2/G3 are excellent.

What are the actual execution times like for those top items?

This weekend I can likely gut the system and put it on the bench and test it that way. I would be surprised if that was it, but like you said, there is not much else that's common between all of these.

I've been trying to find a benchmark to reproduce the problem consistently, and outside of GW2 or ESO, I have yet to find one... until now I think. I've tried Basemark, and Cinebench, and those all seem to run at specc. I've tried Tomb Raider and others like it and those also seem to run at similar speeds to TR tests of various cards. However, I tried GTA V's benchmark earlier, and it appeared to struggle more than it should. Even at 1200/1080p I struggled to maintain 55+ FPS on the highest settings, and had dips in the 30s, judging by tests posted here on TR, and elsewhere, this is not typical performance for this card. I can tell you that my 980 definitely ran it worse than that as well.

So at least I have a more consistent benchmark for this issue now. Now I can more reliably test it.

Ok, the latencymon stats I stupidly forgot to save. Here's another run (after a reboot) for 2 minutes running around in a low populated area in LA:



Also of interest is this video:

https://www.youtube.com/watch?v=qc_bIRzuLmI

This guy gets better average performance while at 4k, than I do at 1080p.

If you have a small hard drive lying around you should do a fresh win install plus the 1-2 games and see if that persists. You did carry over the previous install, I suppose?

ptsant wrote:

If you have a small hard drive lying around you should do a fresh win install plus the 1-2 games and see if that persists. You did carry over the previous install, I suppose?

I have tried this very experiment before to no avail. It's a great suggestion normally thought!

Yup, that's a fine showing in LatencyMon.

Maybe reset the BIOS, touch as few settings as possible, and try again? Most of what's left would seem to be settings that shouldn't be relevant, and that you might habitually set while assuming that irrelevance. Even going through reset BIOS -> fresh OS install w/ local account -> fresh game account with a minimal set of hardware and software and taking care to touch as few settings as possible might be worth it.

synthtel2 wrote:

Yup, that's a fine showing in LatencyMon.

Maybe reset the BIOS, touch as few settings as possible, and try again? Most of what's left would seem to be settings that shouldn't be relevant, and that you might habitually set while assuming that irrelevance. Even going through reset BIOS -> fresh OS install w/ local account -> fresh game account with a minimal set of hardware and software and taking care to touch as few settings as possible might be worth it.

When I did a fresh windows install I did all those things (I'm almost positive I reset the BIOS). I even made sure to never login from my main account in GW2 just to make sure no cross contamination.

Wait, I may not have done a local account. I might try that when I pull the board and card out for a bench test. I've got one of these on the way. Monday night looks like the testing night. It might also be worth it to just try a different OS, like Ubuntu or something.

Here's the thing though, what if results are the same outside the case, with new PSU, on new BIOS reset, on fresh OS with local account? Where would I go from there?

Vinceant wrote:

When I did a fresh windows install I did all those things (I'm almost positive I reset the BIOS). I even made sure to never login from my main account in GW2 just to make sure no cross contamination.

Wait, I may not have done a local account. I might try that when I pull the board and card out for a bench test. I've got one of these on the way. Monday night looks like the testing night. It might also be worth it to just try a different OS, like Ubuntu or something.

Here's the thing though, what if results are the same outside the case, with new PSU, on new BIOS reset, on fresh OS with local account? Where would I go from there?

Process Explorer's per-thread listing of CPU time use, probably? If you could run that on another computer on which the game runs well, to get a comparison, it would at least be interesting. When looking at that, the other thing to check would be how different shadow settings affect it.

GW2 runs on Linux, but purports are that it doesn't particularly well (WineHQ). If you had an AMD card, Gallium Nine would definitely be interesting, but if it's meh on Nvidia's driver there's not much to be done about it. If you could find any particular case where GW2 were faster there than on Windows, it'd definitely be useful, it just seems unlikely that that'll happen.

WineHQ posts mention running the game with -dx9single, which I can't seem to find any references to anywhere else. No idea what it does, but it's a switch that'll be flippable on Windows too, and that sounds worth a shot.

More directed benchmarking of the system could yet turn something up. Poking and prodding GW2 some more before that to narrow down the problem would be good, except that GW2 isn't a sort of game that's liable to take kindly to that (for anti-cheat reasons). I've got at least two more ideas along those (directed benchmarking) lines, but they're half-baked; I'll get back to you on that.

Vinceant wrote:

ptsant wrote:

If you have a small hard drive lying around you should do a fresh win install plus the 1-2 games and see if that persists. You did carry over the previous install, I suppose?

I have tried this very experiment before to no avail. It's a great suggestion normally thought!

Try disconnecting almost everything non-essential. Disable onboard sound at the BIOS. Disable any other peripheral, onboard or not, that you do not absolutely need. Try connecting to the network via cable. Use a simple braindead keyboard and mouse. Record your frame times.

Now tune the settings and record your frame times again. Reduce everything, especially shadows and high-res textures.

Log temps for GPU/CPU.

Run a memory tester.

Try a "performance" power profile on windows.

A friend had a similar situation, turned out that the CPU would throttle occasionally (stock cooler and a lot of dust) and would generate spikes.

My PC (see sig) is definitely slower but I can enjoy GW2 at 100+ fps with medium-high settings in the open world. It feels super smooth with FreeSync. You should be able to achieve much better performance.

Vinceant wrote:

While all of that is certainly true. Wouldn't you agree that I should at least be able to consistently hit 60 FPS while staring at a wall? Because I cannot even do that. Also, wiggling the mouse AT ALL will drop the FPS, even while staring at no players, and no NPC's, to 40s or lower.

Heh, have you tried a different mouse? Or current mouse with just generic drivers and no proprietary software? I dunno, a loss of FPS by moving the mouse tells me that we might be looking in all the wrong places.

ptsant wrote:

Try disconnecting almost everything non-essential. Disable onboard sound at the BIOS. Disable any other peripheral, onboard or not, that you do not absolutely need. Try connecting to the network via cable. Use a simple braindead keyboard and mouse. Record your frame times.

Now tune the settings and record your frame times again. Reduce everything, especially shadows and high-res textures.

Log temps for GPU/CPU.

Run a memory tester.

Try a "performance" power profile on windows.

A friend had a similar situation, turned out that the CPU would throttle occasionally (stock cooler and a lot of dust) and would generate spikes.

My PC (see sig) is definitely slower but I can enjoy GW2 at 100+ fps with medium-high settings in the open world. It feels super smooth with FreeSync. You should be able to achieve much better performance.

I don't know that the logic follows to any on board device, seeing as this has transferred between machines. I may try it in a bit though.

Even though my wifi is awesome (Yay ubiquiti AC Pro), my desktop is wired.

I don't know how to record frame timings.

It's not a temperature issue. Not only has this problem persisted across 3 CPUs and 2 GPUs, I do monitor temperatures and they are very acceptable.

Memtest is something I haven't run on this rig, however, I definitely ran it on the previous hardware set.

I've tried high performance and normal on windows power management. Also, pretty much all "high performance" does on a desktop is not let the GPU or CPU downclock at idle.

I can only break 100 FPS when settings are at absolute minimum and I'm staring at a wall.

kvndoom wrote:

Vinceant wrote:

While all of that is certainly true. Wouldn't you agree that I should at least be able to consistently hit 60 FPS while staring at a wall? Because I cannot even do that. Also, wiggling the mouse AT ALL will drop the FPS, even while staring at no players, and no NPC's, to 40s or lower.

Heh, have you tried a different mouse? Or current mouse with just generic drivers and no proprietary software? I dunno, a loss of FPS by moving the mouse tells me that we might be looking in all the wrong places.

Just tried a wired Logitech M500 (normal mouse is a G700s) while disabling the logitech software. No change. Good suggestion though.

I haven't had a chance to tear the machine out of the case and bench it yet, hopefully going to do that tonight.

kvndoom wrote:

Vinceant wrote:

While all of that is certainly true. Wouldn't you agree that I should at least be able to consistently hit 60 FPS while staring at a wall? Because I cannot even do that. Also, wiggling the mouse AT ALL will drop the FPS, even while staring at no players, and no NPC's, to 40s or lower.

Heh, have you tried a different mouse? Or current mouse with just generic drivers and no proprietary software? I dunno, a loss of FPS by moving the mouse tells me that we might be looking in all the wrong places.

It isn't uncommon for games to cache some view-dependent information and have to burn some CPU time recalculating when the camera moves, it just should be burning at most an order of magnitude less than that, and it shouldn't come into play at all on a couple degrees of camera rotation.

It could be a useful divider between a sort-of bad case and really bad case. Maybe compare CPU use / GPU use / PCIe link use in that static case and while moving the mouse slightly?

(I have encountered one game with serious view-dependent caching problems, that being Shadow of Mordor. It's got three of its six heavy threads dedicated to that kind of work, and it only gets worse from there.)

Finally, some progress!

Ok, so out of the case, new PSU, tests did nothing... on my current windows install.

I was hesitant to try a fresh install just yet as I'm out of bandwidth until the 5th (soon my local electric coop will have unlimited internet with ftth... soon), but I just copied the GW2 dat to a USB drive.

So, fresh install (not sure which build), LOCAL ACCOUNT, defaults on everything else except DSR as my bench's monitor is 1600x900, so I need to use DSR to get a similar res to my proper workstation. And... bam. Same wall I get 70-75 FPS staring at normally, I'm getting 95-115 FPS. It still slows down to the 40s near the bank, but we all know that's an engine limitation. However, I can tell you that the game feels much, MUCH smoother. Turning around, opening the map, none of it is stuttery. It still auto detects to the lowest settings though, but whatever. Setting it to the minimum before granted me 110 FPS, on the fresh install it was in the 200s somewhere.

So, we have a condition where the hardware acts normally. Something to go on. Now, I've done a fresh install before and had no luck, but I must've either used my MS account, or my Nvidia account, or both, and some setting somewhere must have propagated. The question is, where? I've wiped and cleaned all Nvidia settings on multiple occasions, so I'm leaning more toward some windows setting somewhere. I have no idea where I would start.

kvndoom wrote:

I dunno, a loss of FPS by moving the mouse tells me that we might be looking in all the wrong places.

This caught my attention as well, as I had an issue with a wireless device that may be worth mentioning. A few months back, I tried switching to a wireless keyboard with my primary computer, but I discovered that due to my Ubiquiti AC LR being located adjacent to the desk, my keystrokes were intermittently registered in a very laggy fashion. Sometimes, it would work fine, and then suddenly it would get many characters behind. When it did this, the whole system acted flaky, and then all the typed characters would suddenly appear, and the system would act normal again. Since moving the WAP would be a pain, I switched back to a wired keyboard. If you are using a wireless keyboard that is located in the vicinity of your WAP, you may want to test using a wired keyboard or turning off your wireless for a bit to see if this is affecting your game experience.

Aether wrote:

kvndoom wrote:

I dunno, a loss of FPS by moving the mouse tells me that we might be looking in all the wrong places.

This caught my attention as well, as I had an issue with a wireless device that may be worth mentioning. A few months back, I tried switching to a wireless keyboard with my primary computer, but I discovered that due to my Ubiquiti AC LR being located adjacent to the desk, my keystrokes were intermittently registered in a very laggy fashion. Sometimes, it would work fine, and then suddenly it would get many characters behind. When it did this, the whole system acted flaky, and then all the typed characters would suddenly appear, and the system would act normal again. Since moving the WAP would be a pain, I switched back to a wired keyboard. If you are using a wireless keyboard that is located in the vicinity of your WAP, you may want to test using a wired keyboard or turning off your wireless for a bit to see if this is affecting your game experience.

This is perfectly normal in big landscapes where one second you are looking at a wall (1 item to render) vs 50 players and a big boss in the opposite direction (1000s of items to render).

Nice! I've never used a non-local account and have no idea how settings propagate, though, so may not be much help there.

Maybe the advanced system settings area with pagefile and scheduling options?

Another good reason not to use a Microsoft account, I guess.

Vinceant wrote:

Finally, some progress!

Ok, so out of the case, new PSU, tests did nothing... on my current windows install.

I was hesitant to try a fresh install just yet as I'm out of bandwidth until the 5th (soon my local electric coop will have unlimited internet with ftth... soon), but I just copied the GW2 dat to a USB drive.

So, fresh install (not sure which build), LOCAL ACCOUNT, defaults on everything else except DSR as my bench's monitor is 1600x900, so I need to use DSR to get a similar res to my proper workstation. And... bam. Same wall I get 70-75 FPS staring at normally, I'm getting 95-115 FPS. It still slows down to the 40s near the bank, but we all know that's an engine limitation. However, I can tell you that the game feels much, MUCH smoother. Turning around, opening the map, none of it is stuttery. It still auto detects to the lowest settings though, but whatever. Setting it to the minimum before granted me 110 FPS, on the fresh install it was in the 200s somewhere.

So, we have a condition where the hardware acts normally. Something to go on. Now, I've done a fresh install before and had no luck, but I must've either used my MS account, or my Nvidia account, or both, and some setting somewhere must have propagated. The question is, where? I've wiped and cleaned all Nvidia settings on multiple occasions, so I'm leaning more toward some windows setting somewhere. I have no idea where I would start.

this is really interesting!

i'd never thought of local vs microsoft accounts having an effect. perhaps there's some hidden settings being copied around with your microsoft account, and that's interfering with game performance.

now the question is: which setting? perhaps try signing into the two accounts one at a time, and see which one end up having an impact on performance.

how would I sign in to both though?

I guess the first thing to try would be to use a driver cleaner to uninstall everything nvidia related and then reinstall that. What's that utility people use for such a thing? DD something I want to say.