Support NeoGAF

[The Cockatrice]

I tried on reddit but no luck maybe someone on gaf is more savvy than me/reddit. I noticed my 2080 underperforming lately or I dont know maybe its just me but I looked at some dudes videos with a weaker CPU, same 2080 GPU, 1440p all maxed and they had like 20 more framerate than me. I've seen at least 2 different videos from different people. Noticed this on Far Cry 6 and Days Gone.

Then I decided to do a 3DMark test and to my shock the results are piss poor

If you search that gpu and its score most of them are in the 11k range. I'm at the bottom. I had this GPU for 3-4 years now, its from MSI but I never had issues.

Things I tried:
-gpu is well placed in its pcie slot
-tried placing 2 psu cables on it instead of a splitter
-fresh drivers install with DDU, tried 470 driver and then went to latest, same thing.
-power performance is all set to high performance.

Heres my average framerate in Days Gone

Spoiler

and heres a video of some dude having 100 with worse CPU than mine(I got a 12700k) and the exact same GPU


Any ideas?

 

[Shifty]

First step I'd take would be capturing some more in depth perf stats during gameplay - you can use MSI Afterburner to get simultaneous graphs for things like clocks, temps and power usage across both CPU and GPU.

I'd try to make some inferences based on that TimeSpy chart, but it's so low res I can't make out the legend for the Y axis.

Anyhow, that information should contextualize things a bit - could be that you're hitting thermal throttling on either CPU or GPU (have you dusted your case innards recently?) but it wouldn't help to throw further speculation at you without hard numbers.

 

[thief183]

Temps, I solved my problems clening the card and replacing the thermal paste. You might want someone Expert to do it tho.

 

[The Cockatrice]

First step I'd take would be capturing some more in depth perf stats during gameplay - you can use MSI Afterburner to get simultanous graphs for things like clocks, temps and power usage across both CPU and GPU.

I'd try to make some inferences based on that TimeSpy chart, but it's so low res I can't make out the legend for the Y axis.

Anyhow, that information should contextualize things a bit - could be that you're hitting thermal throttling on either CPU or GPU (have you dusted your case innards recently?) but it wouldn't help to throw further speculation at you without hard numbers.

Temps, I solved my problems clening the card and replacing the thermal paste. You might want someone Expert to do it tho.

The problem is I don't have much to compare those stats. I have MSI Afterburner, I've looked at some of the stats but I'm not sure how they help me since I dont have the exact same stats on another PC to check. I thought about throttling but GPU temp does not go above 75(room temp is 30 atm so yeah). CPU Temps are lower at around 65 average or something. However the hotspot for the GPU on hwmonitor is at 100ish but that shouldnt be a problem from what I know. Throttling should only take core and memory temps in consideration, no? I also thought it could be a voltage issue but I am clueless there. I heard about replacing the thhermal plates and removing the plate but yeah might be easier to just save up and get a new gpu.

 

[kraspkibble]

are you on windows 10 or 11? did you recently upgrade to 11? unlikely it's the issue but it could be.

first thing i'd do is clean out my PC and try replacing pads/paste on the GPU if the temps are high. my founders 2080 idles at ~30C and maxes out at about 70-75C when overclocked at playing at 1440p 144hz. edit: wait you said your GPU is 100C??? that's not right.

if cleaning it out doesn't sort it then might be worth a clean windows install.

 

[The Cockatrice]

are you on windows 10 or 11? did you recently upgrade to 11? unlikely it's the issue but it could be.

first thing i'd do is clean out my PC and try replacing pads/paste on the GPU if the temps are high. my founders 2080 idles at ~30C and maxes out at about 70-75C when overclocked at playing at 1440p 144hz. edit: wait you said your GPU is 100C??? that's not right.

if cleaning it out doesn't sort it then might be worth a clean windows install.

Normal temps under load are 75. Hwmonitor also shows you the hotspot which i dont think it matters. Yeah i could try that but never done it before and since temps seem fine so far its either that thise 2 games hate my specific config or something else. I didnt upgrade yet to 11. I just freshly reinstalled win10 4 months ago...cant be that either...i think...i dunno..im out of ideas besides the gpu just showing its age.

 

[kraspkibble]

Normal temps under load are 75. Hwmonitor also shows you the hotspot which i dont think it matters. Yeah i could try that but never done it before and since temps seem fine so far its either that thise 2 games hate my specific config or something else. I didnt upgrade yet to 11. I just freshly reinstalled win10 4 months ago...cant be that either...i think...i dunno..im out of ideas besides the gpu just showing its age.

have you changed anything else in your PC? ram perhaps? ssd? i wouldn't think SSD should affect performance much but RAM definitely will.

it's also possible it's not your GPU but maybe your CPU. i think that benchmark combines both so it could be lowered as a result of CPU. any BIOS updates or changes?

i suppose it could be age but that still seems odd. there are many 2080s still going strong (including mine) and MSI is a good brand.

if it's any help to you i can download 3d mark and do a run to see what my scores are.

edit: also i would say that TimeSpy is quite old. have you tried any other benchmarks? i like to run the RTX one and also unigine superposition.

 

[The Cockatrice]

have you changed anything else in your PC? ram perhaps? ssd? i wouldn't think SSD should affect performance much but RAM definitely will.

it's also possible it's not your GPU but maybe your CPU. i think that benchmark combines both so it could be lowered as a result of CPU. any BIOS updates or changes?

i suppose it could be age but that still seems odd. there are many 2080s still going strong (including mine) and MSI is a good brand.

if it's any help to you i can download 3d mark and do a run to see what my scores are.

edit: also i would say that TimeSpy is quite old. have you tried any other benchmarks? i like to run the RTX one and also unigine superposition.

RAM/CPU/MOBO are all new. Got 32GB DDR5, clocks/mem seems fine on them, at least based on task manager. I did a Cinebench test on the CPU and the score was fine. I only have TimeSpy cuz its free and its the only one out of the free ones thats at 1440p. If you want and when you have time, sure, tho I looked at some other results and it aint looking good for me lol

3DMark.com search

3DMark.com search

www.3dmark.com

Seems like the average is around 11k-12k. Im at 9.

 

[kraspkibble]

RAM/CPU/MOBO are all new. Got 32GB DDR5, clocks/mem seems fine on them, at least based on task manager. I did a Cinebench test on the CPU and the score was fine. I only have TimeSpy cuz its free and its the only one out of the free ones thats at 1440p. If you want and when you have time, sure, tho I looked at some other results and it aint looking good for me lol

3DMark.com search

3DMark.com search

www.3dmark.com

something definitely ain't right. the only thing i can think of is maybe cause you're on DDR5 and a quite new platform there might be bugs in the BIOS.

i just ran a Time Spy benchmark and here are my results....

 

[The Cockatrice]

something definitely ain't right. the only thing i can think of is maybe cause you're on DDR5 and a quite new platform there might be bugs in the BIOS.

i just ran a Time Spy benchmark and here are my results....

You think DDR5 still being relatively fresh, might not sit well with these benchmarks/certain games? Can it impact GPU performance? I dunno what else could it be. Obviously ditching the DDR5 is out of the question.

 

[Xyphie]

Something looks wonky with your GPU memory clocks? Not only is the your memory frequency weird (should be ~1750MHz on a non-OC'd card) but it's a static ~2500MHz during the entire run. Looking at billyxci's graph you can see how the clock throttles back during the intermissions in the run but yours doesn't. 2D GPU clocks don't seem to go down as much either.

 

[kraspkibble]

You think DDR5 still being relatively fresh, might not sit well with these benchmarks/certain games? Can it impact GPU performance? I dunno what else could it be. Obviously ditching the DDR5 is out of the question.

definitely. like i said time spy is old. it's from like 2016/2017. DDR4 only become widely available in 2015/2016 with Skylake intel cpus. I do think it's probably about time that 3D Mark (+ Unigine) start working on more modern benchmarks.

 

[kraspkibble]

Something looks wonky with your GPU memory clocks? Not only is the your memory frequency weird (should be ~1750MHz on a non-OC'd card) but it's a static ~2500MHz during the entire run. Looking at billyxci's graph you can see how the clock throttles back during the intermissions in the run but yours doesn't. 2D GPU clocks don't seem to go down as much either.

i never even noticed that. that is weird there isn't any change. could be a memory issue!

edit: is your GPU overclocked? if so have you tried running a benchmark with no OC?

 

[TrueLegend]

You need to first reset bios, then apply thermal paste again which is a bit of a hectic task on gpu. Then you can format your os. If these three don't solve your problem then you are in a bit of pickle.

 

[The Cockatrice]

i never even noticed that. that is weird there isn't any change. could be a memory issue!

edit: is your GPU overclocked? if so have you tried running a benchmark with no OC?

Something looks wonky with your GPU memory clocks? Not only is the your memory frequency weird (should be ~1750MHz on a non-OC'd card) but it's a static ~2500MHz during the entire run. Looking at billyxci's graph you can see how the clock throttles back during the intermissions in the run but yours doesn't. 2D GPU clocks don't seem to go down as much either.

Hm i dunno... if it helps I tried overclocking the gpu and it made the score worse. I tested this on a game and it looks like overclocking is making things worse for me. My GPU is dying isnt it

 

[TintoConCasera]

Are you on a laptop?

 

[kraspkibble]

Are you on a laptop?

did you bother reading the original post?

 

[TrueLegend]

Just noticed that your igpu is not disabled. You should try to disable it

 

[anothertech]

Dust, paste and psu are what I look at first.

Is your psu pushed to it's limits and becoming heat factor? What kind of cooling do you have on cpu? Your 4 year old gpu been properly cleaned ever? Have you cleaned and reapplied thermal paste?

Strangely, simple things like that can add up and cause real time issues with hardware.

 

[kraspkibble]

Just noticed that your igpu is not disabled. You should try to disable it

good catch.

OP go into your BIOS and disable that shit.

 

[kraspkibble]

Hm i dunno... if it helps I tried overclocking the gpu and it made the score worse. I tested this on a game and it looks like overclocking is making things worse for me. My GPU is dying isnt it

it could be.

perhaps try a benchmark designed just for GPUs. download Unigine Superposition and check your scores. i'll download it and see what mines is too

UNIGINE Benchmarks

Performance benchmarks by Unigine

benchmark.unigine.com

 

[Xyphie]

Is it a used GPU? If it is could be someone used it for crypto mining and flashed some weird mining optimized BIOS on it.

 

[The Cockatrice]

Just noticed that your igpu is not disabled. You should try to disable it

It's disabled but for some reason 3DMark keeps seeing it. I'll try to uninstall it and see what that does.

it could be.

perhaps try a benchmark designed just for GPUs. download Unigine Superposition and check your scores. i'll download it and see what mines is too

UNIGINE Benchmarks

Performance benchmarks by Unigine

benchmark.unigine.com

Alright, I'll check it out in a sec.

Are you on a laptop?

no, its a desktop.

Dust, paste and psu are what I look at first.

Is your psu pushed to it's limits and becoming heat factor? What kind of cooling do you have on cpu? Your 4 year old gpu been properly cleaned ever? Have you cleaned and reapplied thermal paste?

Strangely, simple things like that can add up and cause real time issues with hardware.

It's not very dusty. I'd say it has like 15% dust in places I cant reach. My CPU is cooled by https://noctua.at/en/nh-d15 and it has fresh thermal paste, temps are fine on the CPU.

 

[TrueLegend]

It's disabled but for some reason 3DMark keeps seeing it. I'll try to uninstall it and see what that does.

Alright, I'll check it out in a sec.


no, its a desktop.



It's not very dusty. I'd say it has like 15% dust in places I cant reach. My CPU is cooled by https://noctua.at/en/nh-d15 and it has fresh thermal paste, temps are fine on the CPU.

You need to do that in bios specifically. Although I am not sure that should be causing you problems but still. Turn off hyperthreading and igpu in bios. And then test timespy.

 

[kraspkibble]

The Cockatrice here are my superposition settings/results:

 

[kraspkibble]

You need to do that in bios specifically. Although I am not sure that should be causing you problems but still. Turn off hyperthreading and igpu in bios. And then test timespy.

why would you turn off hyper threading?

i know HT can affect performance in older cpus but it shouldnt with OPs cpu.

 

[TrueLegend]

why would you turn off hyper threading?

i know HT can affect performance in older cpus but it shouldnt with OPs cpu.

Just for troubleshooting

 

[kraspkibble]

Just for troubleshooting

fair enough

 

[Shifty]

The problem is I don't have much to compare those stats. I have MSI Afterburner, I've looked at some of the stats but I'm not sure how they help me since I dont have the exact same stats on another PC to check. I thought about throttling but GPU temp does not go above 75(room temp is 30 atm so yeah). CPU Temps are lower at around 65 average or something. However the hotspot for the GPU on hwmonitor is at 100ish but that shouldnt be a problem from what I know. Throttling should only take core and memory temps in consideration, no? I also thought it could be a voltage issue but I am clueless there. I heard about replacing the thhermal plates and removing the plate but yeah might be easier to just save up and get a new gpu.

Post that shit! Preferably with as many of the graphs turned on and visible as possible. It'll give us much more info to work with, which will help avoid confusing the issue with maybes.

And the hotspot temp shouldn't be anything to worry about - GPUs always reached those temps internally, just never told us about them until recently because the temp sensors were placed on the edge of the board and relied on thermal conduction to get their readings. My 6900XT is rated safe up to 110C hotspot, so it's fine after the initial Oh Shit when you see it brush against triple digits for the first time.

And throttling likely takes all three into account - hotspot temps are more responsive due to being integrated into the business components, which means the hardware can respond faster if it considers itself to be in danger.

Hm i dunno... if it helps I tried overclocking the gpu and it made the score worse. I tested this on a game and it looks like overclocking is making things worse for me. My GPU is dying isnt it

Overclocking isn't guaranteed to increase your final FPS output like it was in the old days when hardware didn't have boost clocks, or thermal throttling to protect itself from unsafe settings. If you push it too far (which in some cases can be any overclock at all - silicon lottery), it'll throttle immediately and reduce your perf instead of melting the card.

If anything, undervolting can yield better results on modern architectures since it gets the same clocks out of less power, resulting in less heat, less throttling, higher boosting, and thus better FPS. YMMV.

So your card isn't necessarily dying. It might just have come from a weaker part of the silicon wafer that doesn't have much OC headroom, or have some other factor (such as dust clog, or an underperforming water loop, etc) slowing it down.

It's disabled but for some reason 3DMark keeps seeing it. I'll try to uninstall it and see what that does.

Don't do this. It's only going to make things better if the games and benchmarks are incorrectly defaulting to your iGPU instead of your dGPU (which is very unlikely, and can be fixed separately in the Nvidia Control Panel), and is going to remove access to your mobo video outputs if disabled or uninstalled.

Also, bear in mind that other users' bench scores aren't going to be a good comparison with your own unless your hardware and software are identical - silicon lottery aside. CPU, GPU, Mobo, PCI lanes, I/O, there are so many variables that feed into it and potentially affect the final outcome.

 

[kraspkibble]

Also, bear in mind that other users' bench scores aren't going to be a good comparison with your own unless your hardware and software are identical - silicon lottery aside. CPU, GPU, Mobo, PCI lanes, I/O, there are so many variables that feed into it and potentially affect the final outcome.

true but OPs cpu/ram are much newer than mine and shouldn't be giving results worse. it's rare you can find benchmarks with the exact set up as yours but they do help.

my 4 year old CPU and DDR4 ram shouldn't really be beating a brand new 12th series CPU and DDR5 ram. the benchmarks shows something is definitely not right.

 

[Shifty]

true but OPs cpu/ram are much newer than mine and shouldn't be giving results worse. it's rare you can find benchmarks with the exact set up as yours but they do help.

my 4 year old CPU and DDR4 ram shouldn't really be beating a brand new 12th series CPU and DDR5 ram. the benchmarks shows something is definitely not right.

Fair point, but it's still possible depending on the other parts - I put a shit hot CPU / GPU on a recent-but-terrible motherboard for the first version of my current build, and only twigged that something was wrong when I ran cinebench and saw an identical CPU absolutely smoking mine in the crowdsourced multicore leadersboards regardless of OC / UV.

Turned out to be crappy voltage regulators causing the CPU to thermal throttle for such short periods that HWInfo was missing the event - I researched up a more reasonable mobo and ended up punching considerably higher than the original rival score, with generally better heat and framerates GPU-side thanks to easing off the CPU bottleneck.

 

[The Cockatrice]

Doing the tests now. Ill reply with results in a sec

 

[Hoddi]

You'll need to find out if it's your CPU or GPU that's underperforming. That scene in Days Gone should be hitting 200fps+ on your CPU if you drop resolution to something like 720p or 960x540.

It's kinda hard to pinpoint the issue without knowing which it is.

 

[PaddyOCanager]

c'mon guys......fix him.

 

[The Cockatrice]

The Cockatrice The Cockatrice here are my superposition settings/results:

You need to do that in bios specifically. Although I am not sure that should be causing you problems but still. Turn off hyperthreading and igpu in bios. And then test timespy.

Post that shit! Preferably with as many of the graphs turned on and visible as possible. It'll give us much more info to work with, which will help avoid confusing the issue with maybes.

And the hotspot temp shouldn't be anything to worry about - GPUs always reached those temps internally, just never told us about them until recently because the temp sensors were placed on the edge of the board and relied on thermal conduction to get their readings. My 6900XT is rated safe up to 110C hotspot, so it's fine after the initial Oh Shit when you see it brush against triple digits for the first time.

And throttling likely takes all three into account - hotspot temps are more responsive due to being integrated into the business components, which means the hardware can respond faster if it considers itself to be in danger.


Overclocking isn't guaranteed to increase your final FPS output like it was in the old days when hardware didn't have boost clocks, or thermal throttling to protect itself from unsafe settings. If you push it too far (which in some cases can be any overclock at all - silicon lottery), it'll throttle immediately and reduce your perf instead of melting the card.

If anything, undervolting can yield better results on modern architectures since it gets the same clocks out of less power, resulting in less heat, less throttling, higher boosting, and thus better FPS. YMMV.

So your card isn't necessarily dying. It might just have come from a weaker part of the silicon wafer that doesn't have much OC headroom, or have some other factor (such as dust clog, or an underperforming water loop, etc) slowing it down.


Don't do this. It's only going to make things better if the games and benchmarks are incorrectly defaulting to your iGPU instead of your dGPU (which is very unlikely, and can be fixed separately in the Nvidia Control Panel), and is going to remove access to your mobo video outputs if disabled or uninstalled.

Also, bear in mind that other users' bench scores aren't going to be a good comparison with your own unless your hardware and software are identical - silicon lottery aside. CPU, GPU, Mobo, PCI lanes, I/O, there are so many variables that feed into it and potentially affect the final outcome.

Well I had to try anyway to see if the igpu and whatnot has any effect. Sadly it does not. You can view the monitoring as well. Im not sure what else to try. I already tried underclocking, no change. Undervolting will most likely yield the same results tho not sure if I want to go there.

 

[The Cockatrice]

You'll need to find out if it's your CPU or GPU that's underperforming. That scene in Days Gone should be hitting 200fps+ on your CPU if you drop resolution to something like 720p or 960x540.

It's kinda hard to pinpoint the issue without knowing which it is.

Yep. Pretty sure it aint the CPU.

 

[ViolentP]

If you use something like Afterburner, make sure your card is set to stock settings. Whenever I benchmark, I always restart, disable gsync, disable vsync, and make sure I am not frame limiting. Compared to other users, it seems your bottleneck is the GPU so I would consider doing clean driver reinstall as well as reviewing your bios settings.

 

[The Cockatrice]

If you use something like Afterburner, make sure your card is set to stock settings. Whenever I benchmark, I always restart, disable gsync, disable vsync, and make sure I am not frame limiting. Compared to other users, it seems your bottleneck is the GPU so I would consider doing clean driver reinstall as well as reviewing your bios settings.

Already done those. Everything is fresh. I even tried an older driver 450 something. No improvement.

 

[ViolentP]

Already done those. Everything is fresh. I even tried an older driver 450 something. No improvement.

Shit, man. Maybe you got a bitcoin miner somewhere in there.

 

[Shifty]

Well I had to try anyway to see if the igpu and whatnot has any effect. Sadly it does not. You can view the monitoring as well. Im not sure what else to try. I already tried underclocking, no change. Undervolting will most likely yield the same results tho not sure if I want to go there.

Oh I'm not suggesting that you should need to overclock or undervolt to get the perf you expect - just that the results you see if you do might not be intuitive since the GPU architecture game has changed a lot in recent years.

Comparing your superposition results to billy's, we can see that his has a minimum temp of ~35, whereas yours is at 50. His GPU clocks are also consistently higher than yours on average during timespy - the pink line on his graph is just above 2000, where yours is just below.

So, we might be dealing with thermals rather than bunk hardware here. I'm not sure if it's the whole story based on the gap in the results, but there is at least a measurable correlation.
Modern GPUs are super aggressive with their boost clocks, and will eke out every last Hz until they hit a certain voltage / temp threshold - if you can bring down either of those, you should see performance rise accordingly.

What kind of cooling are you running? GPU temp will be affected by the case temp, so making sure everything is clean and well maintained will make a difference for air cooling setups. You might be able to take the side panel off and point a desk fan at the GPU as a testing measure to see if benchmark scores are affected by some extra airflow. It could be a case of dried out thermal paste, but that's a pain in the ass to redo on a GPU so it's best to eliminate the other options first.

That's about as much as I can glean without seeing your Afterburner panel, since the timespy graph is pretty coarse and doesn't show reference numbers for anything except frequency.
And we definitely need graphs, for the sake of seeing precise numbers and the difference between idle / load scenarios.

 

[kraspkibble]

definitely not your CPU.

ViolentP possible but made me think of a bug i saw reported on windows recently where the system was under high load when not doing anything. if you went into task manager or other monitoring software then the load would drop and look normal. could be a bug with windows.

if OP is up for it then it might be worth doing a completely clean install of the PC.

 

[The Cockatrice]

Oh I'm not suggesting that you should need to overclock or undervolt to get the perf you expect - just that the results you see if you do might not be intuitive since the GPU architecture game has changed a lot in recent years.

Comparing your superposition results to billy's, we can see that his has a minimum temp of ~35, whereas yours is at 50. His GPU clocks are also consistently higher than yours on average during timespy - the pink line on his graph is just above 2000, where yours is just below.

So, we might be dealing with thermals rather than bunk hardware here. I'm not sure if it's the whole story based on the gap in the results, but there is at least a measurable correlation.
Modern GPUs are super aggressive with their boost clocks, and will eke out every last Hz until they hit a certain voltage / temp threshold - if you can bring down either of those, you should see performance rise accordingly.

What kind of cooling are you running? GPU temp will be affected by the case temp, so making sure everything is clean and well maintained will make a difference for air cooling setups. You might be able to take the side panel off and point a desk fan at the GPU as a testing measure to see if benchmark scores are affected by some extra airflow. It could be a case of dried out thermal paste, but that's a pain in the ass to redo on a GPU so it's best to eliminate the other options first.

That's about as much as I can glean without seeing your Afterburner panel, since the timespy graph is pretty coarse and doesn't show reference numbers for anything except frequency. And we definitely need graphs, for the sake of seeing precise numbers and the difference between idle / load scenarios.

I'll try cleaning the gpu as much as I can and try with an open case and see if theres any framerate difference.

definitely not your CPU.

ViolentP possible but made me think of a bug i saw reported on windows recently where the system was under high load when not doing anything. if you went into task manager or other monitoring software then the load would drop and look normal. could be a bug with windows.

if OP is up for it then it might be worth doing a completely clean install of the PC.

Doesnt seem to be in high-load. I have a fan curve manually set up for the GPU and I'd be able to hear it when it goes under load. Might be as shifty said, the temps. Technically under load 75 should not be a problem, GPU"s only throttle at what, 95+? But Im sure everyone pretty much ran out of ideas. I'll give it a clean tomorrow, do some tests with open-case early in the morning when my room isnt burning from the sun and see if anything changes. If not, well, I dunno. I'll just stick with it until I can afford a new one. :\ plus a new power supply cuz the 3x series require 850+ and mines just at the minimum, 750W plat.

Oh and thank you everyone for the help.

 

[01011001]

have you checked your GPU with GPU-Z?

make sure that the GPU is handshaking correctly.
look if the Bus Interface says x16

I had a BIOS issue with my old motherboard that randomly only detected x2 on boot up, and I needed to boot once and restart to get the full x16

 

[ViolentP]

Do we know if performance has actually deteriorated or is it possible these were day 1 metrics?

 

[PharaoTutAnchAmun]

I have a PC but i dont game on it, and when i read all this i know why. Dont get me wrong, i feel sorry for you having all these troubles, and i hope you can solve them.
Gaming on a PC can be a great experience, but with all does diferent kind of CPU's, Memory, GPU's and so on it can be a burden. I mean theyr are so many drivers and test for all the different parts. Some problems is also trough the OS like different versions of Windows. And when your hardware gets older, then some problems will stick up theyr ugly heads.
Anyway, i hope you can get your PC up and running like it should.

Good luck.

 

[ViolentP]

I have a PC but i dont game on it, and when i read all this i know why. Dont get me wrong, i feel sorry for you having all these troubles, and i hope you can solve them.
Gaming on a PC can be a great experience, but with all does diferent kind of CPU's, Memory, GPU's and so on it can be a burden. I mean theyr are so many drivers and test for all the different parts. Some problems is also trough the OS like different versions of Windows. And when your hardware gets older, then some problems will stick up theyr ugly heads.
Anyway, i hope you can get your PC up and running like it should.

Good luck.

It's ok to not want options and variability but having them is certainly not a problem.

 

[PaintTinJr]

true but OPs cpu/ram are much newer than mine and shouldn't be giving results worse. it's rare you can find benchmarks with the exact set up as yours but they do help.

my 4 year old CPU and DDR4 ram shouldn't really be beating a brand new 12th series CPU and DDR5 ram. the benchmarks shows something is definitely not right.

Actually, looking into the new Heterogenous P/E cores cache arrangements(with l1,l2,l3 info at techpowerup), your regular 12MB L3 cache might be 2MB larger for the P cores (based on speculation by wccftech.com review), and your L1 and L2 to shared L3 cache arrangements go L1 x4 = L2, Then, L2 x4 x core count = L3 which seems like an ideal cache pyramid IMO, Where as on the 12gen P/E cores, it is an odd L1 size of 80KB probably for the E cores needs, Then L1x16 = L2(1.25MB) and then it is slightly opaque, because apparently the igpu shares the L3 cache too, and the thread director divides it up, but it is speculated AFAIK that the dedicated L3 shared for P cores is another core count x 1.25MB, with the remaining 15MB probably for the thread director and igpu.

There is also the possibility that your DDR4 has lower latency - which can be more important than memory clockrate - and your motherboard chipset might be a higher tier too, given that your chip is an i9 I doubt you went cheap. There's too much of the OPs specs we don't know. We don't even know how many memory sticks ,

 

[The Cockatrice]

have you checked your GPU with GPU-Z?

make sure that the GPU is handshaking correctly.
look if the Bus Interface says x16

I had a BIOS issue with my old motherboard that randomly only detected x2 on boot up, and I needed to boot once and restart to get the full x16

Do we know if performance has actually deteriorated or is it possible these were day 1 metrics?

Impossible to know now. Honestly if I didnt randomly decided to check days gone benchmarks I would've never known. I remember checking before cyberpunk benchmarks and metro but they all were in the same area of performance as mine so who knows.

There is also the possibility that your DDR4 has lower latency - which can be more important than memory clockrate - and your motherboard chipset might be a higher tier too, given that your chip is an i9 I doubt you went cheap. There's too much of the OPs specs we don't know. We don't even know how many memory sticks ,

Ram seems fine. XMP 3 also enabled.

 

[Black_Stride]

Silicon Lottery?

And/or disconnect your second monitor and see how it goes.
If you dont have second monitor, your PC is detecting a second display for reasons unknown.


Also maybe try lowering your memory clocks.
On 3D Mark it never drops even when loading tests.....either your memory is prepping to go and/or the drivers/vbios are acting up and keeping them at max this whole time.

 

[Lillie]

You on Win 11 bro how come