Question PC underperforming in games and crashing pc

[yunoisdead]

My pc has been underperforming and fully shutting off while playing games for as long as ive had it. Its worked well in games for about a month or two after i built it and then after that its been getting really bad and crashing the whole pc in games and getting very low fps. Ive done a bunch of benchmarks for all my components and my temps were always fine ive sent back my gpu motherboard cpu cooler and psu but they said that they were all completely fine. all my drivers are updated to the latest update my bios too and ive reinstalled windows through usb and windows itself plenty of times i used to average 400 fps in a league game now i average 140 fps. I have reseached what would possibly be wrong for over 2 years and no one could ever help. im so done with it and i just want it to work well. are you experienced and you think you could help me?? please send me a message my discord and telegram are yukkicio anything helps you can also reply under this thread. i can post some benchmarks of my gpu and cpu for some more info yes ive also tried other games and i also got worse results than ive seen from other people. ive opened so many threads asking so many people for help but it never got anywhere. sometimes my system was fine and worked well with good fps other times not. what do you think that it could be?
https://prnt.sc/CxsGvLzCImN1
specs:
rtx 3080
amd ryzen 9 5900x
16 gb of 3200mhz corsair ram
850w power supply
6 fans for cooling and an aio with 3 fans
https://prnt.sc/SXTfYHvpC8RK

 

[Aeacus]

850w power supply

PSU make and model (or part number) is? Also, how old the PSU is, and was the PSU bought new or used/refurbished?

 

[yunoisdead]

PSU make and model (or part number) is? Also, how old the PSU is, and was the PSU bought new or used/refurbished?

RM White Series™ RM850 — 850 Watt 80 PLUS Gold Fully Modular ATX PSU
PSU was new upon buying and 2 years old now

 

[Aeacus]

Corsair RM-series is mediocre quality PSU. Not the worst but not the best either. So, it could be PSU issue or it could be not. Only way to know for sure, is to test with 2nd, good quality unit.

Though, 850W unit for RTX 3080 is little, due to GPU transient power spikes. It also explains the:

and fully shutting off while playing games

RTX 3080 is 320W GPU and it has GPU transient power spikes of 2 - 2.5 times of GPU TDP.

So, to run RTX 3080, i'd either use:
1000W ATX 2.2 / ATX 2.5 PSU
or
850W ATX 3.0 / ATX 3.1 PSU.

This is needed to create wattage buffer, to soak in GPU transient power spikes. pre-ATX 3.0 PSUs need higher wattage capacity, while post-ATX 3.0 PSUs are already built as such, where they can output double their rated wattage for a short time, to soak in transient power spikes.

What are GPU transient power spikes? A video to watch:

View: https://www.youtube.com/watch?v=wnRyyCsuHFQ


That being said, good 1kW PSUs to go for, are: Seasonic Focus/Vertex/PRIME, Corsair RMx/RMi/HXi/AXi, Super Flower Leadex Gold/Platinum/Titanium.
And here is list of best ATX 3.0/3.1 PSUs, link: https://hwbusters.com/best_picks/best-atxv3-pcie5-ready-psus-picks-hardware-busters/4/

So, 1st order of business: new, good/great quality PSU that can actually sustain GPU transient power spikes.

and getting very low fps.

This is different symptom. May be related to PSU but often, it is not. Still, new PSU would be my 1st step troubleshooting this issue. Once PSU is ruled out and your system doesn't shut down during gaming anymore, then there's reason to further troubleshoot the remaining issue(s), IF there is any left with new PSU.

 

[yunoisdead]

Corsair RM-series is mediocre quality PSU. Not the worst but not the best either. So, it could be PSU issue or it could be not. Only way to know for sure, is to test with 2nd, good quality unit.

Though, 850W unit for RTX 3080 is little, due to GPU transient power spikes. It also explains the:

RTX 3080 is 320W GPU and it has GPU transient power spikes of 2 - 2.5 times of GPU TDP.

So, to run RTX 3080, i'd either use:
1000W ATX 2.2 / ATX 2.5 PSU
or
850W ATX 3.0 / ATX 3.1 PSU.

This is needed to create wattage buffer, to soak in GPU transient power spikes. pre-ATX 3.0 PSUs need higher wattage capacity, while post-ATX 3.0 PSUs are already built as such, where they can output double their rated wattage for a short time, to soak in transient power spikes.

What are GPU transient power spikes? A video to watch:

View: https://www.youtube.com/watch?v=wnRyyCsuHFQ


That being said, good 1kW PSUs to go for, are: Seasonic Focus/Vertex/PRIME, Corsair RMx/RMi/HXi/AXi, Super Flower Leadex Gold/Platinum/Titanium.
And here is list of best ATX 3.0/3.1 PSUs, link: https://hwbusters.com/best_picks/best-atxv3-pcie5-ready-psus-picks-hardware-busters/4/

So, 1st order of business: new, good/great quality PSU that can actually sustain GPU transient power spikes.


This is different symptom. May be related to PSU but often, it is not. Still, new PSU would be my 1st step troubleshooting this issue. Once PSU is ruled out and your system doesn't shut down during gaming anymore, then there's reason to further troubleshoot the remaining issue(s), IF there is any left with new PSU.

the game crashing was very on and off too, only happened on rainbow six siege texas chainsaw massacre and gta 5 and was for like periods of times like one month it would crash a lot and the other month not at all and it would still not explain the fps

 

[Aeacus]

Random reboots/shut downs are mainly caused by 2 things;
1. CPU/GPU overheats and to prevent any damage, system shuts down.
2. PSU fails to deliver enough power to the GPU or fails to keep smooth enough voltage for PC's operation.

If it would've been temps issue, it could've been that your room ambient temp was higher at that period of time, enough for push CPU/GPU into thermal throttle and shut down.
Though, PSU being an issue is more likely, since 850W unit isn't enough for RTX 3080.

Simple math: 320W GPU can spike to 640W-800W. Your CPU is 105W (more if you OCd it). And add the rest of the system to it at ~100W, you're looking all combined with GPU transient power spikes: ~840W-1000W.
850W unit can sustain 840W load. But it can not stand 1kW load. More like ~920W for brief period of time.

1kW PSU can sustain 1kW load. It even can sustain ~1.1kW load for brief period of time, making 1kW pre-ATX 3.0 PSU enough. Or 850W ATX 3.0/3.1 PSU, that is designed to sustain 1700W (1.7kW) for brief period of time.

it would still not explain the fps

Poor FPS could be GPU driver issue, CPU issue (e.g PB2 or PBO values). Or even PSU's high ripple can cause that. Hence why to start with new, good quality PSU. It rules out quite a bit, not to mention curing one of the symptoms (shut down issue).

 

[yunoisdead]

Random reboots/shut downs are mainly caused by 2 things;
1. CPU/GPU overheats and to prevent any damage, system shuts down.
2. PSU fails to deliver enough power to the GPU or fails to keep smooth enough voltage for PC's operation.

If it would've been temps issue, it could've been that your room ambient temp was higher at that period of time, enough for push CPU/GPU into thermal throttle and shut down.
Though, PSU being an issue is more likely, since 850W unit isn't enough for RTX 3080.

Simple math: 320W GPU can spike to 640W-800W. Your CPU is 105W (more if you OCd it). And add the rest of the system to it at ~100W, you're looking all combined with GPU transient power spikes: ~840W-1000W.
850W unit can sustain 840W load. But it can not stand 1kW load. More like ~920W for brief period of time.

1kW PSU can sustain 1kW load. It even can sustain ~1.1kW load for brief period of time, making 1kW pre-ATX 3.0 PSU enough. Or 850W ATX 3.0/3.1 PSU, that is designed to sustain 1700W (1.7kW) for brief period of time.


Poor FPS could be GPU driver issue, CPU issue (e.g PB2 or PBO values). Or even PSU's high ripple can cause that. Hence why to start with new, good quality PSU. It rules out quite a bit, not to mention curing one of the symptoms (shut down issue).

Thank you ill look into a better psu with more wattage. I must note that my room runs really hot and all my family members say that my room is always hot but i never see my temps really high. The fps is my main problem, Ive tried gpu driver updates cpu driver updates and all those things. Also i wouldnt understand why the benchmarks looked decent, unless you believe theyre underperforming.

 

[Aeacus]

Also i wouldnt understand why the benchmarks looked decent, unless you believe theyre underperforming.

Depends on what bench you're running. Cinebench for CPU is good (other two good ones are Prime95 and AIDA64), but for GPU, i'd use Superposition,
link: https://benchmark.unigine.com/superposition

1080p Extreme preset is good one, since that's torture test. But you can use other presets as well and compare the results (score) you got, with other same GPUs. To see, if your GPU, on average, under-performs or is running good.

Then, there's also RAM testing with memtest86. 4 full passes (15 tests in 1 pass) is considered acceptable. But for 32GB of RAM, it takes ~5h per pass and ~20h for 4 passes.
Link: https://www.memtest86.com/

 

[yunoisdead]

Depends on what bench you're running. Cinebench for CPU is good (other two good ones are Prime95 and AIDA64), but for GPU, i'd use Superposition,
link: https://benchmark.unigine.com/superposition

1080p Extreme preset is good one, since that's torture test. But you can use other presets as well and compare the results (score) you got, with other same GPUs. To see, if your GPU, on average, under-performs or is running good.

Then, there's also RAM testing with memtest86. 4 full passes (15 tests in 1 pass) is considered acceptable. But for 32GB of RAM, it takes ~5h per pass and ~20h for 4 passes.
Link: https://www.memtest86.com/

i had 2 screenshots in my thread showing my results you can take a look at that. I had 4 sticks of 8 gigs of ram and did a benchtest that was like from bios or something it wasnt on windows and 3 sticks passed and i had to stop before the last one but now i only have 2 sticks in there right now.

 

[Aeacus]

i had 2 screenshots in my thread showing my results you can take a look at that.

I've already looked at those, that's how i knew that you already ran Cinebench. Though, you only did multi-core test. No test for single-core and GPU.

But multi-core result looks good.
E.g here's similar setup with R9 5900X and RTX 3080,
reddit: https://www.reddit.com/r/pcmasterrace/comments/16o3fao/cinebench_2024_r9_5900x_rtx_3080_benchmark/
With a diff that this person OC'd their chip to get a bit better multi-core result than you did.

As for 3DMark, well, your GPU performs average,
leaderboard: https://www.3dmark.com/search#advan...ck=&maxGpuMemClock=&minCpuClock=&maxCpuClock=

and did a benchtest that was like from bios or something

"or something" is very vague. With PCs, you need to be precise on what troubleshooting steps you actually did and what the result was.

Memtest86 does launch outside of Win as well, but it doesn't test one DIMM after another, instead it tests all DIMMs at once. And in one full pass, there are 15 tests in total, that takes some time. For 2x 8GB, you're looking ~2.5h per one pass or ~10h for 4 passes.

Memtest86 is only way to know, if RAM works as it is supposed to.
Here is full rundown what each and every test in memtest86 does,
link: https://www.memtest86.com/tech_individual-test-descr.html

For absolute result, 32 full passes are needed. Since Test #7 needs 32 passes for it to check every possible option. But no-one in their right mind is going to do 32 full passes, since it takes forever + then some. (For 2x 8GB, ~80h for 32 passes.)
1 full pass is bare minimum. 2 full passes is better, while 4 full passes is considered acceptable when testing RAM. Since RAM testing takes a lot of time, best to let it run overnight (when you're sleeping).

 

[yunoisdead]

I've already looked at those, that's how i knew that you already ran Cinebench. Though, you only did multi-core test. No test for single-core and GPU.

But multi-core result looks good.
E.g here's similar setup with R9 5900X and RTX 3080,
reddit: https://www.reddit.com/r/pcmasterrace/comments/16o3fao/cinebench_2024_r9_5900x_rtx_3080_benchmark/
With a diff that this person OC'd their chip to get a bit better multi-core result than you did.

As for 3DMark, well, your GPU performs average,
leaderboard: https://www.3dmark.com/search#advanced?test=sw DX&cpuId=2758&gpuId=1338&gpuCount=0&gpuType=ALL&deviceType=ALL&storageModel=ALL&showRamDisks=false&memoryChannels=0&country=&scoreType=overallScore&hofMode=false&showInvalidResults=false&freeParams=&minGpuCoreClock=&maxGpuCoreClock=&minGpuMemClock=&maxGpuMemClock=&minCpuClock=&maxCpuClock=


"or something" is very vague. With PCs, you need to be precise on what troubleshooting steps you actually did and what the result was.

Memtest86 does launch outside of Win as well, but it doesn't test one DIMM after another, instead it tests all DIMMs at once. And in one full pass, there are 15 tests in total, that takes some time. For 2x 8GB, you're looking ~2.5h per one pass or ~10h for 4 passes.

Memtest86 is only way to know, if RAM works as it is supposed to.
Here is full rundown what each and every test in memtest86 does,
link: https://www.memtest86.com/tech_individual-test-descr.html

For absolute result, 32 full passes are needed. Since Test #7 needs 32 passes for it to check every possible option. But no-one in their right mind is going to do 32 full passes, since it takes forever + then some. (For 2x 8GB, ~80h for 32 passes.)
1 full pass is bare minimum. 2 full passes is better, while 4 full passes is considered acceptable when testing RAM. Since RAM testing takes a lot of time, best to let it run overnight (when you're sleeping).

ill do a single core test soon, and yes it was memtest i think i had 3 passes if i remember correctly