Question i7-10700K + RTX 3080 Ti + 32 GB RAM = Massive FPS drops ?

Toggle sidebar Toggle sidebar
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
So, I've assembled this PC last year and it worked fine for some games, but I think that those (like Civiization VI) were not very demanding, then this year I've tested Cyberpunk and recently Elden Ring and the experience sucks.
FPS will randomly but regularly drop to less than 10 and I can't figure out why.

Video with CPU, GPU and RAM indicators on screen during a FPS drop.

So far I've tested:
  • Update my drivers and tweak the settings on Nvidia control panel;
  • Format Windows drive and re-installed it with fresh drives;
  • Installed the games on different storage unities (including an M.2 SSD);
  • Tested my RTX on other computer (it worked fine);
  • Tested a 1000W PSU on my PC (didn't make any difference);
  • Tested another RTX 3080 Ti on my PC (didn't make any difference);
I've been monitoring my thermals while playing both games and there are no temperature spikes, I improved my PC ventilation anyway with a better case and added a (freaking) NZXT Kraken cooler to my CPU, the thing is so cool that my RTX fans stay idle whenever I'm not playing.

And yet I can't figure out the problem...
According to my tests, Elden Ring uses only 4 out of my 16 logic cores (but I gess this is expected).
xz4L7rt.jpg


Any help would be greatly appreciated.

Edit - Full specs:
  • Intel(R) Core(TM) i7-10700K CPU @ 3.80GHz
  • Gigabyte GeForce RTX 3080 Ti Vision OC 12GB
  • 32 GB of RAM (4x HyperX Fury, 8GB, 2400MHz, DDR4, CL15)
  • Mobo: Gigabyte B460M DS3H, Intel LGA 1200, mATX, DDR4
  • SSD PNY CS3030 250GB M.2
  • Cougar CMX700 700W 80 Plus Bronze
  • Windows 10 (v. 10.0.19041)
 
Last edited:
Sort by date Sort by votes
Roland Of Gilead

Roland Of Gilead

Titan
Aug 8, 2008
11,404
2,681
80,240
3ernard said:
So, I've assembled this PC last year and it worked fine for some games, but I think that those (like Civiization VI) were not very demanding, then this year I've tested Cyberpunk and recently Elden Ring and the experience sucks.
FPS will randomly but regularly drop to less than 10 and I can't figure out why.

Video with CPU, GPU and RAM indicators on screen during a FPS drop.

So far I've tested:
  • Update my drivers and tweak the settings on Nvidia control panel;
  • Format Windows drive and re-installed it with fresh drives;
  • Installed the games on different storage unities (including an M.2 SSD);
  • Tested my RTX on other computer (it worked fine);
  • Tested a 1000W PSU on my PC (didn't make any difference);
  • Tested another RTX 3080 Ti on my PC (didn't make any difference);
I've been monitoring my thermals while playing both games and there are no temperature spikes, I improved my PC ventilation anyway with a better case and added a (freaking) NZXT Kraken cooler to my CPU, the thing is so cool that my RTX fans stay idle whenever I'm not playing.

And yet I can't figure out the problem...
According to my tests, Elden Ring uses only 4 out of my 16 logic cores (but I gess this is expected).
xz4L7rt.jpg


Any help would be greatly appreciated.
Click to expand...

Hey there,

Can you please list your exact specs in full. Make/model of GPU/Mobo/Ram/PSU too.
 
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
Upvote 0 Downvote
alceryes

alceryes

Splendid
Jun 11, 2004
4,366
942
27,090
Phaaze88 said:
@alceryes
You seeing what I'm seeing? Looks like the OP is burning up their mobo's VRM, which is throttling core clocks.
Click to expand...
The VR thermal alert isn't tripping but the electrical design limitation AND CPU thermals are. Course, it's not clear whether it's MB VR or internal CPU VR that is triggering.
The CPU is dropping down to it's default (read: emergency) frequency of 800MHz.

OP, reset ALL your overclocks, overvolts, timings, and other adjustments and test again.

Edit - @Phaaze88, You referring to the VRM MOS temp? It is a bit high but I think within spec. OP have you adjusted the power profile/amperage/frequency of the VRM of the motherboard? If yes, reset those too.
 
Last edited:
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
alceryes said:
The VR thermal alert isn't tripping but the electrical design limitation AND CPU thermals are. Course, it's not clear whether it's MB VR or internal CPU VR that is triggering.
The CPU is dropping down to it's default (read: emergency) frequency of 800MHz.

OP, reset ALL your overclocks, overvolts, timings, and other adjustments and test again.

Edit - @Phaaze88, You referring to the VRM MOS temp? It is a bit high but I think within spec. OP have you adjusted the power profile/amperage/frequency of the VRM of the motherboard? If yes, reset those too.
Click to expand...
I don't have any OC in this machine, is there any other test I might perform to clarify the matter?
gwYwoHT.jpg
 
Upvote 0 Downvote
Phaaze88

Phaaze88

Titan
Ambassador
Dec 30, 2016
21,500
5,708
104,690
3ernard said:
Sorry but I don't know much of VRM, by burning up you mean "overheating" or "surpassing its power capacity"?
Click to expand...
Overheating.
The cpu isn't pulling much power at all. It's shown staying below 100w most of the time.
VRM MOSfet is at 100C or more for most of the time, yet it's not flagged for VRM thermal throttling... but for the processor to be dropping down to 800mhz is a power issue, not thermal; cpu thermal throttling doesn't drop the clocks down as far - more like 1.2ghz -2.4ghz.
There's also several IA: PROCHOT events, but it's vague. It doesn't specify internal or external sensors, or it could be a faulty sensor.


So far, from what I can tell, the gpu is roasting the VRM, and there may be a faulty motherboard sensor(s).

@alceryes
It's a budget B460 board, it shouldn't have OC features?
 
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
alceryes said:
Oh, wait a min, that PSU is NOT good enough for that GPU.
All your 12V rails are low.

Test with a better PSU - could fix all your problems.
Click to expand...
I tested it with a 1000W PSU and the stutter persisted, the PSU was an early suspect for me too. It might take some days for me to borrow another PSU to test it again.
 
Upvote 0 Downvote
alceryes

alceryes

Splendid
Jun 11, 2004
4,366
942
27,090
Phaaze88 said:
I wish Steve(HWUB) did a B460 VRM thermal roundup...
Instead, there's a B560 roundup with the GB B560M DS3H version, which is listed as throttling around 100C... but I can't use this sample.
Click to expand...
Do you know what adjustment the B460 does have, with respect to Turbo and voltages?
There's no reason to think that a B460 would do BETTER than a B560. So this is definitely a possibility.
 
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
alceryes said:
Have you adjusted ANY BIOS settings or is everything stock?
Click to expand...
I remember setting the CPU fan speed to silent, let me check.

Crap...
FYorcG6.jpg


The CPU, System 1, PCH and VRM MOS temperature warning controls are disabled along with the CPU fan fail warning...
And the fan speed is not on silent mode as I thought.
 
Upvote 0 Downvote
Mcg_mm8

Mcg_mm8

Apr 15, 2022
1
0
10
As far as Elden Ring is concerned, I had the same issues with no defective hardware, and adding -force-vulkan to steam's launch properties fixed it for me.
 
Upvote 0 Downvote
alceryes

alceryes

Splendid
Jun 11, 2004
4,366
942
27,090
3ernard said:
I remember setting the CPU fan speed to silent, let me check.

Crap...
FYorcG6.jpg


The CPU, System 1, PCH and VRM MOS temperature warning controls are disabled along with the CPU fan fail warning...
And the fan speed is not on silent mode as I thought.
Click to expand...
Your CPU is definitely getting temp warnings AND is downclocking to the emergency 800MHz speed. This alone can cause the stuttering you're seeing. Add to that all your GPU rails being low and the PSU being suspect.

Try manually setting your CPU fan to 90% all the time and running another HWiNFO64 test.
 
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
Phaaze88 said:
@3ernard

What is the cooling config, by the way?

Cpu cooler: NZXT Kraken-something... where is the radiator located.

Chassis model.

How many fans, and what's intake/exhaust.
Click to expand...

The cooling setup uses a CoolerMaster MASTERCASE SL600M with the stock fans and a Water Cooler NZXT Kraken X73 360mm, on the top of the case with a pull configuration. The air flows from the bottom to the top (by design there are no openings on the front or back). There might be a low flow of cool air near the VRM, but this problem also happened on the older case (that one had openings on all sides.
3RBOnJW.jpg


My older cooler was also a water cooler, so the VRM never had that passive cooling from a CPU air cooler...
 
Upvote 0 Downvote
3ernard

3ernard

Distinguished
Dec 12, 2015
34
0
18,530
alceryes said:
Your CPU is definitely getting temp warnings AND is downclocking to the emergency 800MHz speed. This alone can cause the stuttering you're seeing. Add to that all your GPU rails being low and the PSU being suspect.

Try manually setting your CPU fan to 90% all the time and running another HWiNFO64 test.
Click to expand...
Will try this out as soon as I get home.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom