[SOLVED] 9700k + 3080 underperforming

Toggle sidebar Toggle sidebar
S

striderjohn

Oct 8, 2022
16
2
15
Hey all- my pc is making me want to pull my hair out.

Specs:
i7 9700k
ID Cooling Auraflow
Zotac Trinity OC 3080 12gb
ASUS Prime z390-a
32gb of Corsair Vengeance RGB Pro @3200 mhz
EVGA 700 GD
PCPartPicker: https://pcpartpicker.com/list/YjFY8r
UserBenchmark: https://www.userbenchmark.com/UserRun/55765951

Basically, I've been playing Spider-Man Remastered and realized my frames are extremely low for my hardware. I'm getting anywhere from low 20s-low 40s depending on the scene, with and without Ray Tracing and DLSS doesn't make much of a difference. I also launched Elden Ring to see how it would perform, and it has a similar outcome, 30-45 fps depending on the scene. It's probably worth mentioning that I've just recently bought the 3080, less than a month ago. Upgraded from a 2070 Super. I've been going through tons of threads on here and found that many people have similar issues, but none of the fixes from those threads seem to help at all.
I'm not sure if it's worth mentioning, but in games like Valorant and Overwatch I get the expected 300-400 fps.
I noticed that in UserBenchmark(linked above) my gpu seems to be underperforming as a whole. My mobo only has PCIe 3.0, so maybe that's why? I don't think that would affect my in-game performance as much as it's lacking though. I do plan on upgrading to Ryzen 7000 series relatively soon but that will take a while to save the money since I have to basically build a new PC for it, so I'd like to get this PC performing as much as possible.
Temps seem relatively normal, CPU normally sitting around low-mid 70s(sometimes going to mid 80s but not for long), and GPU normally sitting high 60s-low 70s.
GPU Usage normally sits in the high 80s-90s while CPU is normally in the 70s.

What I've tried so far:
-Reinstalling drivers
-Uninstalling current drivers(using DDU) and running a previous set
-Uninstalling drivers and GeForce Experience and installed drivers directly from NVIDIA's website
-Fully uninstalling drivers and trying to run games like that(didn't expect much but I'm desperate)
-Updating BIOS
-Reseating GPU, reseating RAM, unplugging and replugging every power cable (8pin CPU, 6pins GPU, 24pin MOBO)
-I fully wiped my PC today as a last resort, and even that didn't affect anything.

I think I included most details but if you have any questions please ask, I've been trying different things for hours on end for 2 days now and am getting to my wits end.



Edit for people who may be having a similar issue, didnt find a proper fix unfortunately. The PCIe slot on my motherboard is locking at x1, rather than x16(to check this, run the render test on GPU-Z, and look at Bus Interface. Should look like 'PCIe x16 4.0 @ x16 4.0 or whatever your GPU supports.), and its happening across two different GPUs. I assume my motherboard is just faulty. Will be upgrading soon.
 
Last edited:
Sort by date Sort by votes
S

striderjohn

Oct 8, 2022
16
2
15
Hi I'd like to correct something- did a test of Spider-Man with afterburner just to get the most accurate stats, here's what was happening:

Graphics on High preset, RT on, quality high, DLSS on balanced.

Depending on the scene, ranged from 20-45 fps. Averaging at around 30.

GPU Temp sat at mid 60s, usage in the 90s
CPU temp sat at low 70s, usage in the high 50s-low 60s

This was after around 10 minutes of just swinging around, testing different scenes.
Not sure if I misremembered or if afterburner is more accurate than task manager/ core temps readings. Sorry about that.
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
Phaaze88 said:
Monitor gpu core frequency, gpu hot spot, memory junction, and Performance Cap Reason too.
To view memory junction, HWINFO or Gpu-Z should be used.

Run the PC without the cable extensions.

Try the gpu in another PC if possible.
Click to expand...

Thanks for the quick reply!
No changes in FPS when cable extensions were removed (honestly kinda happy about this I didnt want to sacrifice aesthetics LOL)
GPU core frequency was at 1995, hot spot maxed at 68.3(averaging 62.5), memory junction maxed at 72(averaging 64.8) and performance limit says utilization is the cause.

Unfortunately don't have another system that I could test my GPU in.
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
spentshells said:
When you say fresh install of windows can you explain this process?

From where I am sitting that should be an amazing setup.
Click to expand...
I assume you mean when I fully wiped my PC? I used the windows reset function, and I selected fully wiping both drives and cleaning.

I thought this would be a great setup too! That's why I'm getting frustrated
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
Phaaze88 said:
Did it hang around that frequency most of the time, or no?
What of Performance Cap Reasons? There's usually at least 1 of the 5 triggered, and some are more favorable than others.
Click to expand...
I noticed that the frequency jumped between 1995 and 2010 a couple times after posting my first reply.
The only performance cap reason labeled 'yes' was utilization. I am new to this application so I could be looking at the wrong thing, heres a screenshot of what I'm looking at.
K4FuFW3.png
 
Upvote 0 Downvote
Phaaze88

Phaaze88

Titan
Ambassador
Dec 30, 2016
21,138
5,552
104,690
Utilization is one of the more favorable ones. It's pretty much always yes.
Reliability voltage has a yes next to it. It's not bad - the term means the gpu core is unable to step up to the next bin on the boost curve because the core will become unstable at the required voltage. So a boost clock of say, 2025mhz likely can't be sustained by the silicon.

Using Gpu-Z:
On the GPU tab, click on the '?' next to bus interface. It should bring up another window - the render test. This allows you to see what the actual PCIe link speed is when the card is under load.
Some boards reduce it to save power when the card isn't in use.

If that checks out fine, then it probably isn't the gpu.
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
Phaaze88 said:
Utilization is one of the more favorable ones. It's pretty much always yes.
Reliability voltage has a yes next to it. It's not bad - the term means the gpu core is unable to step up to the next bin on the boost curve because the core will become unstable at the required voltage. So a boost clock of say, 2025mhz likely can't be sustained by the silicon.

Using Gpu-Z:
On the GPU tab, click on the '?' next to bus interface. It should bring up another window - the render test. This allows you to see what the actual PCIe link speed is when the card is under load.
Some boards reduce it to save power when the card isn't in use.

If that checks out fine, then it probably isn't the gpu.
Click to expand...
I let it run for a couple minutes before this screenshot, but I don't know what I'm looking for here.
HZXLcyJ.png
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
striderjohn said:
Yeah, didnt change at all. I did some research, does this mean its stuck running at pcie x1 rather than x16?
Click to expand...
I switched my gpu from my top pcie slot to the second and launched gpu-z, it was at x16 4.0@ x8 (I forgot the rest I was rushing to get the gpu out because it has almost no clearance because the slot is so low)
 
Upvote 0 Downvote
Phaaze88

Phaaze88

Titan
Ambassador
Dec 30, 2016
21,138
5,552
104,690
striderjohn said:
Yeah, didnt change at all. I did some research, does this mean its stuck running at pcie x1 rather than x16?
Click to expand...
The gpu's Gen 4 bus interface is backwards-compatible with 3.0 boards. Since the motherboard doesn't support 4.0, can you go into the bios' advanced mode and manually change the top slot's link speed to 3.0?
[FYI: x16 4.0 @ x8 is equivalent to x16 3.0 @ x16.]

Bios > Advanced Mode > advanced menu > PCH Configuration. It should be there.
Afterwards run Gpu-Z render test again and see what the bus interface runs at.
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
Phaaze88 said:
The gpu's Gen 4 bus interface is backwards-compatible with 3.0 boards. Since the motherboard doesn't support 4.0, can you go into the bios' advanced mode and manually change the top slot's link speed to 3.0?
[FYI: x16 4.0 @ x8 is equivalent to x16 3.0 @ x16.]

Bios > Advanced Mode > advanced menu > PCH Configuration. It should be there.
Afterwards run Gpu-Z render test again and see what the bus interface runs at.
Click to expand...
Switched it to 3.0, but its still running at x1 3.0.
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
update- I switched in my old 2070 super and it has the exact same issue, so I'm thinking its a motherboard issue? I've messed with about every bios setting I can find and nothing has worked
I also launched spider-man for fun and I'm getting the EXACT performance on this card that I was on my 3080.
sGpuYFE.png
 
Upvote 0 Downvote
S

striderjohn

Oct 8, 2022
16
2
15
Phaaze88 said:
Yeah, looks like the PCIe bandwidth is stuck. That's not caused by the psu, as far as I know.
You MIGHT need to RMA the mobo.

This is a shot in the dark, but you didn't RAID the two SSDs, did you?
How about remove the gpu, blow into the slot, and then reinstall it?
Click to expand...
I didn't RAID them, no. I reseated both gpus multiple times and blew into the slot a few times as well, no dice unfortunately. I plan on pretty much rebuilding the entire pc next month with a Ryzen 7900x, so I think I'll just deal with the lower performance until then. Unless there's risk with my gpu running at x1?
 
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