Question GPU is severely underperforming ?

Oct 27, 2024
20
1
15
Hey all,

I have this one very annoying issue with my 5700 XT GPU (probably due to very low clock speeds), it performs way worse than it should and i can't seem to find a solution. Whenever i look at benchmark videos from people with the same specs as me it looks like they are getting way more fps in games than i am with my card.

I also tried running 3dMark and the score was below average (Mine was 1180 and the average is 2164)

I've tried reinstalling drivers, using older drivers, changing my gpu settings and tried gaming with and without msi afterburner. I even tried updating my bios and tried tweaking around with the settings but this never seemed to help

And overclocking didnt seem to do anything either. The AMD community couldn't figure it out and they told me i should contact Gigabyte for this. I tried contacting gigabyte but i doubt they'll be helpful.

I'm not much of a tech guy so maybe someone else could help me with this

Thanks in advance :)
 
Welcome to the forums, newcomer!

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.
 
Welcome to the forums, newcomer!

When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.
Apologies.

CPU: Ryzen 5 5600X
CPU Cooler: Be Quiet! Dark rock pro 4
Motherboard: Gigabyte B550 Gaming X V2 (Bios Version: FE)
Ram: Corsair Vengeance DDR4 32GB 3600Mhz
SSD/HDD: Samung Evo 990 Pro 1TB, Crucial P5 Plus 500GB, 2x Toshiba HDD 500GB
GPU: Gigabyte 5700XT Gaming OC 8GB
PSU: EVGA 750 BQ 750W 80+ Bronze Semi-Modular (About 2 years old)
Chassis: Corsair 4000D Airflow
OS: Windows 11 Pro
Monitor: Samsung U28E570, LG 22MK400H


Everything in my pc is basically 2 years old except for the GPU since i bought it used (which could be why i am experiencing issues)
 
I've tried reinstalling drivers, using older drivers,
Please elaborate.

I tried completely reinstalling my drivers using DDU and even tried different versions to see if they would be more stable but it just didn't make any difference I also completely reinstalled windows 11 today and that didnt help either
I also tried running 3dMark
Which test did you run and at what settings?
i just use the demo of 3dmark and i just used default settings
Motherboard: Gigabyte B550 Gaming X V2 (Bios Version: FE)
What is the revision for the PCB on your motherboard? If you're on 1.3 or 1.4, I see you have BIOS updates pending;
https://www.gigabyte.com/Motherboard/B550-GAMING-X-V2-rev-14/support#support-dl-bios
I'm on 1.3 and i did see some newer versions but i doubt updating would help since i already had this problem ever since i bought the card and back when i updated the bios FE was the latest version



I'll update the bios right now
 
Last edited:
I'd plug it into some other machine for testing purposes to see if it acts the same? You could try another set of power wires if those are modular, or, if you have a multi meter, you could test and see if you're getting the appropriate voltage. You could check for resistance also with the PSU turned off to see if you have a short somewhere.
 
I'd plug it into some other machine for testing purposes to see if it acts the same? You could try another set of power wires if those are modular, or, if you have a multi meter, you could test and see if you're getting the appropriate voltage. You could check for resistance also with the PSU turned off to see if you have a short somewhere.
I'll try it at work tomorrow

I don't have another set of power wires and i don't know if using random power wires is a good idea since people always told me to use the wires that come with the PSU

Don't have a meter sadly


Any ideas what else could be the problem?
 
Okay, without a multimeter you can't really troubleshoot your PSU then. But if you have a modular PSU, they do sell replacement wires in case yours are bad, it happens sometimes. I mean I would replace that vs buying a new psu..
 
Okay, without a multimeter you can't really troubleshoot your PSU then. But if you have a modular PSU, they do sell replacement wires in case yours are bad, it happens sometimes. I mean I would replace that vs buying a new psu..
Mine is Semi Modular whatever that means..


I do still have a warranty for the PSU so that's good
 
Last edited:
semi means that only some wires can be detached from the unit itself. Semi modular PSUs usually offer everything else to be detached from it except for that fat 24-pin plug + the CPU power cables. If it was a fully modular, like mine, everything could be detached from it.

It could be something simple, like a bad connection. I never had any issues with EVGA PSUs, which is why I keep buying them.

Reading your other messages, you already tried a fresh windows install with new drivers, which didn't help. So, look on the physical level of things, are connections tight?

You bought a used GPU, probably a faulty one. At this point, you need to try the card on some other computer and/or the same on your PC, do you have another card you can test on?

The 5700xt - does it look like it was taken apart? I wonder of someone tried to redo the thermal pads on it and did a shotty job that they had to sell it? It could be anything really.
 
semi means that only some wires can be detached from the unit itself. Semi modular PSUs usually offer everything else to be detached from it except for that fat 24-pin plug + the CPU power cables. If it was a fully modular, like mine, everything could be detached from it.

It could be something simple, like a bad connection. I never had any issues with EVGA PSUs, which is why I keep buying them.

Reading your other messages, you already tried a fresh windows install with new drivers, which didn't help. So, look on the physical level of things, are connections tight?

I could try and look but the PSU isn't easily accessible so i'd have to take it all apart
You bought a used GPU, probably a faulty one. At this point, you need to try the card on some other computer and/or the same on your PC, do you have another card you can test on?

I have a very very old nvidia card like one of those small ones but i'll try testing the gpu at work
The 5700xt - does it look like it was taken apart? I wonder of someone tried to redo the thermal pads on it and did a shotty job that they had to sell it? It could be anything really.
Well funny thing i had the thermal pads done after buying it because the temps were getting too high and ever since then the temps have been fine except for the weird junction temperature that's at a 110c but AMD said that it's "To be expected" with these cards so i assume that's fine

Anyways i don't think it's been opened up before that i mean i bought it from this kid who said that he bought a better card
 
the temps have been fine except for the weird junction temperature that's at a 110c but AMD said that it's "To be expected" with these cards so i assume that's fine
110c is not a fine temperature lol. What do you have to do to get it to that temperature? The highest I ever see on mine is like 70-75c and that's on full load.
 
110c is not a fine temperature lol. What do you have to do to get it to that temperature? The highest I ever see on mine is like 70-75c and that's on full load.
It's the hotspot temp and believe me AMD said that is was normal with these cards

Idk i stressed about it alot but i can't do anything about it since thermal pads only helped with the "main" thermals and then i read abt ppl saying that its normal

The "main" thermals are around 70 when its on full load

I couldn't have a better airflow in my case either




https://community.amd.com/t5/gaming...e-more-control-over-gpu-power-and/ba-p/418629

you can read about it here if you want to
 
Last edited:
Your card is running abnormally hot, which leads me to believe you are getting thermal throttled given your 3dmark scores assuming you aren't having any other issues. The hotspot on my card never exceeded 70-75 degrees C. My case has 2 140mm fans in the front and 3 120mm fans, 2 up top and one in the rear.

This is my fan curve: https://prnt.sc/bqZlIDuOyHiW
 
Your card is running abnormally hot, which leads me to believe you are getting thermal throttled given your 3dmark scores assuming you aren't having any other issues. The hotspot on my card never exceeded 70-75 degrees C. My case has 2 140mm fans in the front and 3 120mm fans, 2 up top and one in the rear.

This is my fan curve: https://prnt.sc/bqZlIDuOyHiW
But AMD saying that it's normal doesnt mean anything then? I mean why would they be lying about it
 
But AMD saying that it's normal doesnt mean anything then? I mean why would they be lying about it
well lets make some things clear
gpu has some base clock, for rx 5700xt that would be 1605MHz
it will then boost when theres some higher load, how far it will boost will depend on temperatures/voltages and wattage
now if youre hitting 105+C, GPU will thermal throttle, if gpu clock is at or above 1605MHz, then its within specs, but then dont complain that its not boosting to ~2GHz under load and giving you lower perf then expected, if clock is below 1605MHz then your gpu isnt running within specs

anyway, just repaste ur GPU
 
well lets make some things clear
gpu has some base clock, for rx 5700xt that would be 1605MHz
it will then boost when theres some higher load, how far it will boost will depend on temperatures/voltages and wattage
now if youre hitting 105+C, GPU will thermal throttle, if gpu clock is at or above 1605MHz, then its within specs, but then dont complain that its not boosting to ~2GHz under load and giving you lower perf then expected, if clock is below 1605MHz then your gpu isnt running within specs

anyway, just repaste ur GPU
Don't think repasting is gonna help since that already has been done not too long ago
 
But AMD saying that it's normal doesnt mean anything then? I mean why would they be lying about it
Wondering what AMD would say about my card lol. So, you're getting thermal throttled, hence poor performance due to such high temperatures, which is most likely caused by a poor thermal pad replacement job, which I'm still trying to understand why it was needed anyway. My 4 year old card is still on stock pads/paste. But if you think AMD is totally right then I suppose /thread.
 
Wondering what AMD would say about my card lol. So, you're getting thermal throttled, hence poor performance due to such high temperatures, which is most likely caused by a poor thermal pad replacement job, which I'm still trying to understand why it was needed anyway. My 4 year old card is still on stock pads/paste. But if you think AMD is totally right then I suppose /thread.
Well the thermals were even worse when i bought it so thats why i did it
 
You said it was worse before.. I don't know how much worse it can be but what this screenshot is showing is horrible. im gonna stress test mine and post it tomorrow for you to see.

Did you replace the pads or someone else did? What pads and what thickness? The thickness of thermal pads is different.

I'd say you need to do it again with quality thermal pads and proper size/thickness.