• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Symptoms of an overheating GPU?

Associate
Joined
16 Nov 2005
Posts
111
Hey all, hoping I might get some advice and not really sure the best place to post this but figured I’d try here and apologies for the war and peace

I’ve got a powercolor 6800xt red dragon that regularly hits junction temps of between 100-110deg C when playing RDR2 and the game freezes requiring a hard reset

appreciate there’s a lot written about RDR and freezes and assumed its the game but now wondering if the temperature might be the cause

I’d always thought a GPU would cause artefacts when it overheats and that the card would throttle before that and I’m not getting any of this

I’ve tried the only other game I’ve bothered to install since rebuilding which was stellaris but obviously it doesn’t push the card or the cpu anywhere near as hard and temps are a lot lower with no stability problems

Air flow in the case isn’t great (phantek evolv x) and previously I was using an r9 290 and 5820k oc’d @ 4.5ghz in a custom loop with cooling via 240 slim and 360 deep rad and never had any issues with temps

unfortunately there’s no water block specifically for the card so I’m having to rely on the air cooler and I also have a pci WiFi card installed directly below which is less than ideal but I can’t really do anything about

in order to fit the GPU I’ve had to strip the pc apart and lose the 360 rad so
Cpu is on its own and using the 240 slim with temps into the 70’s vs 60’s previously

for context specs as follows

5820K @ 4.5ghz
X99 krait
16gb hyper x
Corsair rm850
6800xt
PCI WiFi card

The case currently has the 240rad installed in the top with 2x Corsair spl 120’s intaking into case

3x Corsair spl 120’s installed in the front as intakes however I have an ek d5xres 140 fitted against the middle fan which isn’t helping air flow

and finally a Corsair af 140 in the rear as an exhaust

I’ve tried cracking the case open and having a tower fan blowing into it and not sure how much this has helped as I’m still getting freezes (again not sure if this down to temps or just the game)

Could some one give me some advice on how I can diagnose this via a suitable bench mark software or something

or tell me it’s all fine and it’s just the game lol

if it is overheating I’m at a loss on how to fix this short of buying a new case with better air flow, going with a usb WiFi card or going out on a limb with the Red Devil water cooler and hope it fits the red dragon (I checked images of the PCB’s and they did look different but not sure if it’s enough to make it not fit)

none of these are particularly great as it means me having to explain why yet more boxes from Overclockers are arriving to the wife after the UW screen, card psu etc etc

advise greatly appreciated...
 

TrM

TrM

Associate
Joined
3 Jul 2019
Posts
744
first amd are designed to hit 100-110 junction temp. since junction temp is just the hotest spot on the die and not the overall temp. amd only throttle to keep below 110 degress on the hostest part of the die

what is the overall temp on the 6800xt? i have a 6800xt gigabyte gaming oc and my junction temp never hit above 90 degress. and i have the same case.

what is your fan speed on the card? have you used afterburner or in the amd drivers to adjust the graphic card fan speed? whats the gpu clock speed?
 
Associate
Joined
12 Jan 2021
Posts
1,296
I haven't tried it but I imagine RDR2 allows the use of Radeon Chill. You could use it to set fps at say 60 min, 65 max. This will reduce the GPU temps and you can see if the game still crashes with lower temps.

https://youtu.be/94ti1ZuIFhA
 
Associate
OP
Joined
16 Nov 2005
Posts
111
Thanks for the replies

trm, I guessed for it to throttle at 110 it must be within spec but it just feels crazy hot in the case

if by overall temp you mean the core temp it’s usually in the 90’s but I’ll need to check next time, same with clock speed

I havnt messed with the fan speeds but will give it a try in the driver which truth be told I havnt really looked at been using hwmonitor for values

Im interested what your arrangement in the case is thou

Asmb,

that’s not a bad idea I’ll drop all the settings and give it a go - currently I’m only getting 60 fps but that’s with everything maxed at 3440x1440

I’m suffering massively from buyers remorse at the moment especially as for an extra 50quid I could have bought a red devil and had a water block on the thing
 
Associate
OP
Joined
16 Nov 2005
Posts
111
So did some testing based on playing with the fan speed in the drivers

did 7 runs of the rdr benchmark back to back and ended up with the following with me adjusting max fan speed %

100% fan, max clock 2463, core temp 64, HS temp 87

80%, 2463, 68, 91

70%, 2482, 72, 94

60, 2482, 75, 97

at this point I was feeling good as temps looked fine and had heat soaked enough so decided set up a custom fan profile based on what sounded like a compromise in terms of acoustics got half way through with temps looking fine on overlay before it locked and rebooted itself - not good

During all the previous runs cpu package temp was maxing out in and around the 60’s and hadn’t been tracking it, I’ve got Icue set to change the keyboard and mouse colour based on temp with red set to anything over 75

And surprise, surprise during last run keyboard turned scarlet just before it reset...

Guess I need to look into a different avenue to resolve this
 
Associate
Joined
2 Feb 2021
Posts
206
Over heating should reduce framerate but not lock the GPU, I think you need to start looking at RMA your card (try a different card first)

Set it on a loop on 3D Mark and see if it freezes or errors out, if it does and doesn't on another card you have a RMA problem.
 
Associate
OP
Joined
16 Nov 2005
Posts
111
I think the problem is with the cpu had an overlay up when it crashed out and GPU temps were fine

need to do some more testing on the cpu but have a feeling a new case is going to be required
 
Associate
Joined
12 Jan 2021
Posts
1,296
There aren't any decent reviews of this gpu. I did read a review of the 6800 version and it had an overclock with a large increase in power usage, nearly the same usage as the 6800xt. Have you tried the silent bios switch, I assume it has one, and looked at the temperatures. You could also look to reduce the clocks back to the reference card, or reduce the power use slider in radeon software, sucks getting a non reference card though that might have a bios overclock which might cause such high temperatures. A search of the internet did find someone else with your kind of temperatures so not sure a new case would help
 
Soldato
Joined
24 Sep 2013
Posts
2,890
Location
Exmouth, Devon
Buyers remorse - well, bet there's 1,000's itching to be in your position. You're just not enjoying at the moment due to these annoying crashes.

Get a length of CAT6 cable from your router to your PC and take the WiFi card out. Is it a game freeze and it returns to desktop after a while or a full pC lockup where you need to reset or restart the PC? USe Whocrashed (free) that can be offer more specific data to the crashes. This only works for full system lockups, blue-screens etc. Are the 2 GPU power cables separate to the PSU or a Y connector?

When it locks up, note the time, then use Event Viewer (search in taskbar)> custom views> administrative events, to see what the errors are at that time. Share these here too.

Annoying after you've built a PC but forums are here to help so you are in the right place.
 
Associate
Joined
13 Sep 2010
Posts
1,970
During all the previous runs cpu package temp was maxing out in and around the 60’s and hadn’t been tracking it, I’ve got Icue set to change the keyboard and mouse colour based on temp with red set to anything over 75

And surprise, surprise during last run keyboard turned scarlet just before it reset...

Guess I need to look into a different avenue to resolve this

75 isn't too hot for a CPU, but it would be interesting to know how far above that it is going. You could enable logging in Coretemp to log the results to an excel file and hopefully it will have time to record the data before crashing!

Given the OC on your CPU (and potentially the high temps), I guess it could be drawing a lot of power, so in my mind the next step would be a high CPU only load such as Prime95 or similar (apologies if you've already tried this, wasn't sure from what you have described);
This will tell you whether the CPU is crashing due to temp, or if not, I suspect the total power draw may be too much (even though 850 is over the recommended spec), given the pump/fans etc. (and taking the graphhics load out of the equation will bring it much lower)

Also have you checked your CPU thhermal paste application recently?! Might be wort a look if it's dried out over a few years

Hope that helps!
 
Associate
OP
Joined
16 Nov 2005
Posts
111
Thanks for the replies before I answer some of your points by way of an update

I’ve just spent the last hour running rdr with both side panels of the case cracked open and the front panel and dust filter removed with tower fan blowing into it and temps were absolutely fine and no crashes and when I say crash I’m not talking about crash to desktop or blue screen literally frozen screen and audio with no response to key board inputs only hard reset assuming it hasn’t rebooted itself

I’m now pretty certain it’s cpu temps causing the problems, I’m a firm believer if something looks right it usually is, and the cooling setup for the cpu basically doesn’t

I went from a 240 x 25mm deep and 360 x 65mm deep rad cooling my previous loop. To just the 240 cooling the cpu

obviously the 360 was doing all the work as 240 in the top of the case gets f all airflow and that’s pretty evident now as it’s clearly not cutting it with just the cpu in the loop

previous loop used to take a litre worth of coolant, current arrangement is less than a third so that’s a lot of thermal mass to loose

Getting rid of the WiFi card and running cat 6 isn’t gonna happen soon unfortunately, I live in a town house with the pc now relegated to the top floor and the router in an extension on the side of the house so short of running it externally up the house its a none starter (and believe me I thought about doing exactly this a few times since lockdown began)

I mentioned buyers remorse and that stems from having given up on any form of gaming for the last 3 years due to 2 newborns
So decided to treat myself on some long overdue upgrades and cost never really came into to it so yeah guess it was a good position to be in but for an extra 50quid I’d have bought a Red Devil and it would have been under water and I doubt I’d be having these problems

To resolve this I need to beef up the cooling to the cpu, I might try flipping the rad fans to exhaust from the case but can’t see how that’s likely to help other than airflow across the rad might increase but it’ll be with a worse air on temp due to the case being obviously hotter than ambient

really it needs a bigger rad and that won’t fit in the current case so looks like that’ll be the next purchase

first world problems lol
 
Associate
OP
Joined
16 Nov 2005
Posts
111
Congrats on your bairns!

What does event veiwer or Whocrashed say?
https://www.resplendence.com/whocrashed
sorry forgot the link before

cheers mate, should have been clearer I gave up gaming when they were born but their now 2 1/2 and the other 4 next week and hence finally having time to get back into gaming, albeit for about an hour every other night - still their worth any sacrifice

I’ll take a look at event viewer or d/l whocrashed and try and replicate the issues as I’m curious to know the cause and will report back
 
Soldato
Joined
30 Mar 2016
Posts
2,884
Location
East Lothian
My instinct from everything you've described is that it's not temps causing your problem. 75 degrees at 60% fan speed sure doesn't sound like there anything wrong with GPU, that's really quite good. I would check your RAM sticks are seated correctly (I've been caught out twice with this - you hear the click and think they're in but sometimes they're not fully locked down). Often these issues turn out to by PSU related either through a fault or power spikes it can't handle. I would definitely test on a couple more games that will push your system too. It's unlikely to be just a problem with RDR2 I would say...
 
Associate
OP
Joined
16 Nov 2005
Posts
111
Yeah I think the GPU is actually ok, especially after i sat and did some proper testing with It

I think the previous high junction temps was due to heat soak but looking at the recent recent results the core temp seems to track 20 degrees below so bit more relaxed about it now

It would be massively irritating if it was down to the psu given I only just bought it when I took delivery of the card

I’ll reseat the ram to rule it out and check the cables from the psu

If I’m honest I’m not massively happy with the way I applied the thermal paste when I stripped the block down to clean it out but I doubt that causing the issues

I’ll do some further testing on the GPU and cpu and keep an eye on power draw and see what’s causing the lock ups
 
Soldato
Joined
11 Aug 2012
Posts
2,592
Location
Scotland
I was getting crashing with my 3060ti in a system that was stable with my old 980ti. Eventually put my ram back to stock by disabling xmp and was fine with no crashes. Tried adding voltage with xmp but still crashed. I then down clocked ram xmp from 3200 to 3000 and it's been fine for the last week. I assume the new card exposed ram instability.
 
Associate
OP
Joined
16 Nov 2005
Posts
111
To be fair hasn’t really considered the ram also don’t know much about it other than turning xmp on certainly isn’t over clocked beyond it’s rated speed

I’ll add to the ever growing list of things to check

availability aside at this point Im wishing I’d upgraded the cpu, mobo and ram
 
Soldato
Joined
11 Aug 2012
Posts
2,592
Location
Scotland
To be fair hasn’t really considered the ram also don’t know much about it other than turning xmp on certainly isn’t over clocked beyond it’s rated speed

I’ll add to the ever growing list of things to check

availability aside at this point Im wishing I’d upgraded the cpu, mobo and ram
I was also getting the odd blue screen with random memory errors but sometimes game would just crash. That's what led me to the ram.
 
Soldato
Joined
30 Mar 2016
Posts
2,884
Location
East Lothian
Yeah I think the GPU is actually ok, especially after i sat and did some proper testing with It

I think the previous high junction temps was due to heat soak but looking at the recent recent results the core temp seems to track 20 degrees below so bit more relaxed about it now

It would be massively irritating if it was down to the psu given I only just bought it when I took delivery of the card

I’ll reseat the ram to rule it out and check the cables from the psu

If I’m honest I’m not massively happy with the way I applied the thermal paste when I stripped the block down to clean it out but I doubt that causing the issues

I’ll do some further testing on the GPU and cpu and keep an eye on power draw and see what’s causing the lock ups
Yeah you'd like to think the PSU would be ok but I've had a faulty PSU out of the box before. Unlikely but it can happen. But yeah I think these issues are usually either RAM or PSU related before temps. Without trying to be a bad influence, I upgraded to a 5800X, B550 mobo and nice RGB high speed RAM using Overclockers Finance option, worked out at £24 a month over 3 years... :D
 
Back
Top Bottom