I bought the graphics card a year and a half ago, and it used to work fine. Recently I noticed that fps in games became much lower than before. I checked the temperature - the card heats up to 89 degrees and starts throttling. Could it be a thermal paste issue or something else?
Clean the dust off your graphics card - that's the most common cause of overheating and throttling :) Dust clogs the radiator and fans, and the card starts heating up even under lighter loads. Yeah, thermal paste can dry out over a year and a half, but try cleaning it first - it's simple and often does the trick. If the card still runs hotter than 80 degrees in games after that, then think about replacing the thermal paste or checking the entire cooler.
First of all, check if your cooling system is clogged up - yeah, it's obvious, but 89 degrees while throttling is a pretty big hint! Grab a can of compressed air and blow out the radiator and fans, just do it outside or over a bathtub, otherwise dust will go everywhere. If the temperature doesn't drop at least 5-10 degrees after cleaning, then yeah, the problem is the thermal paste - you'll have to replace it, but that's trickier and riskier without experience.
Life hack: when you're cleaning, listen to see if the card's fans are making weird noises or clicking sounds - that could mean the fan blade is starting to hit the case because of dust buildup. And make sure you power everything down and wait about 5 minutes so static doesn't fry your components!
Before cleaning, make sure the card is actually getting enough airflow in the case - sometimes the issue isn't dust on the heatsink, but poor air circulation inside the PC. If the airflow is fine, then yeah, blow out the heatsink with compressed air.
After that, it makes sense to think about thermal paste - a year and a half isn't a critical timeframe, but if the card was under heavy load, the paste could have degraded. That said, replacing it takes care: you need to disassemble the cooler, remove the old paste, and if you're not confident doing that, it's better not to risk it.
89 degrees is already a sign of serious cooling failure, but there's one thing people often overlook. First, check if the thermal paste between the chip and heatsink has dried out. A year and a half is about the time when paste can lose its properties, especially if the card runs at full load a lot. That said, to do this you'll need to remove the heatsink, which requires care. Before you take it apart, it makes sense to clean the accessible parts - fans, external parts of the heatsink.
To add to what's already been said about dust and airflow in the case: make sure the fans themselves are spinning properly. Sometimes they get stuck from oxidation or dust in the bearings and can spin slower than they should, even though they look fine visually. If a fan spins slowly or with difficulty - that reduces cooling way more than it seems. Check in utilities like GPU-Z or in your motherboard's BIOS what RPMs the cooler's showing under load. If it's all below 60-70% of max at these temperatures - there's your problem.
And one more thing: make sure the graphics card is seated properly in the slot, the contacts are clean, and the PCI Express power cable is plugged in tight. Poor contact can cause voltage spikes and excessive heat in the power circuits, though this affects the memory and core temperature sensors indirectly. If everything above is fine - yeah, open up the card and replace the paste.
89 degrees is not quite critical for the RTX 4090 - throttling usually kicks in higher than that, so it's more likely a matter of dust buildup or dried-out thermal paste. I had a similar experience with my graphics card: it suddenly started running hot and losing performance, I initially figured it was a hardware issue, but then realized the fans had just gotten so clogged with dust that almost no air was getting through.
Start with the simplest thing - open up your case and take a visual look at the card's heatsink, dust and hair often accumulate between the fins in a way that's not visible from a distance. If there really is a buildup, blow it out with a can of compressed air or even a vacuum on exhaust mode (externally, don't stick anything inside). If the temps don't drop at least 5-10 degrees after cleaning, then start thinking about replacing the thermal paste - a year and a half of use could have dried it out, especially if the card's been running under heavy load a lot.
When you're cleaning the card, pay attention to how the fans are seated on the mounting - sometimes they shift slightly and start hitting something or spinning freely, which also kills cooling :) If the temperature doesn't drop after cleaning, then yeah, deal with the thermal paste.
A year and a half isn't exactly ancient for a graphics card, but it's a long enough time for thermal paste to degrade. That said, 89 degrees while throttling suggests it's not just that. Guys, I'm on board with cleaning, but I'd suggest doing it systematically and in order.
Start with the simplest thing: power down the card, shut off the power supply, and blow out the heatsink with a can of compressed air. Dust builds up fast, especially if the case isn't closed or the airflow is poor. While you're at it, check the card's fans - they should spin freely without any grinding. If there's a lot of dust and throttling keeps happening after cleaning, then we think about thermal paste.
The second thing people often miss is air circulation inside the case. If the graphics card is wedged between other components, lying on its side, or there's just no decent airflow, even a clean heatsink won't help. Check if there's space around the card, whether the case fans are working, and if the system isn't clogged with cables. Often just reseating the card in a different slot or upgrading the case cooling does the trick faster than messing with thermal paste 😊
A temperature of 89 degrees on its own isn't critical for the 4090 - the card handles higher temps - but if it's throttling at the same time, that means the thermal protection kicked in, and the culprit is poor heat dissipation. Yeah, dust and dried-up thermal paste are the usual suspects, but don't rush to take the card apart: first, blow it out with a can of compressed air right there in the case, check if air circulates freely around the card, and make sure no cables are blocking the air intake.
If nothing changes after cleaning - that's when you think about replacing the thermal paste, but that's a more involved operation that requires disassembly. Try the simple stuff first, especially if the card was running fine for a year and a half and suddenly started acting up.
Based on your description, it's most likely a cooling issue, but I wouldn't rush into taking the card apart. First, check if there's enough airflow in your case - if other components are also running hotter than before, the problem is in circulation, not the GPU itself. While you're at it, look at your monitoring software and see if the card's fan speeds are jumping around erratically, because sometimes they get stuck and won't ramp up properly even at high temperatures.
If airflow in the case is fine and the fans are spinning smoothly, then yeah - blow out the heatsink with a can of compressed air. It'll take about five minutes and often fixes the problem completely.
Don't rush to take the card apart if your case is stuffy - first check if air is flowing properly through the whole system. Sometimes a GPU gets hot not because it's clogged, but because hot air just isn't leaving the case and keeps circulating. Make sure your front fans are pulling air in, the rear ones are exhausting, and there's at least a couple of centimeters of clearance between the case and the wall.
Are the fans on the card spinning normally themselves or making some weird noise? :) To put it shortly, yeah, 89 degrees while throttling almost always means clogging, but don't forget about power - sometimes the PSU starts sagging on voltage over time, and the card doesn't get stable current, which is why temps climb. Before you take the card apart, just make sure the power cables are seated properly, the contacts are clean, and the PSU isn't making any weird sounds.
I had something similar with my graphics card - I cleaned the radiator, but it didn't help much, turned out the fans just wore out and were spinning slower than they should 😅
check the fan speeds in monitoring (in MSI Afterburner or similar) - if they're spinning at 50-60% even under load, then the coolers can't handle it. try manually raising the fan curve in the settings and see if the temperature drops. if it does - then it's definitely a cooling issue, not dried-out thermal paste. if not, then yeah, you need to replace the paste or even reseat the radiator
Thermal paste can definitely dry out over a year and a half, especially if the card was under load, but before you replace it, first check the simplest thing - whether air can circulate normally in your case around the graphics card. Sometimes a card overheats not because the heatsink is clogged, but because hot air just stagnates inside the case and the fans have nothing to cool. If your case is full of cables or sitting in a closed cabinet, that could be the culprit.
Log into answer.
