Your gaming PC ’s graphic lineup , otherwise hump as a GPU , is one of its most important components , and it ’s authoritative to periodically condition its temperatures to guarantee longevity . Temperature is the primary factor affecting its disintegration , and you need it to last a foresightful prison term — particularly in today ’s securities industry , where even thebest art cardsaren’t cheap .
Fortunately , keeping an eye on its temperatures is n’t unmanageable and can be accomplished in just a few consequence with helpful tools . But there are a few other thing to keep in mind , so let ’s toil in .
How to monitor your GPU’s temperature
By far , the easiest instrument to check your GPU ’s temperature in Windows 10 can be found by burn up Windows Task Manager and jumping to thePerformancetab . At the bottom of the lean , you canfind out what GPU you have in your system , where it is listed with its temperature between brackets . habituate this to contain your GPU temperature under a heavy gaming freight , ensuring it is operating at or very close to 100 % utilization .
But this is n’t the best tool , as it ’s difficult to supervise in - game . The moment you Alt + Tab out of a game , the GPU temperature deteriorate immediately , so you wo n’t be able to view a unrecorded reading like this unless you have a second monitor lizard , making it unmanageable to see what your GPU ’s temperature is like under load .
However , there is another tool that you may prefer that ’s built into the Nvidia GPU drivers , and AMD ’s GPUs have a similar utility . To get to Nvidia ’s overlay , verify you have GeForce Experience install and are sign in . UseAlt + Zto open up the overlay in - game , click theCogicon on the right hand , and congeal up the sophisticated public presentation monitor in the recess of your choosing . It will calculate like the above windowpane , and you may spy the GPU temperature on the quaternary bloodline from the bottom .
AMD ’s GPUs also have an in - game overlay for monitoring hardware , and there are also alternate third - party monitoring tools such asHWiNFO , MSI Afterburner , andEvga ’s Precision X1 .
The real test is synthetic
All that being said , the aforementioned tools and tests generally do n’t shorten it completely , as different games may top to different temperatures . Or , if your processor is n’t up to sniff , some games may not even tax your GPU at 100 % , which may give a false sentience of security . Instead , the real run for whether a GPU ’s cooling result is adequate is to runFurmark , a unrelenting synthetic torture test build to convey out the worst of your GPU .
The test is prosperous to scat — just instal Furmark , open it , and operate the test at standard configurations . You should see a screen like picture above . Let this guide for about 10 to 15 minutes until the temperature has stabilize and you have a good estimation of where you ’re at . While running the bench mark , Furmark also tells you the GPU ’s clock speeds , temperature , and more .
Do keep in mind that Furmark is harder on the GPU than your average game , so it will bring out the worst and lead in blue clocks than games generally attain . Therefore , it is also a speculative test to run , and it does monish you about this before set off a run , but as long as you keep an center on the temperature and see to it they do n’t exceed 90 degrees Celsius , there is nothing to revere .
Today’s GPUs target: 85 degrees Celsius
The immense legal age of today ’s computer graphic board have a target temperature of about 82 to 85 degree Celsius , have in mind they will either overclock until they reach that temperature or have turn over their maximum overclock . The rationality for this is simple : If there is thermic headroom , the GPU will seek to run at a higher frequency so as to boost performance and/or cut fan speeds to subdued thing down if the power limit has been reached .
But , this can make it a bit unmanageable to tell whether your GPU ’s temperature is within norm . Of course , if the temperature you blob is below 80 degrees Anders Celsius , you loosely have no reason to worry — all is good — but if it is at incisively 85 degrees , it could be fine , but it could also be a star sign of a job .
This is because just as a GPU can overclock when it has n’t reached the thermal verge yet , it can also underclock itself as aegis from overheating when it need to . Therefore , if your GPU temperature is at or very closemouthed to 85 grade , you will want to double - check whether its clock speeds match the design specifications .
Say you have a GPU with a base clock of 1605 MHz and a boost clock of 1770 megacycle , which it happily boost to 1890 MHz inHorizon Zero Dawn , and in the synthetic Furmark examination , it runs at a perfect , on - point 1605 MHz , all at about 72 degree . This is a healthy GPU .
However , if its temperature were to bound to 85 point , andits accomplish clock amphetamine under load were to drip far below the factory base clock of 1605 MHz , then there would be a problem .
One more thing to keep in mind is that although many prick do n’t display it , some may display a “ junction temperature ” that comes out eminent than the GPU temporary read you see elsewhere . This economic value measure the hottest temperature on your GPU rather than the GPU ’s edge temperature , and while the higher physical body may await shivery , it ’s not one you generally postulate to worry about unless you’reoverclocking your GPUwell beyond factory specification .
So, what if the temperatures aren’t right?
If you find that your GPU endure at or above 85 point and its clock speed under load is lower than the advertised base clock , then there is piece of work to be done . The first thing to delay is whether there is enough air flow in your example and whether the debris filters have been cleanse . At the same clock time , agree the GPU ’s heatsink for junk buildup , and transfer it if present . Clean your personal computer , reassemble , and test again . It ’s also worth update orinstalling a new GPU driver , just in case .
What if this does n’t clear your GPU temperature trouble ? There is a high luck of bad contact between your GPU ’s cooler and the GPU ’s gist itself , either from a forged install from the mill or dry out - out paste from old age . Under heavy manipulation , even thebest caloric pastewill begin to dry out out and degrade after about two days , and from then on , it can conduct to significant performance exit from about four years onward .
figure above is a GTX 780 Ti that was disassembled after about four year of leave behind it untouched , and you could clearly tell that the thermal spread is no longer in ordination and need to be replaced . After replacement , not only did the GPU run ice chest , it also ran much quieter , as the GPU was finally able to properly move its hotness into the heatsink for profligacy , which means its rooter and thePC case coolingdidn’t have to work as hard . best yet , this also led to a significant performance addition .
Keep in brain that dismantle a GPU is a specialist procedure and is intimately done by someone with an kinship for disassembling electronics . If you ’re not well-to-do admit electronics apart and your GPU is still within its warranty period , your best stakes is to send the card to the manufacturer for haunt or convey it to a local PC computer hardware shop class .
Understanding your GPU and its temperature
With all of this digested , you should be capable to see that GPUs have an optimal temperature and a maximum temperature . Just like you could overclock a CPU , a GPU can also be overclocked , and many of the same constraints are in station .
If you are plan to overclock your nontextual matter card , be sure to swear on thebest GPU overclocking toolsto keep the mental process unproblematic , and your GPU temperature fully monitored .
