[SOLVED] Temps after 9 months of gaming.

[moepkid]

Hey all,

So I have been running this prebuild lenovo pc for the last 9 months, and my temps are a bit higher now than that they used to be (Due to higher room temps of 25 degrees). I heard that 90 degrees is the safe temp for a video card, and I just saw a reading of 91 Can somebody explain to me if this is a cause for worry, or that I can safely game further? This is all on stock air cooled.

These temps are after running FFXIV for about 3 hours, with something for shaders on (GPOSE). I noticed that when enabling these shaders, my GPU will run at 100% and achieve the following temps (in celcius):

RTX 3080
max clock: 1935mhz (avg 1700)
max memory: 1187mhz (avg 1100)
max GPU temp: 76 (avg 71)
max Hot Spot: 91 (avg 85)
max Memory Temperature: 86 (avg 82)
Max fan speed: 78% (avg 73) and 77% (avg 71) (2670 and 2620 RPM)
GPU load: 100% (avg 88)
Power consumprion: 102% (avg 90)
CPU: 73 (avg 61)



I also play bf2042, overwatch and the new borderlands. Temps are a bit lower after the same amount of time (around 85 max), So I am scared now that I see one of 91.
Will it still function after a year? I am planning to replace the PC when its out of warranty (so september next year). I know we cant predict the future, but will these temps destroy my pc? I still have warranty luckily, because I dont know anything about PC's and opening them up.

 

[Karadjgne]

Warm. Even heavy gaming shouldn't take that gpu past 83-85°C or so. But being a pre-built you are at the mercy of not-so-fantastic airflow in the case, it's all about the look, not function.

Better airflow through the case will supply the gpu with more colder air, which is more effective at cooling, and also remove the exhaust heat.

You can try twiddling with the case fan speeds and/or gpu fan speeds using MSI Afterburner and any Lenovo software or bios. See if you can't raise the gpu upto 90% fans at above 80°C and case fans should be near maximum once case temps reach @ 45°C - 50°C as read by motherboard temp.

 

[moepkid]

Warm. Even heavy gaming shouldn't take that gpu past 83-85°C or so. But being a pre-built you are at the mercy of not-so-fantastic airflow in the case, it's all about the look, not function.

Better airflow through the case will supply the gpu with more colder air, which is more effective at cooling, and also remove the exhaust heat.

You can try twiddling with the case fan speeds and/or gpu fan speeds using MSI Afterburner and any Lenovo software or bios. See if you can't raise the gpu upto 90% fans at above 80°C and case fans should be near maximum once case temps reach @ 45°C - 50°C as read by motherboard temp.

Thank you for the info. With GPU you mean all the temps? So including memory and stuff? I saw on the nvidia website that it was listed as 93 degrees, but I dont know which temp they mean. I see other posts stating that 76 is good, and above 90 it isnt but i dont know which temp I should be looking at when they say these ranges. Sorry if im sounding stupid.

 

[ttower2020]

You are fine. Your GPU is actually running at max temps of 76 Degrees. You can ignore the Hot Spot information, as that is from whatever single part of the GPU is the hottest, which includes parts of the GPU that can be run that hot, or hotter. GPU Memory can run up to 110C, which would show on the Hot Spot. The GPU Temp is the thing to look at, and 76 Degrees is perfectly normal. That is the temperature of the main GPU chip itself, which is how GPU temps are measured.

 

[Karadjgne]

Hotspot is on die, the processor itself, not hottest component or they'd just call it memory temp. At 76° it's warm, not hot, gpu won't start throttling until it hits 83°C 😁

 

[KyaraM]

The 91°C limit does not apply to the hotspot, but the general/average GPU temp. It's fine. VRAM is a little at the higher side based my own 3070Ti, maximum here is around 100°C as well iirc, but not dangerous in any way and different coing solutions can account for the difference. Hotspot max is 110°C I was told, so you are still a way off of it being anywhere critical. Nothing is overheating. I wouldn't expect that system to die any time soon because of temperatures. But if you are still concerned, try undervolting. No, you won't break anything.

100% GPU usage only tells you that the GPU is the limiting factor in the system, which isn't surprising if you play in resolutions above FHD and got a high-end CPU.

Btw, I also noticed that FFXIV has some of the higher CPU temperature readings of all the games I play, which is really interesting since I would expect games like Anno 1800 with its higher CPU utilization to run quite a bit hotter, but it's actually a bit cooler... even when taking into in account that FFXIV only really uses 2-3 of my cores. I think it might be connected to online play and Windows optimization...

 

[Karadjgne]

Utilization and usage are not the same. Usage isn't the amount of the cpu used, but the amount of time it spends idle between working. So if you see 40% usage, there's 60% of any given time period that the cpu is doing nothing. That's all the gaps in the code, spaces between lines, time between frame completion and frame transmission, lack of space in a core bandwidth for the next thread etc. All those tiny waiting periods. It's essentially a measure of cpu capacity, not capability or ability.

When you hit 100%, the cpu has zero time for anything new added, so extras get put into a waiting queue to be added next. That kills fps because it takes the cpu far longer to create a frame that has all that added info. Online/multi-player adds a ton of Ai, every object susceptible to multiple manipulations from 2 or more pc's and that's added on top of the users pc, something else for the cpu to deal with. So you'll get high temps on individual cores, but since software reports hottest core activity, it's a false reading, you see cpu at 70°, but really thats only 3 cores, the rest are much cooler.

FFXIV has a huge amount of foreground and object info, but being contained to only 3 or so threads, that can hurt fps. You'll see individual cores at 100% usage, but the coding doesn't allow for spillover, it won't use additional free cores.

 

[moepkid]

You people are amazing. Thanks for taking some of my worries away. Again, thanks

 

[moepkid]

The 91°C limit does not apply to the hotspot, but the general/average GPU temp. It's fine. VRAM is a little at the higher side based my own 3070Ti, maximum here is around 100°C as well iirc, but not dangerous in any way and different coing solutions can account for the difference. Hotspot max is 110°C I was told, so you are still a way off of it being anywhere critical. Nothing is overheating. I wouldn't expect that system to die any time soon because of temperatures. But if you are still concerned, try undervolting. No, you won't break anything.

100% GPU usage only tells you that the GPU is the limiting factor in the system, which isn't surprising if you play in resolutions above FHD and got a high-end CPU.

Btw, I also noticed that FFXIV has some of the higher CPU temperature readings of all the games I play, which is really interesting since I would expect games like Anno 1800 with its higher CPU utilization to run quite a bit hotter, but it's actually a bit cooler... even when taking into in account that FFXIV only really uses 2-3 of my cores. I think it might be connected to online play and Windows optimization...

Utilization and usage are not the same. Usage isn't the amount of the cpu used, but the amount of time it spends idle between working. So if you see 40% usage, there's 60% of any given time period that the cpu is doing nothing. That's all the gaps in the code, spaces between lines, time between frame completion and frame transmission, lack of space in a core bandwidth for the next thread etc. All those tiny waiting periods. It's essentially a measure of cpu capacity, not capability or ability.

When you hit 100%, the cpu has zero time for anything new added, so extras get put into a waiting queue to be added next. That kills fps because it takes the cpu far longer to create a frame that has all that added info. Online/multi-player adds a ton of Ai, every object susceptible to multiple manipulations from 2 or more pc's and that's added on top of the users pc, something else for the cpu to deal with. So you'll get high temps on individual cores, but since software reports hottest core activity, it's a false reading, you see cpu at 70°, but really thats only 3 cores, the rest are much cooler.

FFXIV has a huge amount of foreground and object info, but being contained to only 3 or so threads, that can hurt fps. You'll see individual cores at 100% usage, but the coding doesn't allow for spillover, it won't use additional free cores.

So just a quick update. I noticed that my fps was quite high (165 fps). Turns out, a friend of mine set it to that because I have a 165hz gsync screen. I set it to 82 fps and did not notice anything. However, my temps are waaaaaaaaay lower and fans are also really quiet.

GPU: 64
Hot Spot: 75
Memory: 76
CPU: 61
Power: 51%
Fans: 31% max

Sorry for the stupid question, but did I break anything perhaps by using this card to the fullest for the past 9 months? I feel kind of really dumb to not notice this sooner, and now I am quite worried that I degraded the lifespan of my system.
Also, can an fps limit to half of the refresh break something? Or is it better? Somehow my temps were way lower in BF5 or bf2042 because I think i did not hit 165Hz constantly. In FFXIV however, the counter was at 165 all of the time. Also in overwatch, but maybe not because I did not really check.

Again, im sorry if I sound stupid. I worked so much for this PC and i am afraid I will damage it by not knowing so much.

 

[Karadjgne]

Didn't break anything. Not even close. If there was any potential for damage due to overheat situations, the gpu would first cut its boost, and failing that, shutdown completely. Since none of that happened, you are good.

 

[moepkid]

Didn't break anything. Not even close. If there was any potential for damage due to overheat situations, the gpu would first cut its boost, and failing that, shutdown completely. Since none of that happened, you are good.

Aah good to hear. Im really paranoid with moving stuff like fans, So i thought 75% is way too much to handle. I will try and set it off my mind and just enjoy the games thank you so much for your help, i Really appreciate it.

 

[KyaraM]

You also can't break anything by limiting FPS, btw. So if it makes you feel less worried, you can do that without issues.

 

[moepkid]

You also can't break anything by limiting FPS, btw. So if it makes you feel less worried, you can do that without issues.

Thank you so much for answering this!

Sorry to bother, but do you also perhaps know how come that my GPU power usage is above 100%? I saw it hit 102 once in bf5 and FFXIV when the fps was so high (max not constantly).

I read that it is possible overclocking, but I (and the friend who helped me pick it) didn't do anything related to dat that. Does a GPU overclock itself a bit when the temperature allows it? So that the MHz numbers can go a bit high? I use a function on Lenovo software that says performance mode when gaming.

Sorry if this sounds tupid, I'm really curious with all these numbers.

 

[KyaraM]

Thank you so much for answering this!

Sorry to bother, but do you also perhaps know how come that my GPU power usage is above 100%? I saw it hit 102 once in bf5 and FFXIV when the fps was so high (max not constantly).

I read that it is possible overclocking, but I (and the friend who helped me pick it) didn't do anything related to dat that. Does a GPU overclock itself a bit when the temperature allows it? So that the MHz numbers can go a bit high? I use a function on Lenovo software that says performance mode when gaming.

Sorry if this sounds tupid, I'm really curious with all these numbers.

Iirc that is possible depending on the model. My own GPU is hard-capped to 100% maximum, can't even adjust it up manually, only down. What I can tell you is that the GPU would indeed clock as high as it can within its parameters unrelated to what the vendor, eg MSI, guarantees you it can deliver. For example, without manual OC, my 3070Ti continuously clocked at 1950MHz despite Gainward only guaranteeing that it can boost to 1770MHz. That's almost 200MHz more without me doing anything.

 

[moepkid]

Iirc that is possible depending on the model. My own GPU is hard-capped to 100% maximum, can't even adjust it up manually, only down. What I can tell you is that the GPU would indeed clock as high as it can within its parameters unrelated to what the vendor, eg MSI, guarantees you it can deliver. For example, without manual OC, my 3070Ti continuously clocked at 1950MHz despite Gainward only guaranteeing that it can boost to 1770MHz. That's almost 200MHz more without me doing anything.

I just found some other posts on other sites with people having the same thing, so it seems normal I realy should worry less and game more hahaha

 

[moepkid]

You are fine. Your GPU is actually running at max temps of 76 Degrees. You can ignore the Hot Spot information, as that is from whatever single part of the GPU is the hottest, which includes parts of the GPU that can be run that hot, or hotter. GPU Memory can run up to 110C, which would show on the Hot Spot. The GPU Temp is the thing to look at, and 76 Degrees is perfectly normal. That is the temperature of the main GPU chip itself, which is how GPU temps are measured.

Didn't break anything. Not even close. If there was any potential for damage due to overheat situations, the gpu would first cut its boost, and failing that, shutdown completely. Since none of that happened, you are good.

Iirc that is possible depending on the model. My own GPU is hard-capped to 100% maximum, can't even adjust it up manually, only down. What I can tell you is that the GPU would indeed clock as high as it can within its parameters unrelated to what the vendor, eg MSI, guarantees you it can deliver. For example, without manual OC, my 3070Ti continuously clocked at 1950MHz despite Gainward only guaranteeing that it can boost to 1770MHz. That's almost 200MHz more without me doing anything.

I want to thank you all for having such detailed answers. Its the first day in these 9 months that I actually dont worry while gaming all to your advice and knowledge. Thanks again!