• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

GPU over heating on some WUs

Joined
Apr 12, 2010
Messages
571
So my GPU is overheating but only on some WUs,whats weird is that most of the time its fine,I have my fan speed set at 80% and it will run around about 65C,I noticed that once a WU finishes and it grabs another one that as soon as it starts processing the new WU my card will sometimes start whining like crazy,so I check the temps and it was at 105C so I shut it down.

Is this normal?Its only on some WUs cause some times it will grab a new one and all will be fine,my card has coil whine normally but I know to check the temps when it starts getting really loud really fast and setting the fan to 100% doesnt even phase the temps,they just keep getting higher and higher.


does any one know if this is normal or could something be wrong with my card?
 
So my GPU is overheating but only on some WUs,whats weird is that most of the time its fine,I have my fan speed set at 80% and it will run around about 65C,I noticed that once a WU finishes and it grabs another one that as soon as it starts processing the new WU my card will sometimes start whining like crazy,so I check the temps and it was at 105C so I shut it down.

Is this normal?Its only on some WUs cause some times it will grab a new one and all will be fine,my card has coil whine normally but I know to check the temps when it starts getting really loud really fast and setting the fan to 100% doesnt even phase the temps,they just keep getting higher and higher.


does any one know if this is normal or could something be wrong with my card?

no its not normal so a couple of things for you to check and do.

1st - it may be a bit louder but just have the fan set to 100% all the time, its one less thing to worry about when your card is running hot - been there, done that, got bored

2nd - pull the card from the case and clean out the heatsinks - you will be amazed at how much crap will be on/in it a good clean and blow through will work wonders.

3rd - check your case temps, clean all the case fans and then see how it runs - if its still hot add extra fans and make sure current ones are working and not blowing hot air back onto the card.

4th - some WU do run hotter - the 450 pointers are the hottest and there is one other, its either the 587 or 787 pointers - my rig is off at the mo so i can't check.

Try them and see how you get on - any more problems we are here to help
 
So my GPU is overheating but only on some WUs,whats weird is that most of the time its fine,I have my fan speed set at 80% and it will run around about 65C,I noticed that once a WU finishes and it grabs another one that as soon as it starts processing the new WU my card will sometimes start whining like crazy,so I check the temps and it was at 105C so I shut it down.
There are certain series of WUs that utilize the GPU more heavily than others. Consequently, they induce higher temperatures on the cards running them and the systems they're installed.

Is this normal?Its only on some WUs cause some times it will grab a new one and all will be fine,my card has coil whine normally but I know to check the temps when it starts getting really loud really fast and setting the fan to 100% doesnt even phase the temps,they just keep getting higher and higher.
Are your cards single or dual-slot? The single-slot cards have inferior cooling and exacerbate the higher temps of certain WUs. You may benefit from reapplying thermal paste and or upgrading the coolers, but you cannot completely avoid all the increases in temperature these WUs create. I have 120mm fans blowing on the sides of all my cards.

does any one know if this is normal or could something be wrong with my card?
If it's only on some WUs, it might be normal depending which WUs in question and if you have single-slot or dual-slot cards. Let us know the GPU WU projects that you believe are causing the temp problems.
 
ok,ill take the card out and open it,I have read that the heatsink and shroud are one piece,so in order to clean it out I will have to remove the heatsink,unless any one knows how to remove the shroud from it,its a GTX 260,also,how much TIM should I put on it,I have some noctua NT-H1,I have never taken apart a GPU before,any pointers?
 
ok,ill take the card out and open it,I have read that the heatsink and shroud are one piece,so in order to clean it out I will have to remove the heatsink,unless any one knows how to remove the shroud from it,its a GTX 260,also,how much TIM should I put on it,I have some noctua NT-H1,I have never taken apart a GPU before,any pointers?

The plastic shroud should either be clipped or screwed on along the top and bottom of the card, if its clipped be very careful as the clips are very prone to breaking - a couple of very fine screwdrivers and lots of patience is needed. If its screwed the screws are very fine and cheap and will damage easily - either way it can be removed just relax and take your time.
 
how to I check to see what WU its on?
Look in your log file, FAHlog.txt and you will see every WU your client worked on. If you cannot remember the WU project numbers during these episodes, this won't come of any use. If you can't find the WU type then post again whenever another unusual WU is worked on by your card.
 
I actually paused it on this last one,I think this is it Project: 6600 (Run 10, Clone 950, Gen 220)

That is the last one that I was working on when it hit 105C
 
the overheating is on the gtx 260?

i would check your fan speeds.. if you have evga precision or afterburner.. see if the fans actually ramping up like its suppose to.. especially if your using the 258.96 drivers a few people have been reporting that it screws the fan speeds up on the gtx 260's this might be whats happening to you.. if so just manually set the fan to 60% or 70% and you should be good to go..

and yes the p6600's will make your card run hotter then any of the other WU's.. but shouldnt be making it hit 105C.. my gtx 260 never went over 75C at 60% fan and a 600/1512/1106 overclock..
 
ok,this is just weird,I cleaned out the card with some air,it was pretty dusty,idle temps dropped from 59C to 39C,fired up the GPU client at 100% fan speed and the temps jumped up again,like to 75C with the same loud coil whine that was happening before but stabilized there at 75C

So I entered the configure page then to the advanced tab and adjusted the slider higher than where it was(was around the 4th mark from the end) and started the WU again and now the temps dropped to 57C?!?

Wouldn't upping the CPU usage slider percent increase the activity on the GPU therefore the heat it produces?It went from 85% GPU usage to 98% and the temps dropped?That does not seem to make sense to me.also the coil whine is very faint now.
 
I have both of my 260's overclocked/fans at 65% and even on the 450 point WU's they don't really go above 80. They were running much hotter then that until I took the air compressor to them. Dropped a good 10 degrees off of the temperatures.
 
ok,this is just weird,I cleaned out the card with some air,it was pretty dusty,idle temps dropped from 59C to 39C,fired up the GPU client at 100% fan speed and the temps jumped up again,like to 75C with the same loud coil whine that was happening before but stabilized there at 75C

So I entered the configure page then to the advanced tab and adjusted the slider higher than where it was(was around the 4th mark from the end) and started the WU again and now the temps dropped to 57C?!?

Wouldn't upping the CPU usage slider percent increase the activity on the GPU therefore the heat it produces?It went from 85% GPU usage to 98% and the temps dropped?That does not seem to make sense to me.also the coil whine is very faint now.


because your limiting the cpu usage it may be requiring more gpu power.. theres honestly no reason to limit the cpu usage on a nvidia card though.. it only uses between 1-4% of the cpu and usually sits at 1% most of the time..

also 100% fan = useless.. you dont need to go any higher then 80% last 20% doesnt do shiot except make the fan louder..
 
Last edited:
Im sure your fans-heatsink need's cleaned or you have poor case airflow. i run duel 260 in one box 702/1544 60% fans and only run low 70's at most.
 
I cleaned the fans and heatsink,my case has ok air flow.

I have been checking it and the only time this happens is on the project 6600 WUs,I dont know why,but as soon as they start the temps just climb and climb,they dont stop,all other WUs are fine and my temps are inline with what most of the others are getting.I dont know what to do,if I pause the WU and restart it,or make any change in the config menu the activity stops then once it restarts the temps normalize to around 65C making me think that the fan might not be spinning up,but that cant be the case cause I have it manually set to 80%.
 
if you have a couple bucks buy a hvac thermometer. $10 see if its really overheating or if your thermal sensor is going bad.
 
I have been checking it and the only time this happens is on the project 6600 WUs,I dont know why,but as soon as they start the temps just climb and climb,they dont stop,all other WUs are fine and my temps are inline with what most of the others are getting.I dont know what to do,if I pause the WU and restart it,or make any change in the config menu the activity stops then once it restarts the temps normalize to around 65C making me think that the fan might not be spinning up,but that cant be the case cause I have it manually set to 80%.
It's very possible that your fan isn't operating at optimal speeds. You can try increasing to 100% and if that doesn't make a difference then it's likely you have a failing fan. It happened to me at least twice. Once that occurs your only recourse is a new HSF unless you can find a replacement fan. Regarding the 6600s, they aren't the hottest WUs, that's why I had wondered which ones were giving you problems.
 
sorry to resurrect a old thread,but I have completely stopped GPU folding,which sucks cause I feel like I could be doing more in this race with Evga,but this damn card is getting way too hot (only sometimes when folding) and I have no idea what to do,when I game the temps are fine,80%ish gpu usage and only around 65C,cleaning the card really helped with temps while gaming

Ive cleaned out the card,put a fan aimed right at it,left the case door open and nothing seems to help the temps,the fans seems to be ramping fine,If I adjust the fan speed I can hear it speed up and down,I have it manually set.

Could there be some sort of driver conflict,software conflict or something,cause the card temps are fine while gaming and its fine while folding for the most part,but I am just not comfortable leaving the card to fold while I am not home and since I work 2 jobs,I am not home alot,eat,sleep,work,I dont really have time to worry about a gpu melting down(or setting my house on fire lol)

Also is the GPU supposed to buzz while it is folding,not a coil whine type of noise but a steady buzz It has different pitches depending on the WU.I am still new to folding,so this may be completely normal,I just want to make sure.
 
Sounds like you need FAH GPU Tracker V2

Havent tried this part of it, but looks like you can set speeds per project to deal with it!

original.jpg
 
sorry to resurrect a old thread,but I have completely stopped GPU folding,which sucks cause I feel like I could be doing more in this race with Evga,but this damn card is getting way too hot (only sometimes when folding) and I have no idea what to do,when I game the temps are fine,80%ish gpu usage and only around 65C,cleaning the card really helped with temps while gaming
What are the exact temps the card is attaining when folding?

Ive cleaned out the card,put a fan aimed right at it,left the case door open and nothing seems to help the temps,
I have a card with some cooling issues. I even removed the HSF and reapplied thermal paste but had no effect. Then I removed the secondary HSF on my EVGA 8800GT Akimbo card and installed it on the back of this problematic card, so it has the regular HSF and this passive low profile sink with a heat pipe on its backside..Unfortunately, it only lowered the temps by about 3°C. It's better than nothing and I just left the card as is. This card is in an open case and has a 120mm fan blowing on it from the side as well, but it is a single-slot card. I would see much better temps if this card had a dual-slot cooler.

Regarding the noise you're hearing, it could be a number of things.
 
The highest it has gotten is 105C and that is only because I heard the card screaming for its life and jumped up off the couch to check what was going on,generally when the card is going to get very hot I can tell because the card just starts screaming as soon as it starts the WU so I come check afterburner and just watch the temps rise and rise hoping they will stabilize,they dont stop so there is not exact temp cause I usually just stop it around 99C.

I'll take a look into FAH gpu tracker.
 
OK, you have some serious issues with your card. That is much higher than the card I was mentioning, which I thought was running a little on the warm side, yet my card is under 90°C. Your card probably even throttles, which will reduce folding production so there's little point running it.
 
That is not what I wanted to hear,but is kind of what I was expecting,I have another GTX 260 that should be here tomorrow,ill see how that one does.

It is most likely throttling as you have said,when running this card as well as the smp client I was only getting around 9kppd running it and a Phenom II 955,I was under the impression I should be getting atleast 6.5 to 7k points out of the card alone,cant remember how many points the 955 should average but right now its around 5.5k

Thanks for all the help,is there anything I could do,like check specific parts on the card maybe or do you think I should just give up on it?
 
That is not what I wanted to hear,but is kind of what I was expecting,I have another GTX 260 that should be here tomorrow,ill see how that one does.
If this card is going to replace the one you have now, I am very curious what the difference will be, especially if they're similar cards.

It is most likely throttling as you have said,when running this card as well as the smp client I was only getting around 9kppd running it and a Phenom II 955,I was under the impression I should be getting atleast 6.5 to 7k points out of the card alone,cant remember how many points the 955 should average but right now its around 5.5k
I do not have a GT200-series card, but I've seen people recently posting that their GTX260s get close to around 9000 PPD on their cards alone with the best WUs, not including whatever other clients may be running on their machines. So, yeah, you are definitely losing production. There's no doubt in my mind your card was throttling once it hit ~100°C.

Thanks for all the help,is there anything I could do,like check specific parts on the card maybe or do you think I should just give up on it?
Well, aside from replacing the HSF for an aftermarket unit like something from Thermalright, you've done pretty much all that could be done except for one thing. Removing the HSF, cleaning the GPU and applying the best thermal paste you can find. Then reinstall the HSF making sure it is contacting well against the card. Other than that, get the card replaced.
 
That is not what I wanted to hear,but is kind of what I was expecting,I have another GTX 260 that should be here tomorrow,ill see how that one does.

It is most likely throttling as you have said,when running this card as well as the smp client I was only getting around 9kppd running it and a Phenom II 955,I was under the impression I should be getting atleast 6.5 to 7k points out of the card alone,cant remember how many points the 955 should average but right now its around 5.5k

Thanks for all the help,is there anything I could do,like check specific parts on the card maybe or do you think I should just give up on it?


depending on your overclock the gtx 260 should do between 6.5-8k PPD.. my guess is you are running the gpu3 client and getting periodic gpu3 non fermi WU's which kill PPD on the g200's.. as far as the temp problems its probably happening when you get the p660x gpu2 WU's which will make the card run much warmer but not 105C..

you mentioned possible drivers problem, what drivers are you running? i know the 258.96 drivers where causing issues for some gtx 2xx owners and was causing the fan to completely stop then start back up and stop again in a cycle.. only way around that was manually setting the fan.. when i was running my gtx 260 i was running it with the fan at 60% which is pretty quiet.. 70% i could here it and 80% or higher was just plain annoying.. at 60% fan and running 602/1512/1161 my temps sat at about 68-75C depending on the WU..

though since you did clean out the card.. my best guess is that if none of these idea's above help your temps then the next best thing would be to remove the heatsink and replace the stock thermal paste on the gpu.. XFX is pretty good about people removing the heatsinks and using aftermarket stuff so i doubt changing the thermal paste will cause you to void the warranty if thats something you are worried about..
 
The card I am getting is an EVGA gtx 260 core 216 55nm,it is identical to my current card,only the brand is different.bought it off a fellow [H] member.Wow I was losing alot of ppd then,wish I would have known that earlier.

My current card was bought used as well,so who knows what may have been done to it before I purchased it.I will have to contact XFX to see if they can do anything for me,not holding my breath though cause the card seems to run fine while gaming,though the newest game that I have been playing is mass effect and I don't know if that game is very demanding on the card.

I wonder how well the card would do under a more demanding load?would MSI kombuster push the card enough to see if the temps are stable?
 
What I would do is SLI the two cards since the old card is still good for gaming, and just run the folding client on the new card providing there aren't any issues with it. Stress test it by folding. You'll soon see if the temps will be OK. It shouldn't take more than 5 minutes or so.
 
@ sir monkey,

I am pretty sure that I am using the GPU2 client and I was on the 197 drivers while this was happening,I have since moved to the 258 drivers but I have the fan speed set manually,I was hoping it was a driver issue,but the problem still persists

How can I check to see what version of the GPU client I am using?
 
@ sir monkey,

I am pretty sure that I am using the GPU2 client and I was on the 197 drivers while this was happening,I have since moved to the 258 drivers but I have the fan speed set manually,I was hoping it was a driver issue,but the problem still persists

How can I check to see what version of the GPU client I am using?

in the log file(if your using HFM right click the client then click view cache log) at the top.. if its the gpu3 client it will say:
Folding@Home Client Version 6.30r2

i believe the gpu2 client says 6.23 but i havent used the gpu2 client in a while so i cant remember for sure.. basic easy way to check.. (again if your using HFM) go to tools/benchmark viewer scroll the bottom and see if theres a WU called p10632 thats a gpu3 non fermi WU.. if you dont see the WU then you know for sure you are using the gpu2 client..

if i were you id just take the heatsink off and remount it.. cheapest and easiest way to solve the problem.. because if you try to RMA it for temp issues they will test it and find it plays games/passes tests fine and just send you the card back or you get a wonderful refurb sent back instead..
 
if i were you id just take the heatsink off and remount it.. cheapest and easiest way to solve the problem

I agree with this. There's no way a GTX 260 should hit the temps you're seeing when folding unless something is physically wrong with the card. Check the heatsink and fan and make sure everything is installed correctly, reapply thermal paste, etc.

The only other consideration is the ambient temperature. Is the room temp above 80F?
 
well,I just got the new card installed and boy what a difference it has made I can definitely tell that there was something seriously wrong with my xfx card.

while running the XFX card it would constantly buzz or whine I thought it was normal for the F@H client to do that to a card,this EVGA card is dead silent,I can not even hear the fans running at 70% speed and the temps are only 65C and it is a OC edition card and I was not expecting that.Im going to give the shaders a bump as soon as the current WU finishes to try and make up for some of the points I have lost over the last few weeks.

how much does OCing the shaders improve performance?will it make a noticeable difference?

also put in a ticket to XFX to see what can be done about the other card.
 
well,I just got the new card installed and boy what a difference it has made I can definitely tell that there was something seriously wrong with my xfx card.

while running the XFX card it would constantly buzz or whine I thought it was normal for the F@H client to do that to a card,this EVGA card is dead silent,I can not even hear the fans running at 70% speed and the temps are only 65C and it is a OC edition card and I was not expecting that.Im going to give the shaders a bump as soon as the current WU finishes to try and make up for some of the points I have lost over the last few weeks.

how much does OCing the shaders improve performance?will it make a noticeable difference?

also put in a ticket to XFX to see what can be done about the other card.


you can do the shader bump while its on the WU but its up to you.. card should do 1512 shaders no problem if its the OC edition.. but if your worried start off at 1458 and then work your way up from there.. i think the next shader step after that is 1482 then 1512..

if the xfx cards coil whining id just rma it instead of screwing with it but its up to you..
 
I am just going to try and RMA the XFX card,hopefully they can do something for me.cause it would be freaking awesome to get both of these cards folding in my rig.

I have the card right now running at 626/1584/1053 running fine and dandy so far,man I could not even dream of upping the shaders this far before and I have not even touched the core or memory clocks,the last card was a freaking turd when it came to any adjustments and the temps are 64C!!!

I forgot to say thanks to every one that offered their advice to me thus far,I really appreciate it there is not doubt in my mind that [H] is the #1 team on the planet regardless of what rank we hold!!
 
Awesome, what's the PPD like with the new card so far?
 
I am just going to try and RMA the XFX card,hopefully they can do something for me.cause it would be freaking awesome to get both of these cards folding in my rig.

I have the card right now running at 626/1584/1053 running fine and dandy so far,man I could not even dream of upping the shaders this far before and I have not even touched the core or memory clocks,the last card was a freaking turd when it came to any adjustments and the temps are 64C!!!

I forgot to say thanks to every one that offered their advice to me thus far,I really appreciate it there is not doubt in my mind that [H] is the #1 team on the planet regardless of what rank we hold!!


just be warned if you do game.. whats stable in F@H no no where near stable in gaming :p but damn i wish the evga gtx 260 i had even remotely overclocked that high.. highest it would go was 630/1350/1161 for gaming and 600/1512/1161 for F@H.. not sure how much higher your card will go but 1584 is about the norm for those evga OC versions in F@H.. as far as your clocks go i would go with 630/1584/1100 though the vram should do 1200 no problem it doesnt make to much of a difference PPD wise maybe 10-15 PPD.. same with the gpu at best you will gain 100ppd trying to hit 700 mhz while adding 10C to the temps... but have fun experimenting with it and see what you like..
 
Awesome, what's the PPD like with the new card so far?

8053 ppd is where HFM.NET puts the card at,running the shaders at 1620 now,12,900 ppd total,cant wait to get an H70 to throw on on the phenom II and really push it.

having a card that works correctly makes me want to purchase another,I really hope XFX can help me out with the other,I have never had a card or heard of a card a card that was so unstable while being stable at the same time lol
 
Back
Top