Safest Maximum Temperature for CPU?

[defense]

I was just wondering on "average" what the safest Temperature for a CPU would be? I check my system BIOS' hardware temperature sometimes, and I just noticed that my computer is maxing up to 105 degrees Farenheit or 40 degrees Celsius! That's seems pretty extreme, especially considering my case has SIX FANS, not including the one on the mother board or Video Card for a total of (8) fans.

My motherboard was at about 88 degrees Farenheit or 31 degress Celsius. Can anyone confirm if that is MUCH too hot for a system? The highest I've ever seen my system was about 90 degrees or so, until now that is. If anyone can advise on this, i'd appreciate it.

[WeedVender]

WOAH that many fans ehh?

Well the fans need to be instrategic spots to maximize effiency.
A CPU running at 105 is pretty high...

Mine is running at 65 using a HUGE FAN and some special heat sink...
my other system is a VapoChill.. and the CPU runs at -27 .F
I got no fans on that one..

115 is over kill for a CPU and consider replacing your fans..

[rallynavvie]

I don't know if that's true or not. According to the monitoring software issued by my system board manufacturer my CPUs run average 45ºC and plateau at 55ºC during long MPEG encoding sessions (just did my first multithread encode with TMPGEnc today with both CPUs getting up to 95% load). I have massive heatsinks and fans on each CPU, one on the system chipset, 2 on the 600W PS, 2 intake chassis fans, and one large exhaust chassis fan. Do I count the one on my VPU?

I'm getting pretty good circulation in my case with all rounded IDE cables and minimum of tangle. I guess it's not in a very cool part of my house (some people like having them in a room where you need a jacket to work comfortably) but it gets cool air off the window next to it.

I've got a DigitalDoc5 thermal management center for my box. It has 8 temp sensors I can place anywhere in the case and can control up to 8 fans (as well as monitor 5v and 12v power). I'm planning on using some of the free-floating thermistors to map the airflow and cooling efficiency of my case to see if I have any hot-air vortices being caused by all the fans. I looked into liquid-cooled systems but I think those are overkill in a home system, even an enthusiast system like my own. But if I ever get that hardware encoder I may look into them.

[WeedVender]

same area...
I mean in Farenheit
By the way..

Defence.. how many outlets do you have on your motherboard for fan power. Or do use something to power the fans no using the motherboard.

[zodiac]

i dont know what a good temp for a p4 is,but under 50c is a good thing.
defense 40c is not hot.
for other chips over 50c is warm,over 60c is hot.
my amd 1900+(1.6ghz) is at 55c which is warm,but it can stand more then that.

[defense]

thank you ALL for your replies! I did some research on this as well because I had never seen my CPU that hot, and I found out that some people actually go up to 150 degrees Farenheit and aren't concerned! I was like WHOA! Anyway, Weedvendor..lol....smooth handle by the way :icon6: ....but my motherboard has a fan on it, and it powers itself.

I know a guy that owns a computer store down here and he gives me everything basically half price and sells old equipment that I have for me. This case I have is really HELLIFIED! It goes for OVER $250 and he sold it to me for a little more then $100. It's got those see through windows which are Neon Blue on the sides and the top. You can slide the ENTIRE inside out by unscrewing a few twist screws, and the doors have two twist screw each which make it very simple to take them off.

As far as the fans go, there are two in the front, two in the back, two on the top, and one on my motherboard AND videocard, for a total of 8. Not to mention that my computer is super quiet. I can't even hear it running with all of those fans.

[rhegedus]

I've got a P4 2.26 oc to 2.53.

32 degrees C @ idle
48 degrees C @ 100% load after 1 hour (it gets to 48C after a few minutes but stays rock steady hours after, and cools down in a couple of minutes)

Intel - don't you just luuuuurve them !?!

Reagrds,

Rob

[defense]

I've got a P4 2.26 oc to 2.53.

32 degrees C @ idle
48 degrees C @ 100% load after 1 hour (it gets to 48C after a few minutes but stays rock steady hours after, and cools down in a couple of minutes)

Intel - don't you just luuuuurve them !?!

Reagrds,

Rob

Rob, what is your multiplier locked at and what are all the settings you changed to increase your chip by 3 ghz? Did you just increase the FSB and VCore Voltage? If so, what did you increase the FSB "TO" ..."FROM" and what did you increase the VCORE VOLTAGE "TO" .."FROM." Thanks in advance

[rallynavvie]

I think there's more going on to keep your CPUs that cool rather than just being Intels Just because my CPUs spontaneously catch fire means nothing

[rhegedus]

Rob, what is your multiplier locked at and what are all the settings you changed to increase your chip by 3 ghz? Did you just increase the FSB and VCore Voltage? If so, what did you increase the FSB "TO" ..."FROM" and what did you increase the VCORE VOLTAGE "TO" .."FROM." Thanks in advance

3ghz?????

CPU Host Freq (MHz) 149 (was 133), multiplier = 17
Memory Freq (MHz) 298 (set automatically)
CPU V Core 1.500V (no adjustment - maybe automatic?)




Regards,

Rob

[defense]

Rob, lol..I mean 300 MHZ..I was dead asleep when I responded to that post. Anyway,that's interesting you didn't adjust that VCORE voltage at all, because all the "guru's" which I read posts from on other sites, said they honestly believe Vcore voltage is the KEY to overclocking. I did increase my VCORE voltage from 1.525 to 1.70.

Everything was running so smoothly at 2.94 GH'z which is where it was maxed out, but those damn VOBS had issues. I wish I knew why that was. Thanks again for the info though.

[rhegedus]

From my own experience, I've increased the FSB untill the PC becomes unstable, then up the Vcore by a notch or so to stabilize it and then repeat until either the temp or stability becomes an issue. However, I'm happy with my 10% or so increase in performance from a simple oc from 2.26 to 2.53 so I don't feel the need to risk my PC.

Regards,

Rob

[defense]

From my own experience, I've increased the FSB untill the PC becomes unstable, then up the Vcore by a notch or so to stabilize it and then repeat until either the temp or stability becomes an issue. However, I'm happy with my 10% or so increase in performance from a simple oc from 2.26 to 2.53 so I don't feel the need to risk my PC.

Regards,

Rob

So you didn't change ANY of your memory frequency? Just the Vcore Voltage and the FSB, right? If that is the case, that is EXACTLY what I did. I went up slowly and at from 133 to A WHOPPING 166 FSB. At 166 the system locked up on the desktop, but down ONE NOTCH to 165, I was crusin' at 2.94 GH'z!

Talk about LIGHTNING QUICK! I have this crazy case which has 8 fans so my temperature NEVER increased past 110 degrees Farenheit. I thought everything was great, until as I said, those damn VOBS played with pixelation and screetching. That is the ONLY reason I decreased my FSB back to where I am at 2.54 ghz from 2.4. I want to go back to 2.94 because at 165 FSB I'm maxed out with no known problems EXCEPT those VOBS playing the way they do. I guess I'll just have to join one of these forums for overclocking where I don't need a PERMISSION SLIP..lol.

This is the ONLY problem preventing me from staying at that INSANE PROCESSING SPEED. I wish I knew why those VOBS played like that, but from what I'm getting on this forum, it looks like the hard drive, I just wish someone knew and could explain "why" the HD would do that. Anyway, thanks for all your info, it was definitely helpful.

[rhegedus]

I'm having a few doubts about it being the hard drive - the read/write speed would be the same no matter what the processor was set to. I'm watching some vobs on my HD at the moment and the cpu is running between 10 and 20%, so it's not really 'processor hungry'.

Have you opened the case to see if you can identify the source of the noise?

Regards,

Rob

[defense]

I'm having a few doubts about it being the hard drive - the read/write speed would be the same no matter what the processor was set to. I'm watching some vobs on my HD at the moment and the cpu is running between 10 and 20%, so it's not really 'processor hungry'.

Have you opened the case to see if you can identify the source of the noise?

Regards,

Rob

What's up Rob? Thanks again for the reply...to answer your question, I guess I was inspecific and maybe that's why noone is able to comment on what the noise or pixelation might be. The noise isn't coming from inside the case...it's coming from the "speakers" or more specifically, the sound of the movie when I play the Vobs from my hard drive. To explain in detail, what happened at 2.94 GH'z is or 165 FSB is when I played a movie VOB from my HD the picture had "pixelation." Also, while the music of the movie was playing, there was like a high pitch screetching noise mixed in with the sound of the movie.

I am not sure if "pixelation" is the right term to use, because the ENTIRE picture wasn't pixelated, but there were "little" tiny dots scattered throughout the entire picture. I can say very confidently, whatever the cause of those "tiny dots" were, was the same cause of the screetching, because they went hand in hand. Again, if this is any help, I was COMPLETELY maxed out at 165, and when I went up one more notch to 166, my system locked up on the desktop when Windows Started. I decided to drop back down to 145 FSB and the problem went away completely, but I was still concerned about something else occurring so I dropped back down to about a 140 FSB.

Even at 2.75 GH'z, the VOB problem went away, but I decided to drop down further because I didn't want to take a chance of something else occuring.

Well, with all of this info do you have ANY ideas at all as to what it can be? I'm sure your advice is much more incisive then mine. From what you've said thus far though, you don't believe it is the Video Card, and you don't think it's the Hard Drive...any other ideas...suggestions? Thanks once again in advance.

[rhegedus]

Well, I'm now stumped as to what the cause could be. Since the sound is coming from the speakers (I presume it is not present if the speakers are turned off or disconnected) it would appear to be some sort of interference probably as an artefact to the above normal voltages going through the motherboard. Anyone else have any ideas?

Regards,

Rob

[defense]

it would appear to be some sort of interference probably as an artefact to the above normal voltages going through the motherboard. Anyone else have any ideas?

Now that is VERY, VERY interesting regarding the voltages!

The reason I say this is because I got a "hardware" error shortly after increasing the FSB by that much..but I soon realized the error was due to my VCORE voltage being set too high. I upped it from 1.525-1.725 which is a HUGE increase, however, from what I read..that should of been about right when increasing a chip 500 MHZ from 2.4 to 2.9 ghz as I did.

I did sign up to a hardware forum and ask this very question to those guru's and one of them responded saying that when I increase the FSB other things increase as well such as the "AGP/PCI" for the Video Card and only "CERTAIN" motherboards can lock in a AGP/PCI setting and allow you to increase the FSB.

I have to check if mine is one that can lock it in. I"m very confident that it's one of the "two" things that caused that..if not a little of both. This is not something I want to keep playing with, so if I can get a sure shot answer then I'll decide on whether or not to go back to 2.9 ghz and leave it there for good, or just stay where i'm at.

When you get a chance, let me know what you think about that "AGP/PCI" possibility.

[scottrempel]

Hi, I am not any kind of guru but I have been reading a lot of forums because I am purchasing a new computer soon and was learning about overclocking. The forum is at nforcershq.com and so relates mostly to nforce motherboards for AMD's. However I am sure everything is very similar between the way mobo's work. The problems you are having is from exactly what you are saying with the increased voltage.
Do you use onboard sound or a soundcard?
If a soundcard, then you are having the problem where your PCI can not be locked and you are pushing it beyond its limit, causing sound distortion.
If onboard sound, then it is a similar problem except that its the chip that controls the sound thats getting pushed to hard.
You should be perfectly fine running at the highest stable speed. For example, if you are stable at 160FSB with no errors then that should be fine. There are many programs that can be run to test your comps stability.
I would look into the AGP/PCI thing though as the board I am getting locks them and hence they can not be hurt. If yours does not lock them you could end up frying them which would be bad.
Hope some of this helps! Good luck!

[Bob W]

40 degC is not hot for any cpu. I believe you will find most AMD cpu's run 40 to 50 degC with a stock heatsink/fan and the Intel's may run a little cooler. I would start to worry if the cpu goes above 55 degC.

[SaSi]

It's interesting how this "keep-the-CPU-cool" thread turned into a "how-hot-can-I-go"

Some of my experience in overclocking.

Defense, the sound problems you are having remind me of some problems I had quite a while ago with sound playback on OC system. It turned out to be that the sound card couldn't take the increased PCI frequency. To be more specific, since we can't unlock the CPU multiplier, we have to increase CPU frequency. Most motherboards tie this with the PCI frequency, increasing it by similar steps. Although most PCI cards will happily operate in increased PCI rates, the limit is not defined and sound cards (especially old ones) are most prone to have problems.

Overclocking can't really offer more than 10% more CPU power. Also, it's recommended (www.overclockers.com) that CPU VCore is not increased to more than 10%. Actually, you don't have to increase VCORE by more than 7% to go to 10% higher clock.

Another issue with increasing the CPU clock is that again in most m/b it is again tied with memory clock. If you have DDR 333 (as I do) this means that nominal CPU clock gives 166MHz to the memory. I can increase CPU clock to almost 2.939 from 2.67GHz but that would make RAM operate at 185MHz --> 370MHz. RAM won't take this so I have to drop RAM speed to DDR266 specs so the increased CPU clock won't overstretch RAM.

To that extent, I have found out that CPU and RAM intensive jobs (like encoding) are not accelerated by much. I gain almost 10% by CPU clock increase but loose almost half of it in slower RAM speed, giving me an overall encoding performance increase of 6~7%. Its not really worth it, I think.

Also, OC CPUs get much hotter, less reliable, etc.

To give an answer to the original question, I have seen PIII systems that had heat-sinks poorly installed (not enough silicon paste and not perfect fit of the heatsink on the CPU) run at 100 degrees Celcius (VERY VERY HOT). They could almost take it. They locked up at approx. 105 degrees Celcius. Correcting the CPU heatsink fit solved the problems and the CPUs were not permanently damaged. (this is not a single case, the bad fit was a problem on 12 machines).

CPU temperature can be as high as 70~80 degrees Celcius on a P4 and not cause any harm at all, however a nominally clocked P4 with a decently good heatsink (e.g. the Intel provided one) will never exceed 45 ~ 50 degrees Celcius.

A very interesting video is posted at www.tomshardware.com along an article on how do CPUs handle heat-sink failure. They actually removed heatsinks from CPUs (P4, Athlon and Celeron, I think). The P4 just decreased performance at almost 20% of nominal to cool itself down (cool isn't it?), the Celeron, I think locked up and the Athlon reached a temperature of almost 300 degrees Celcius and then exploded along with the motherboard - that's what I call a "hot-CPU" :P

[flaninacupboard]

there's an interesting video on the VIA website of the same thing, a C3-Erza CPU at 800mhz, they remove the heat sink and fan, and nothng happens, 10 minutes later nothing happens, 24 hours later and quake3 is still running perfectly. nice

[Faustus]

flaninacupboard I'd really love some linkage on that video if ya have it.

[Ducatti20]

I'm not sure about Intel CPU's, but AMD chips are rated from 70C to 110C as a no cross temp. My XP 2200 Thunderbird runs happily along at ~40 to 50C idle and loaded, with a stock AMD heatsink/fan combo. Results from other users vary, most reporting instability at temps over 60C. Rule of thumb seems to be, if it's stable, it's cool enough. The cooler you go, the longer the lifespan. Average lifespan seems to be 3-4 years for AMD chips running at ~50-60C, and over 9 cooler than that. How old is my processor? 3 months, and it won't last a year before I change it out. If you're really concerned, try running SETI @ Home http://setiathome.ssl.berkeley.edu/ or Folding @ Home http://folding.stanford.edu/download.html. Both of these will run your CPU hard as long as you have them open. [/url]

[defense]

@ scottrempel..thank you for that post..it was very informative and is what it seems like more people say is the problem...that the voltage was too high and causing this. I don't have onboard sound..I have a sound card..the Sound Blaster Audigy 2...and the "distortion" sounds like the BEST way to describe the problem. Although, it appears the only reason the sound was distorted, was because the "video" was. I will test out other methods now because i'm curious, but it looks like I can possibly overclock more. Thank you for that infornatiion..it is gonna be very helpful.

@SaSi, thank you very much for that very informative response....quickly..as i mentioned above..I don't have an older sound card..it's the Sound Blaster Audigy 2, and I think your info on the no "more" then 10% increase in VCORE voltage is gonna be VERY HELPFUL. I increased the VCore Voltage by more then that..and I think I shouldn't of gone passed 1.675 which would of been 10% from 1.525. 7% would probably of been the best bet from the info you posted...so I will have to see. Thanks again for the info

[rallynavvie]

I just put a pair of Thermalright SK7 copper heatsinks on my CPUs with big 80mm fans on top of them. Now after an hour of encoding my temps only get to 48ºC (was 55ºC with OEM aluminum heatsinks). I'd say that's a pretty good deal. Now if only I could get my motherboard to cooperate so well

[flaninacupboard]

flaystus
check this page at the bottom, video is called "beat the heat"

http://www.via.com.tw/en/Products/movie.jsp

[musher70]

go here defense, lots of stuff about OC here...
http://www.3dgameman.com/
end up in forums, see em go wild...
p4's up to 4 g's.....

[Raen]

Tom's Hardware has alot of cool things on it. There's even a video on replacing fans on a CPU, motherboard, etc with a water pump. Keeps everything amazingly cool(way beyong anything mere fans could do) and is silent.

It sounds like a really cool idea. You'd use distilled watered so nothing shots out incase of a leak, but if you put it together right nothing will ever leak. I'd try it myself if the smaller kit alone wasn't 300 bucks. If you're a serious overclock fiend a water pump system is the way to go.

No fan noise and the system stays cooler

[defense]

go here defense, lots of stuff about OC here...
http://www.3dgameman.com/
end up in forums, see em go wild...
p4's up to 4 g's.....

What's up Musher..how's it goin'..I haven't seen you on here in a while. Thanks for that link...I'll check it out. Thanks