X2 3800 running at 113C under full load. :|

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
I just recently installed a new upgrade into my computer. An AMD Athlon 64 X2 3800+ was the CPU I chose. I've been fooling around and trying stuff out and it performs fantastically, but I have a bit of an issue. I'm part of a group for distributed.net, which is similar to the Folding@Home situation. CPUs run at 100% whenever they can to help achieve something. I was planning on overclocking my CPU to improve the number of keys I could crack in distributed, knowing that the 3800+ is well-known for overclockability, but decided to check my CPU temperature to see how much room I had. Sisoft SANDRA 2007 reports my CPU temperature to be at 113 degrees Celsius (254 Fahrenheit) when being under 100% load after a few hours. I saw this and I almost flipped! That seems incredibly high, and my comp should have blew up by now. I stopped distributed from running and after a few minutes checked again. My CPU temp was down to 65 idle. Right now it's at 50C idle.

This seems like it could really damage my computer. Last time I heard, your CPU shouldn't be over 50C period to ensure you don't shorten its lifespan. I'd appreciate some input here. I'd like to know how much damage I've already done and if I should consider getting extra cooling and maybe underclocking a bit.
 

repoman0

[H]ard|Gawd
Joined
Apr 7, 2006
Messages
1,216
Up to 70C is fine, but 113 is ridiculous. Either your monitoring software is reading it wrong (use Core Temp) or you need to remount the cooler..
 

daytripper67mi

Limp Gawd
Joined
Feb 22, 2004
Messages
364
check your mount,idle should be much lower 30ish anyway..my 3800 with a/c 64 freezer would get 55c under load so i went to watercooling,also check your voltage setting in bios.
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
I figured it was rediculous. My old Athlon XP 2600+ ran in the 30s idle, 40s under full load.

Just as a side note, my CPU case has no case fans (I know, I need to get some) and I'm using the stock heatsink and CPU fan. Just a quick question, and I'm not trying to seem n00bish here, but by "cooler" you do mean the heatsink right?
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
Okay, I went into my BIOS and got this info from the PC Health area:

CPU Tcontrol: 65C
System Temperature: 61C
CPU Fan Speed: 0 RPM
System Fan Speed: 3020 RPM
CPU Vcore: 1.23V
VDIMM: 1.87V

The fan speeds reminded me of an issue I faced while assembling the parts. The CPU fan's power connector has 3 pin holes. The CPU fan power connector on the motherboard has 2 pins. I asked my more knowledgable friend over MSN about this and he said to look for an alternative for plugging the power connector into. I asked him if the "AUX power" connector on the motherboard, which had 3 pins like the CPU's, would be okay, and he said it would. So I took his word for it and plugged the CPU fan into the AUX power connector on the motherboard.

Therefore, the CPU fan is running at 3000 RPM right now. This seems low because when I was in my brother's computer's BIOS I remember his fan running at 4700 RPM. Maybe this is the reason my CPU is overheating?
 

daytripper67mi

Limp Gawd
Joined
Feb 22, 2004
Messages
364
on my 3500+ cpu fan speed is 3125rpm so the stock cooler is running about the right speed id guess,id check the hsf mount if it was me...and buy a couple case fans
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
The mount is secure... I made sure it snapped onto the heatsink securely. The two are inseparable, so I don't see how I could improve that. I guess my only option is to buy a case fan or two, like you suggested. That or purchase a better quality heatsink. I get the impression that Zelman coolers are popular here on the board. :p
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,906
It sounds like your monitoring software is reading temps incorrectly...unless...you did use thermal paste, right?
 

trick_m0nkey

Moderator
Staff member
Joined
Oct 11, 2005
Messages
4,228
Then that would be why. You need to apply thermal paste on that proc quick.
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
Found the problem:

Upon reading the manual, I noticed that it mentions thermal paste being already applied. I unmounted the heatsink, and checked for it. There was this thick solid plastic cover still attacked to the bottom of the heatsink! Apparently, I went a little too quickly while assembling the parts and didn't notice the plastic covering. This was the first time I've put together a new CPU and motherbaord in my computer myself (the alst time I did, I was 13 and had my dad assemble that), so I can use that as an exuse I guess :p. Either way, I was an idiot and in the future I'll go slower and check for plastic coverings on CPUs prior to putting them onto a motherboard.

I removed the plastic, it wasn't melted in any way (thank god), and now Sisoft SANDRA reports the idle temp to be 29 C down from 67 C. As for peak temp, 10 mins of full load has the CPU running at a cool 38 C down from the 113 C earlier. I have plenty of overclocking room now. ^_^ Though I do plan on buying a few fans and a high quality thermal paste in the future just for good measure.

p.s. To the guy who asked about cut-off temperatures in X2's, I noticed in my BIOS that the default cut out temperatures were disabled by default. To all people owning EliteGroup Sys (ECS) motherboards, I advise you check your BIOS's concerning the cut-off temps.
 

rmed64

Limp Gawd
Joined
Mar 22, 2005
Messages
318
LOL holy crap. I know exactly what you mean by those plastic coverings over the thermal pads. Wow dude, you are lucky that CPU is still working. 113c? WOW :eek:
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,906
Hehe, I called it!

If those temps were real, I figured he either wasn't using any thermal paste, or that plastic film was still on. :p
 

Elios

Supreme [H]ardness
Joined
Aug 12, 2004
Messages
7,282
113f or 113c BIG diffrence

113 degrees Fahrenheit = 45 degrees Celsius
113 degrees Celsius = 235.4 degrees Fahrenheit
 

finalgt

Supreme [H]ardness
Joined
Aug 3, 2002
Messages
5,506
How the hell did that proc survive? Not only were you not cooling it, you were practically insulating the heat. That's pretty impressive CPU engineering, TBH.
 

Unabomber

Supreme [H]ardness
Joined
Jan 1, 2005
Messages
5,244
Good to hear that your solution only takes a few calories (food calories, not heat!). :)

Your CPU should still be OK, but I'd put it through some thorough testing, if I were you.

On another note, you have my respect for having the guts to confess to such a thing. Many folks would have tried to make it sound as if it weren't their own faults.
 

SippieCup

Gawd
Joined
Feb 11, 2007
Messages
762
funny story to tell to yer kids... i bet you have a pretty decent overclocking cpu as well. if ti can handle 113* then you might be able to clock it up quite a bit
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
lol ya it was kind of embarrassing admitting that. There's no way I mention this to my friends or I'll probably lose what little geek cred I have :p

To the guy who mentioned it, it was 113 C, not F. I'm Canadian, so I pretty much always use Celsius eh.

As for the board not shutting the comp down at such high temperatures, that thought occurred to me too after I fixed it. But like I posted earlier, the BIOS has the shut down thresholds disabled by default (NOT safe, bad engineering on ECS's part).

I'm surprised the CPU withstood all that abuse. I had it running at 100% for 6 hours straight once. AMD continues to show quality engineering with this chip as an example!
 

Citizen86

Supreme [H]ardness
Joined
Feb 20, 2007
Messages
4,572
Well I think newer CPU's usually have an automatic shutdown temperature to ensure that they shut down before they actually BURN up and start a fire ;)

but haha good story, I enjoyed it. Gotta be careful EVERYTIME you put a computer together, but live and learn, eh? I just put together my 3rd computer a couple months ago. Got it all put together in like half an hour, booted up first try, haven't had a problem since.

I'm surprised that the plastic hadn't melted all over your CPU though, weird. Either it's really tough plastic or you had the heatsink mounted wrong also, and it wasn't making ANY contact :eek:
But that might have actually been better for it, haha.

Edit: First thing you should do when installing/reinstalling a heatsink, and ESPECIALLY a new CPU is check the temperatures, or else you may run it for a couple days at 100C+ ;)
 

Matt Welke

Limp Gawd
Joined
Mar 17, 2004
Messages
354
The plastic was a thick cover (not some peel off thing) at least 1mm thick, so that explains why it didn't melt.

This was my first time putting together a new motherboard and CPU. Hopefully the second time will be a little smoother and problem-free. Like you said, we live and learn. :p
 

Unknown-One

[H]F Junkie
Joined
Mar 5, 2005
Messages
8,906
This gets me wondering...just how slow (and more importantly, how little voltage) would you need to run an Athlon 64 to run it without a heat sink?

I'm betting you could do 800MHz at 0.6v on a good chip (single core Sempron so there's less cache) and not need a heatsink...who wants to test this out? :D
 

(cf)Eclipse

Freelance Overclocker
Joined
Feb 18, 2003
Messages
30,027
This gets me wondering...just how slow (and more importantly, how little voltage) would you need to run an Athlon 64 to run it without a heat sink?

I'm betting you could do 800MHz at 0.6v on a good chip (single core Sempron so there's less cache) and not need a heatsink...who wants to test this out? :D

i can run my fx-55 at 800mhz, .95v without a heatsink. load temps get up to around 60-65c with just a fan blowing over the motherboard
 

Fendrix

2[H]4U
Joined
Jul 9, 2006
Messages
2,263
I'm Canadian, so I pretty much always use Celsius eh.

LOL the "eh" killed me.

I was actually close to doing what you did with the plastic covering... I could see how it'd be invisible and one might forget to take it off. Thank the Lord it's still okay - maybe you got the F3 stepping haha.
 

Easykill1978

Limp Gawd
Joined
Jan 19, 2006
Messages
479
well its not supprising really that it didnt shut down, after the amd thunderbird error/issue, amd have sinceplaced measures in the cpu to enable wait states when the temp climbs too high. i bet that is what happened, motherboards turn themselves off if things go wrong not th cpu.

also most dektop pcs are derrivatives of server cpus, and i know for a fact that under full load, 1au systems get real toasty before the fan cranks way up.
 

MrWizard6600

Supreme [H]ardness
Joined
Jan 15, 2006
Messages
5,779
LOL the "eh" killed me.

I was actually close to doing what you did with the plastic covering... I could see how it'd be invisible and one might forget to take it off. Thank the Lord it's still okay - maybe you got the F3 stepping haha.

you know, I'm canadian, and I actually use "eh?", but theres more shocking news yet: americans say it too, its just a distant cousin, its called "EAHN". It's really too wierd a sound to accuratly discript with english characters. I heard it on orange county choppers, paul SR said "Yeah why don't you move that rear fender down a bit, yeah that looks good eahnn".

:p

anyways, about the 113C.

you may have just completed what we would call a burn in, at least in part. When your temps get that hot things that arnt supposed to liquify, liquify, and air starts to get out, even on the processor. What your left with is usually a better bond between your processors heat shield and the die, and a better bond between the heat shield and the heat sink. In your case, only the former applies due to you 113C fiasco, but the better bond between the heat shield and the actual processing core will yield better cooling.
 

Citizen86

Supreme [H]ardness
Joined
Feb 20, 2007
Messages
4,572
I'm Canadian, I say "eh", but just usually at the end of a sentence, and usually to use as asking a question. And my American friends pick right up on it, and I don't even notice it anymore.

On Topic:

I've never gotten a processor that hot... Maybe I should try a burn-in? yay. I thought that a burn-in was just running the processor at full for like 24 hours to get it burned in...
 

AltTabbins

Fully [H]
Joined
Jul 29, 2005
Messages
20,277
To the OP..

Wow.. You just topped my darwin award list. Pretty close to The customer who forced ddr2 into a ddr slot and another who forced his s775 proc in backwards..

You earned the spot in my sig. Grats.
 

Citizen86

Supreme [H]ardness
Joined
Feb 20, 2007
Messages
4,572
To the OP..

Wow.. You just topped my darwin award list. Pretty close to The customer who forced ddr2 into a ddr slot and another who forced his s775 proc in backwards..

You earned the spot in my sig. Grats.

Mean! haha. I have a feeling he's not the first one to do that... I think FORCING a stick of ram into a slot that it DOES NOT fit is worse... and forcing a CPU is pretty bad too. NO FORCE PLS
 

(cf)Eclipse

Freelance Overclocker
Joined
Feb 18, 2003
Messages
30,027
and you tried this.... why?
why not? ;)

Maybe I should try a burn-in? yay. I thought that a burn-in was just running the processor at full for like 24 hours to get it burned in...
running it that hot is never good, regardless of the circumstances, unless you're trying to kill it quickly. burn-in is generally a myth anyway.

I think FORCING a stick of ram into a slot that it DOES NOT fit is worse

i've put ram in backwards SO many times, it's amazing. especially ddr2 T_T
 

(cf)Eclipse

Freelance Overclocker
Joined
Feb 18, 2003
Messages
30,027
Well yeah, you put it in, see that it doesn't fit... I think everyone would do that. Or have you broken slots that way :eek:

hm, like this?

DSCF6448.jpg


my desk grew a memory slot!
 
Top