AMD X2 4400 OC'ed temps 40/55 high?

B

BoB-O TiVo

Limp Gawd
Joined
Mar 25, 2004
Messages
147
Hey all,

This is my first OC and it seems to be going well and stable. Right now, I'm at 240FSB (2640 CPU) and 1.5v. That gives me a dead idle temp of 40C and load temp of 55C with average usage around 44C. I'm using an XP-90 and slow speed Nexus fans to keep the box quiet (it also houses a RAID5 array), so I know that I'll have higher temps, but would you worry about running a box 24/7 with these temps?

Thanks,
BoB
 
55 is a little high and I would look into some other types of cooling for it also try and I wouldn't leave this box on 24/7 with those temps that is gonna shorten it's life time pretty severely. put a couple fans at the top as exhaust for hot air and a couple at the bottom to intake the cool air. if your board can take an xp-120 that would probably help
 
umbra said:
55 is a little high and I would look into some other types of cooling for it also try and I wouldn't leave this box on 24/7 with those temps that is gonna shorten it's life time pretty severely. put a couple fans at the top as exhaust for hot air and a couple at the bottom to intake the cool air. if your board can take an xp-120 that would probably help
Click to expand...

I'm wondering why you think those temps will shorten the lifetime of the machine? They're clearly inside the thermal limits of all components involved.

Thanks,
BoB
 
BoB-O TiVo said:
I'm wondering why you think those temps will shorten the lifetime of the machine? They're clearly inside the thermal limits of all components involved.
Click to expand...
for every 10c increase of temps, life is reduced by half. it's not like a sudden threshold where the chip goes from fine to quick death. ;)
 
(cf)Eclipse said:
for every 10c increase of temps, life is reduced by half. it's not like a sudden threshold where the chip goes from fine to quick death. ;)
Click to expand...

AMD lists the max temp at 70C. What is the lifetime at that temp? Also, are you talking about lifetime wrt 24x7 operation at those temps? Like I said, I'm about 40C at idle. I only get into the hight 50's running the Prime95 mega torture test.

Thanks,
BoB
 
Well you asked if we thought the temps were high, and you were given an answer you apparently didn't want to hear. 55 C is the maximum I would ever consider allowing any AMD chip to run. It's not going to burn up and die tomorrow, but like others have said, it's just not good for the chips lifespan. AMD's 70 C rating is the thermal threshold temperature for the chip-- any hotter than this and your chip will die very quickly. You don't want to come any where near this, and 55 is too close for comfort.

If you need better cooling and don't want to turn up the fan on your XP-90 (I am in the same boat) consider going with water.
 
Zinn said:
AMD's 70 C rating is the thermal threshold temperature for the chip
Click to expand...
correct.. for some cpu's.
most clawhammers = 70c, but FX-55 was dropped to 63c
newcastle = 70c
winchester = 65c
venice, san deigo, manchester and toledo have a varying rating.. anywhere from 53c to 69c it seems: http://www.xtremesystems.org/forums/showthread.php?t=71044&highlight=temperature+59c
the mobile a64's are rated to 95c


any hotter than this and your chip will die very quickly
Click to expand...
this is where i have issues. every cpu is different. just as various cpu's overclock better than others, some are more resistant to heat than others. the reason for amd's thermal limit is actually for stability, not cpu life. there are 3 things that can effect stability: mhz, voltage, and temperature... all of which sorta match up to sorta make a curved plane of stability. have voltage and temperature be the X (side to side) and Z (in and out) axis, with mhz being Y axis (up and down)

more voltage at a constant temperature brings the max stable mhz up.
more temperature at a constant voltage brings max stable mhz down.

for this reason, amd defines a rated speed at a given voltage AND temperature. typically you can exceed one or more without breaking the stable threshold, but by messing with them, you can get mhz way up. this is why phase change can get so much higher than air at a given voltage with the same cpu
and.. it's why i simply stated that for every 10c increase, you find that the chip's lifespan will reduce something like 1/2. how long will it last at a given temp, voltage and mhz? there's only one real way to find out.


that said, i'll add a disclaimer. i'm a mechanical engineer, not electrical. those of you who know your shit can probably add a lot more clauses and detail, but this should suffice without my stepping out of my bounds :D
 
Zinn said:
Well you asked if we thought the temps were high, and you were given an answer you apparently didn't want to hear. 55 C is the maximum I would ever consider allowing any AMD chip to run. It's not going to burn up and die tomorrow, but like others have said, it's just not good for the chips lifespan. AMD's 70 C rating is the thermal threshold temperature for the chip-- any hotter than this and your chip will die very quickly. You don't want to come any where near this, and 55 is too close for comfort.

If you need better cooling and don't want to turn up the fan on your XP-90 (I am in the same boat) consider going with water.
Click to expand...

You're right about the fact that it isn't the answer I wanted, but I'm just trying to learn about *why*. Rather than parrot something back when someone asks me the same question later on in life, I'd like to be armed with a little background.

Thanks all!
BoB
 
One other point, we keep talking about 55C, but like I mentioned above, this is only during the Prime95 torture test (large FFTs). Normally, I'm running between 41C and 45C. Are people still leery of these temps?

FWIW, I actually think my problem is case ventilation. I'm using an Antec SLK3700-BQE and there's almost no airflow up at the top of the case. It gets quite hot. When I run without the side panel, temps never go above 50C.

BoB
 
Well, I've got the same CPU up to 2.5GHz on an XP-90C on my Abit AN8 SLI. I'm using BIOS 17 with the recalibrated temperatures. So they're no longer reporting about 10c higher than most other boards.

2.5 GHz @ 1.5v vcore, I'm seeing 47c under dual-prime95 load when room temperature is 72F.

However, when the room temperature gets to 85F, I'm seeing dual-Prime95 load temps of 55c.

It's 24-hour Prime stable at these speeds/temps. So, I'm not going to worry about it.

If yours is stable, and staying under 60c, I wouldn't worry about it either. Yes, you may shorten the life of your CPU. . . but how long do you intend to keep it? CPUs are designed to run for many years. So, even if you chop its life in half, I doubt it'll be around long enough for it to fail on you.

Just my 2 cents.

H
 
BoB-O TiVo said:
You're right about the fact that it isn't the answer I wanted, but I'm just trying to learn about *why*. Rather than parrot something back when someone asks me the same question later on in life, I'd like to be armed with a little background.
Click to expand...
i guess you started your post before i finished mine. read and learn :p
 
Hurin said:
Well, I've got the same CPU up to 2.5GHz on an XP-90C on my Abit AN8 SLI. I'm using BIOS 17 with the recalibrated temperatures. So they're no longer reporting about 10c higher than most other boards.

2.5 GHz @ 1.5v vcore, I'm seeing 47c under dual-prime95 load when room temperature is 72F.

However, when the room temperature gets to 85F, I'm seeing dual-Prime95 load temps of 55c.

It's 24-hour Prime stable at these speeds/temps. So, I'm not going to worry about it.

If yours is stable, and staying under 60c, I wouldn't worry about it either. Yes, you may shorten the life of your CPU. . . but how long do you intend to keep it? CPUs are designed to run for many years. So, even if you chop its life in half, I doubt it'll be around long enough for it to fail on you.

Just my 2 cents.

H
Click to expand...

I have same board and have very similar results 3800+ though. I'm not worried either ;).
 
Hurin said:
Well, I've got the same CPU up to 2.5GHz on an XP-90C on my Abit AN8 SLI. I'm using BIOS 17 with the recalibrated temperatures. So they're no longer reporting about 10c higher than most other boards.

2.5 GHz @ 1.5v vcore, I'm seeing 47c under dual-prime95 load when room temperature is 72F.

However, when the room temperature gets to 85F, I'm seeing dual-Prime95 load temps of 55c.

It's 24-hour Prime stable at these speeds/temps. So, I'm not going to worry about it.

If yours is stable, and staying under 60c, I wouldn't worry about it either. Yes, you may shorten the life of your CPU. . . but how long do you intend to keep it? CPUs are designed to run for many years. So, even if you chop its life in half, I doubt it'll be around long enough for it to fail on you.

Just my 2 cents.

H
Click to expand...

Yes, my house is air conditioned, but my office is upstairs and has other computers running. With that factored in, it's held at about 78F in the heat of the day.

Right now, I'm using an Asus A8N-SLI Premium. I'm running 1.5Vcore, 2.6Vdimm. I'm running 250FSB x 10.5, with a 5:6 mem div 2-3-2-6-1T. My XP-90 is running a Nexus silent fan. My case is an Antec SLK3700-BQE w/ Seasonic S12 430W. The latter is my problem.

There's no exhaust in the top of the case and the bleed-out heat from the PS is swamping the top of the case. That, combining with the heat from the RAM is putting temp pressure on the HSF. I'm going to have to move up to, at least, the slow speed 92mm Panaflo. I'll use some acoustipak to absorb the extra noise. My only other problem is venting the top of the case. I'm tempted to just put a blowhole at the top of the side panel as that would be easiest, but I'm not sure what's the best at this poitn.

BoB
 
Dude, some people trip out on temps too much. If it is below AMD's rating for the chip...and runs 100% stable...I say, fuck it.

The mobo sensors are usually so far off up OR down anyways so you need to take that with a grain of salt, not to mention the OEM rigs, well, they typically have much less cooling than any of us enthusiasts....and I guarantee they run ALOT hotter in say, a Dell, than they do in the average pc guru that builds his own rigs.

I don't give a damn about temps too mucu as long as they are in the proper range for whatever type of cooling you are using. From previous experience I know my small and quiet water cooling system is good enough for my X2. On my Abit AN8-SLI with the 17 BIOS I get a max of 52c load temps under dual instances of Prime, dual Folding only nets me 48c, and a game (single core used) only hits 44c at most. I have a warmer than average room and live in HOT ass Texas...when I went to, say, QuakeCon 2004 and 2005, and it's rather cool air conditioned convention center, I typically saw a 3-5c drop in temps across the board compared to my bedroom.

On the AMD X2 3800+, max temp is 67c according to AMD....so 55c aint jackshit to worry about... :cool:
 
Dude I have the same setup and see the same temps. I have an XP-120 and it will not drop your temps at all. The issue is with the voltage supplied by the Asus A8N-SLI Premium. Anything above 1.4625 will jump your temps from around 32deg idle to 40+ deg idle on that chip. The max stable OC on air you can expect to get with that voltage is about 2.55Ghz. Test it out and see! I have played with this setup for a couple of months so I'm sure yours is going to yield the same results. Also I'm willing to bet your load temp of 55 degs is only running one instance of prime95. If you run 2 instances of prime95 your temps are gonna go up to about 67-70 durning a 24 hour run. The difference from 2.55 to 2.64 is not even noticable unless you are a benchmark freak. I opted for a slower clock and a cooler running chip. The extra .09mhz is not worth 10-12 extra degrees IMO. Under full dual prime load my temps never break 48degs @ 2.55Ghz. Here's a link to another thread about the vcore on the A8N-SLI.

http://www.hardforum.com/showthread.php?t=947019&highlight=premium
 
snoops said:
Dude I have the same setup and see the same temps. I have an XP-120 and it will not drop your temps at all. The issue is with the voltage supplied by the Asus A8N-SLI Premium. Anything above 1.4625 will jump your temps from around 32deg idle to 40+ deg idle on that chip. The max stable OC on air you can expect to get with that voltage is about 2.55Ghz. Test it out and see! I have played with this setup for a couple of months so I'm sure yours is going to yield the same results. Also I'm willing to bet your load temp of 55 degs is only running one instance of prime95. If you run 2 instances of prime95 your temps are gonna go up to about 67-70 durning a 24 hour run. The difference from 2.55 to 2.64 is not even noticable unless you are a benchmark freak. I opted for a slower clock and a cooler running chip. The extra .09mhz is not worth 10-12 extra degrees IMO. Under full dual prime load my temps never break 48degs @ 2.55Ghz. Here's a link to another thread about the vcore on the A8N-SLI.

http://www.hardforum.com/showthread.php?t=947019&highlight=premium
Click to expand...

Yeah, I know about the vcore issue. I am running at 1.5V (despite the BIOS setting of 1.475). Anyhow, I am quoting 55C at two instances of Prime95 torture test (large FFTs), but that's with the case side off. With the case side on, I do hit 60C. I'm going to swing by Fry's today and see if they have the Panaflo fans. If they don't, I'll order one. I'm hoping that double the CFM will bring the temps back in line. I also should re-apply my AS5, since I think I used too much, but the temps don't seem *that* off, so I may not.

BoB
 
You must log in or register to reply here.
Back
Top