I have an AMD XP 3000+ and it runs (min load) arround 48.5 deg C

but when it is punding out some of my engineering simulatiosn i have hit 75 deg C, the only reason that i haven't gone above 75 is because i have a safe guard in plance where if the temp gets jumps to 75.1 deg C the computer shuts down and restarts

If a BIOS update doesn't give revised temperatures then it indicates a problem with the system's cooling.

I've got an XP3200+, and Seti runs 24/7- according to SANDRA the CPU temp is 50°c and that's with an ambient temperature of 32°c 62% RH. I've got the case off of the computer as that drops the temperature another 5° or so (otherwise the CPU would now be around 55°) which comes in usefull when the ambient gets around the 37° mark.

---

Grant
Darwin NT

Yesterday evening to test the temps on my PentM, I set up my 4 channel temperature meter, this has serial link o/p, and connected this to my P3 to record the graphs at 5 min intervals, the simple software will only do 100 samples.

The supplied heatsink has a silicon buffer that surrounds the cpu to prevent uneven pressure loads, PentM's do not have heatspreader, as fitted to P4's etc. So put one sensor in this space, one on heatsink directly above cpu, one in case and last one monitored the ambient temp. I also checked the reported cpu temp using Everest, the mobo supplied one is unstable.

At all times from about 10 mins after this started, with BOINC running, the internal case temp was 1C above ambient, the heatsink was 12C above ambient, and the space close to cpu was 18C above ambient. And from the infrequent readings I took the reported CPU temp was 30.5C plus/minus 1C above ambient.

Even by doing the stupid thing of holding heatsink on CPU, rather than screwing down, running computer/BOINC etc, then removing power from psu and quickly measureing the temperature on top of cpu the temperature was 42C, about 10C less than that reported by Everest.

So, unless you can physically monitor voltage from internal cpu diode sensor, that would entail knowing pins, attaching wires etc, ny conclusion is it is almost impossible to measure the cpu temperature and verify the reported temps unless they are reported lower than physically measured.

Fot those wishing to monitor sensor diode voltage, it is 0.65V at 20C and decreases 0.02V/degree C.

---

Last night I was going to check the temp of my Toshiba laptop. However, the only tmp reading showed by SpeedFan is the HHD temp which is or was at the time 26C. The program could not detect the CPU temp.

This laptop runs 24/7 running S@H and P@H.

EDIT: This laptop has a Intel Celeron M. The fan only runs about 60 to 65 % of the time the unit is on.

The Dell has always had a heating problem and it sometimes shuts down for the any reason. Even if it's not a heat problem it will shutdown even when a program crashes! I don't ever see any kind of a BSOD. It just locksup and then shuts down. When I retart it I sometimes have to restore the Active Desktop. If anyone out there has any ideas I'd be glad to hear them. I have given up on DELL they are a pain in the ___________. (Fill in with your choice of word or words) :)

Both laptops are checked weekly for spyware and virus update and OS updates.

Ricky

---

On my AMD 64 3000+ OCed to 2ghz I run in the low to mid 30 degrees C. Thats while running SETI 24/7. Although this summer it hit about 39 or 40 once in awhile, but now that its cold its in the low 30's most of the time. I have 2 120 mm case fans, and the stock heatsink and fan on the cpu. Powersupply has 2 fans as well.

---

Join the team, SETI.USA We are growing and could use your help to overcome SETI.Germany...www.setiusa.net

Hi. From experience I have found that AMD processors work fine until their temp gets to about 75-80C.

I have an Athlon64 3000+ (not overclocked) running Seti 24/7 and the max temp I have seen this get to is 56C

---

The Dell has always had a heating problem and it sometimes shuts down for the any reason. Even if it's not a heat problem it will shutdown even when a program crashes! I don't ever see any kind of a BSOD. It just locksup and then shuts down. When I retart it I sometimes have to restore the Active Desktop.

See if there's anything in the system or application logs.
Put the laptop on a hard surface (eg table top) with a desktop fan blowing on it & see if it becomes stable- if so then it's definately a heating problem & the heatsink/vents may be clogged. When running hard, can you hear the CPU fan?
If it's not over heating then it's most likely a power supply problem- either the external "brick" unit or it's own internal regulators.

---

Grant
Darwin NT

The Dell has always had a heating problem and it sometimes shuts down for the any reason. Even if it's not a heat problem it will shutdown even when a program crashes! I don't ever see any kind of a BSOD. It just locksup and then shuts down. When I retart it I sometimes have to restore the Active Desktop.

See if there's anything in the system or application logs.
Put the laptop on a hard surface (eg table top) with a desktop fan blowing on it & see if it becomes stable- if so then it's definately a heating problem & the heatsink/vents may be clogged. When running hard, can you hear the CPU fan?
If it's not over heating then it's most likely a power supply problem- either the external "brick" unit or it's own internal regulators.

The fan runs almost all the time based on the project running at the time. I run C@H, E@H and S@H and when running the C@H the only time the fan stops is when C@H is writing to the hard drive. During E@H it runs the about 80% of the time and with S@H about 70%. The laptop is always placed on a hard table top. I try using one of those "Chill mats" but the fans turn the wrong way. The fan on the laptop pulls the air in and the "chill mat" wants to pull the air down from the laptop so the two fight each other. I never understood why dell but the fan on the bottom. My Toshiba's is on the right side of the laptop.

Ricky

---

ricky. Have you tried putting it up on something, i put my laptop on dvd cases. 2 high on 2 corners so that air flows under. It improved my times by 10-20% just by doing that.

---

Join the team, SETI.USA We are growing and could use your help to overcome SETI.Germany...www.setiusa.net

.... During E@H it runs the about 80% of the time and with S@H about 70%.

Odd.
My CPU runs at 100% load all the time, so the temperature is always high & stable. I would expect the same thing with the laptop.

The laptop is always placed on a hard table top.

As Joe Joe suggested, have you tried placing it on some small spacers- give it extra air flow.
Tried the fan trick?

---

Grant
Darwin NT

.... During E@H it runs the about 80% of the time and with S@H about 70%.

Odd.
My CPU runs at 100% load all the time, so the temperature is always high & stable. I would expect the same thing with the laptop.

The laptop is always placed on a hard table top.

As Joe Joe suggested, have you tried placing it on some small spacers- give it extra air flow.
Tried the fan trick?

Right now I only have the back of the unit up about an inch based on something I read in another thread about overheating. As long as the room temp does not go above 75 the Dell is ok. I may have to set it so it only runs the projects in screen saver mode instead of Run Always.

EDIT: Yes I did try the desktop fan trick but didn't work. I am thinking of making a stand for it with a 80mm fan to pull air in under the unit and cover it with a A/C filter to keep out dust.

Ricky

---

Hi. From experience I have found that AMD processors work fine until their temp gets to about 75-80C.

I have an Athlon64 3000+ (not overclocked) running Seti 24/7 and the max temp I have seen this get to is 56C

Most processors are rated to around 100C. Traditionally, they have a lifetime of 10y, which is where the manufacturers once decided they could accept the economic risk based on the reliability bath tub curve. More recently the acceptable lifetimes have been pulled in (at least internally to companies like Intel) to 7 y and to 5 y. Again they are accepting more risk by pushing the operating conditions. (If the customer dumps his laptop after 4y why warrant anything beyond 5y, is the argument.) In the same vein, they have convoluted 'market segment use condition' with lifetime, which on the surface is very objective, but proves a risk for gonzo's like the people in here who are using their laptops as desktops. For example, the Pentium M was meant for a laptop which has a totally different use condition (always on) than a desktop (frequently off, little number crunching), so it's performance and reliability 'curves' were adjusted accordingly.

With that long winded preamble, when I read the comment above regarding AMDs' pooping out at 75-80C, it makes me wonder what AMD has fiddled with on the reliability side to get the performance. (They don't have a good track reliability track record, at least a few years ago.) These temperatures are on the low side, especially since lifetime failure mechanisms are thermally activated (exponentially turned on for those folks in Rio Linda). Could we be looking at a rash of early wearouts in two years, in part because we gonzos have them going full blast all the time? (Well some of you do; I've stuck with Intel for now.) Something to ponder.

---

May this Farce be with You

For example, the Pentium M was meant for a laptop which has a totally different use condition (always on) than a desktop (frequently off, little number crunching), so it's performance and reliability 'curves' were adjusted accordingly.

The above statement looks backwards to me. Shouldn't it be:

For example, the Pentium M was meant for a laptop which has a totally different use condition (frequently off, little number crunching), than a desktop (always on) so it's performance and reliability 'curves' were adjusted accordingly.

Bur then I could be wrong if so i'm sorry.

Ricky

---

Most processors are rated to around 100C. Traditionally, they have a lifetime of 10y, which is where the manufacturers once decided they could accept the economic risk based on the reliability bath tub curve. More recently the acceptable lifetimes have been pulled in (at least internally to companies like Intel) to 7 y and to 5 y. Again they are accepting more risk by pushing the operating conditions. (If the customer dumps his laptop after 4y why warrant anything beyond 5y, is the argument.) In the same vein, they have convoluted 'market segment use condition' with lifetime, which on the surface is very objective, but proves a risk for gonzo's like the people in here who are using their laptops as desktops. For example, the Pentium M was meant for a laptop which has a totally different use condition (always on) than a desktop (frequently off, little number crunching), so it's performance and reliability 'curves' were adjusted accordingly.

With that long winded preamble, when I read the comment above regarding AMDs' pooping out at 75-80C, it makes me wonder what AMD has fiddled with on the reliability side to get the performance. (They don't have a good track reliability track record, at least a few years ago.) These temperatures are on the low side, especially since lifetime failure mechanisms are thermally activated (exponentially turned on for those folks in Rio Linda). Could we be looking at a rash of early wearouts in two years, in part because we gonzos have them going full blast all the time? (Well some of you do; I've stuck with Intel for now.) Something to ponder.

I think you answered your question regarding what AMD has fiddled with on the reliability side to get the performance yourself. If Intel now make their CPU's with an estimated life of approx 5 years, AMD probably do the same. 10 years ago, the processor of the day was a Pentium running at 133-200Mhz. Today this system would be usless at running the latest apps/games and would not even be able to run some of the BOINC projects. Even a 5 year old processor would struggle.

A system with the correct heatsink/fan fitted should never get anywhere near 65C, let alone the 75-80c when things start going wrong.

---

With that long winded preamble, when I read the comment above regarding AMDs' pooping out at 75-80C, it makes me wonder what AMD has fiddled with on the reliability side to get the performance. (They don't have a good track reliability track record, at least a few years ago.)

It's been a long time since the K6, and even then there wasn't a reliabilty problem with those CPUs, just a stability one. And most of those problems were chipset related, not CPU related.

---

Grant
Darwin NT

Most processors are rated to around 100C.

My quick scan of CPU specs shows that most CPUs are rated at around 70° case temperature, it is only the mobile ones that have the 100° case temperature rating.

---

Grant
Darwin NT

@Ricky: Yes you are right; I typed it wrong.

My point is that using these chips outside of the market segment use conditions should cause premature failures (if the manufacturer's reliability folks are earning their money and not overspec'ing things). If someone is actually running an AMD (see below) and is getting soft or otherwise errors, this may be an early warning.

Case temps are one thing, but the chip's junction temps are spec'd, at least internally. The junction temp (what the heat sink is trying to cool) is probably 100C. Electromigration and related rel specs are based on this junction tempeature, too.

Oh, yeah, blaming the chip set instead of the cpu manufacturer is hardly honest. The system fails when it fails.

---

May this Farce be with You

My point is that using these chips outside of the market segment use conditions should cause premature failures (if the manufacturer's reliability folks are earning their money and not overspec'ing things).

I think that both Intel and AMD have been making processors long enougth now to know what they are doing.

If someone is actually running an AMD (see below) and is getting soft or otherwise errors, this may be an early warning.

This would be the case for any make of processor.

Case temps are one thing, but the chip's junction temps are spec'd, at least internally. The junction temp (what the heat sink is trying to cool) is probably 100C. Electromigration and related rel specs are based on this junction tempeature, too.

Any processor would fail at much above 75C. Even some of the motherboard monitoring software stars giving an alarm at 70C and then shut down the system at 75C by defult.

Oh, yeah, blaming the chip set instead of the cpu manufacturer is hardly honest. The system fails when it fails.

What about the Ali chipset on some cheap socket 7 motherboards, that was total rubbish.

---

...
Have recently upgraded to the latest BIOS release, this has made no difference to the temp. reading. If anything the temp.actually went up 1-2 'C and I wasn't dully concerned about that!
I am using Gigabite's own EasyTune 4 for 'on the fly' tuning and monitoring

Is it possible (of course anythings possible) that EasyTune4 is missreading the temp. values?

A I reported earlier, my Athlon64 3500+ o/c, runs in the low 30's. This 'puter is running 24/7, doing Seti/Boinc awell as shoot'em'ups, only problems that occur is when both Seti and Games are running at the same time.

---

Hi. From experience I have found that AMD processors work fine until their temp gets to about 75-80C.

I have an Athlon64 3000+ (not overclocked) running Seti 24/7 and the max temp I have seen this get to is 56C

Most processors are rated to around 100C. Traditionally, they have a lifetime of 10y, which is where the manufacturers once decided they could accept the economic risk based on the reliability bath tub curve. More recently the acceptable lifetimes have been pulled in (at least internally to companies like Intel) to 7 y and to 5 y. Again they are accepting more risk by pushing the operating conditions. (If the customer dumps his laptop after 4y why warrant anything beyond 5y, is the argument.) In the same vein, they have convoluted 'market segment use condition' with lifetime, which on the surface is very objective, but proves a risk for gonzo's like the people in here who are using their laptops as desktops. For example, the Pentium M was meant for a laptop which has a totally different use condition (always on) than a desktop (frequently off, little number crunching), so it's performance and reliability 'curves' were adjusted accordingly.

With that long winded preamble, when I read the comment above regarding AMDs' pooping out at 75-80C, it makes me wonder what AMD has fiddled with on the reliability side to get the performance. (They don't have a good track reliability track record, at least a few years ago.) These temperatures are on the low side, especially since lifetime failure mechanisms are thermally activated (exponentially turned on for those folks in Rio Linda). Could we be looking at a rash of early wearouts in two years, in part because we gonzos have them going full blast all the time? (Well some of you do; I've stuck with Intel for now.) Something to ponder.

Never had a problem with any of my AMD processors. And that is starting with the 486DX100, a K62 300, a K63 400, Athlon Tbird 1.2, Athlon XP 2400, Athlon XP 2600, Athlon XP 2800, Athlon XP 3200, Athlon 64 3400, Athlon 64 3800 (Laptop), and my Athlon 64 X2 4400.

The only ones I still have are what are in my sig. The rest I have sold or donated to churches and schools ect. But i have never had one fail.

---

AMD Athlon 64 X2 4400+
AMD Athlon 64 3800+
AMD AthlonXP 3200+

If you want to know the details of your AMD CPU including max temps then see this chart at Tom's Hardware.

---