• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Tomshardware's "Heat Emergencies" Article Now Applies to Athlons.

A

AGodspeed

Diamond Member
http://www.anandtech.com/showdoc.html?i=1600&p=9

This is ASUS' KT333 motherboard. It can read the temperature from the CPU's internal thermal diode, and halts the CPU in case of overheating. On the ASUS stand, a video of what happens if you remove the heatsink during operation was shown. This video was very similar to the video of AMD CPUs burning up after heatsink removal that was posted on Tom's Hardware Guide a while ago. But in this case, the CPU survived, and there was no damage done. It remains to be seen whether independant testers will be able to reproduce the positive result of this risky experiment - but if this is the case, then it clearly shows that monitoring the CPU temperature through the internal diode is far superior to the "measurement below CPU" method.
 
Good to hear. I also noticed Asus' new dual AMD board based on the 760MPX chipset has that same feature. I was curious as to how effective it was..looks promising..🙂
 
too bad its one out of many motherboards, 99.999% of which do not support diode or thermal protection...

Until amd forces MB manufacuterers to implement thermal protection like this, or Makes the protection on-CPU, there will be no real solution that is universally appliable to all amd based systems.


Mike
 
Agreed. If I remember correctly, wasn't it a little while before the motherboards properly read the P3 coppermine's diode too? Maybe as time goes on this feature will be properly supported. At least it seems Asus is using it on all their new Athlon boards, so maybe others will follow.
 


<< too bad its one out of many motherboards, 99.999% of which do not support diode or thermal protection...

Until amd forces MB manufacuterers to implement thermal protection like this, or Makes the protection on-CPU, there will be no real solution that is universally appliable to all amd based systems.


Mike >>

Who cares? Who can complain about the Athlon's thermal protection when they can get a board from a top-tier board maker like ASUS for a decent price that will react similarly to a P4 when catastrophes like heatsink failure occur.

It's being implemented by more than just top-tier mainboard makers like ASUS, and so that's progress. Still, if anyone were ever concerned about the miniscule possibility of their heatsink failing or falling off, they'd have a few boards like ASUS's KT333 solution among others to choose from.
 
Personally, I think this whole issue of overheating Athlons is totally overblown. The percentage of HSF that "fall off" is virtually nill. Having the HSF fail is closer to real world occurances and even in those cases the Athlon core has demonstrated impressive temperature ranges.

For example, two weeks ago I was testing a spare CPU in my MP3 box. I got careless and forgot to connect the HSF on the stock HSF and left the computer running at 100% load (Crunching UD) for over 45 minutes with no fan! The exercise resulted in a system halt, but the CPU was completely undamaged and is currently running F@H 24/7. While I wouldn't recommend this to anyone, I am confident that a proper HS installation combined with motherboard shutoff settings will protect most if not all running Athlons.


edit: can't spell...
 
Agodspeed,

I care. It should be ON-CHIP. Look at the excuses certian motherboard manufacturers are using regarding Internal Diode support (Abit is claiming lack of chipset support, Gigabyte is citing patent issues, etc). There's 2 "top-tier" level manufacturers (yes, abit can be argued, but that's beside the point). Then there are the boards that have "unknown" diode support. This will continue, and let me give you one reason why:

Internal diode Athlon temps are high wtih most "non-overclocker" heatsinks. Look at the a7vtroubleshooting page (the only internal diode tests done with reasonable accuracy). CPU temps ranging from 67C to 90C.

Not every person buys Asus motherboards. How many asus boards do you think make it into prebuilt computers?
Dedicating just a little bit of on CPU space to thermal management wouldn't kill AMD's bottom line, and would save everyone the hassle of making sure the motherboard has support for it.

yes, overheating on an AThlon is overblown. I will agree with that. But this is an issue that should be resolved. People still burn their AMD chips with regularity. And, as it is, we'll see people comparing internal diode Athlon Temps versus socket-thermistor temps, which will even further the current confusion. People will be claiming that the diode reading boards are "reading too high", etc.


Mike
 


<< Personally, I think this whole issue of overheating Athlons is totally overblown. The percentage of HSF that "fall off" is virtually nill. Having the HSF fail is closer to real world occurances and even in those cases the Athlon core has demonstrated impressive temperature ranges. >>



The problem isn't the HSF falling off, it's the newbie system builder that doesn't get it put on properly in the first place that is the problem. Even some motherboards that have thermal shutoff don't enable it by default so that first boot can be dangerous.
 


<< Agodspeed,

I care. It should be ON-CHIP. Look at the excuses certian motherboard manufacturers are using regarding Internal Diode support (Abit is claiming lack of chipset support, Gigabyte is citing patent issues, etc). There's 2 "top-tier" level manufacturers (yes, abit can be argued, but that's beside the point). Then there are the boards that have "unknown" diode support. This will continue, and let me give you one reason why:

Internal diode Athlon temps are high wtih most "non-overclocker" heatsinks. Look at the a7vtroubleshooting page (the only internal diode tests done with reasonable accuracy). CPU temps ranging from 67C to 90C.

Not every person buys Asus motherboards. How many asus boards do you think make it into prebuilt computers?
Dedicating just a little bit of on CPU space to thermal management wouldn't kill AMD's bottom line, and would save everyone the hassle of making sure the motherboard has support for it.

yes, overheating on an AThlon is overblown. I will agree with that. But this is an issue that should be resolved. People still burn their AMD chips with regularity. And, as it is, we'll see people comparing internal diode Athlon Temps versus socket-thermistor temps, which will even further the current confusion. People will be claiming that the diode reading boards are "reading too high", etc.


Mike >>

I agree with you Mike, the Athlon XP should have on-die thermal protection not dependent on mainboard support, perhaps a solution similar to the Pentium 4's integrated heat spreader. It shouldn't have to rely on mainboard manufacturers, I agree. However, this may be a moot point if the majority of mainboard makers start supporting the AXP's thermal diode. But we'll assume for now that it won't be a reality anytime soon.

There must be cost reasons or supply/demand reasons why AMD hasn't implemented some type of on-die thermal diode not dependent on mainboard support and implementation, I just don't have a clue as to exactly what the holdup is?
 
Agodspeed,

Looks like it'll have to be on the hammer's, since there's no socket-space for a socket-thermistor 😉. But regarding current support, it kind of sucks that even with a single manufacturer, diode support usually isn't across the board. Like Asus, the KT333 and A7m-266D support it, but the Sis board doesn't. the KT333 MSI does, but its Dual board doesn't? The Abit KT333 doesn't, but its dual does? Wth is going on with these manufacturers 😀.



Mike
 
Wth is going on with these manufacturers 😀.

I haven't a clue. I wish AMD and mainboard makers would address it though...
 
Well, with all due respect, this lovely set of pictures shows what happens when it doesn't work. Granted, the only protection this board had was fan rpm monitoring shut down.

You got a problem with that? Flame the motherboard, it seems to like it
 
Stan, an Athlon T-bird or Palomino won't fry when a fan fails if it's setup properly. We're talking about a total heatsink failure (or dissatachment), which is much worse and much more rare.
 
All things equal it would surely be best to pickup a mainboard with thermal protection if you grab a TBred this year. It's gonna be a hot one in the old town tonight...
 


<< Yeah, I guess it's just coincidence I've seen 3 such failures in the last 2 months.

Uh huh >>

Uh, if you've had 3 Athlons/Durons fry because of fan failure, then you just messed, plain and simple. All Socket A motherboards that I've ever used (or heard of) have an option in the BIOS to deal with fan failure. Did you set up these systems by yourself? The BIOS should have the settings on by default anyway, what did you do?
 
There must be cost reasons or supply/demand reasons why AMD hasn't implemented some type of on-die thermal diode not dependent on mainboard support and implementation, I just don't have a clue as to exactly what the holdup is?

Dunno how much it would cost, but I'd be more than happy with a feature like "thermtrip" (ie the cpu halts by itself when it's about to fry) that's been on Intel chips at least since the PII.

BTW, I wonder if the PIV heatsink mechanical specs are patented. If I was the guy in charge for that at AMD, I'd just adopt the PIV mechanism for future Athlon CPUs.

Alternatively, I wonder why nobody sells 'boxed' Athlons with permanently attached (thermal epoxy...) HSFs. Resellers would love this.
 
Can you read Agodspeed? Nowhere did I insinuate I was responsible for the 3 failures nor was I. I got to deal with the unhappy owners of the failed boards with nice, toasty, dead cpus.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top