P4 Burnout??? Noooo! it cant be!

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Matt,

Yes, in the event of a heatsink removal, however, if you then restarted that machine without a heatsink, it toasts.

recycling the power will toast it
So if you sit there and remove the heatsink, it should shut off. Of course then you know it is off. If it were in your case, and the heatsink had dislodged or come off due to clip failure then you would not know this. 99.95% of PC users would, as their first course of action after a PC shutdown/turned off in the middle of their using it, turn it back on again. In that case, the cpu toasts. If the heatsink dislodged during moving and you turn it on, it will toast.

What I'm saying here is that the solution is real, but it is in no way as complete nor as 'safe' as Intel's. Don't get me wrong, I have a pair of XPs merrily churning through data on my home server/workstation - but I have a fair idea what I'm doing and don't carry the box around all over the place.

I know what you are trying to say, however there is no way under anyone's Sun that you can argue that the motherboard implimentation of thermal protection can be as secure and effective as having the cpu do it itself.

-* <font color=red> !! S O L D !! </font color=red> *-
To the gentleman in the pink Tutu<P ID="edit"><FONT SIZE=-1><EM>Edited by pEtEb on 05/29/02 10:35 AM.</EM></FONT></P>
 
<A HREF="http://www.au-ja.org/artikel-asus-cop-test.phtml" target="_new">http://www.au-ja.org/artikel-asus-cop-test.phtml</A>

In german, so unless your german is up to it, you'll need babelfish.

Their test (okay, one site) shows the Asus implimentation (COP) protects against fan failure and heatsink removal, but fails when the PC is booted without a heatsink - i.e. heatsink removal + restart.

The _assumption_ is that BIOS is required to interface with the processor to shutdown and that unless the PC is already running by the time BIOS loads, the cpu is toast.

Your mileage on this may vary, and if you have anything documented that shows otherwise, or if other manufacturers have implemented more successfully, then I'd be happy to see it.

-* <font color=red> !! S O L D !! </font color=red> *-
To the gentleman in the pink Tutu
 
Asus cop is not amds specced design!

Asus cop is designed by them iirc, amds specced design prevents any frying even on reboot without hsf(why the hell the mobo makers dont just use it is beyond me, maybe cause they are to fin cheap to use the extra chip required(adding 2 bucks to cost etc) instead op for a bios solution, which is flawwed)

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
do you know of any manuf. that does follow spec. and would tolerate starting without a heatsink?

-* <font color=red> !! S O L D !! </font color=red> *-
To the gentleman in the pink Tutu
 
I definately have to agree, that it's a scam and a fake. Looking at the pictures, the whole thing looks fishy.

JLBIGGUY, I also agree. I too have never heard of someone's p4 dying.

xxsk8er101xx, if your CPU dies, it's not always your fault. For example, if your HSF suddenly dies, and so does your CPU, then it's not your fault, now is it? Also, if you bought it, yes it's yours, but that doesn't mean you're completely responsible for everything the CPU (and your hardware) does.

Chipdeath, Matisaro, and Eden: yes, in terms fo convenience, I would rather have my computer crash. But, if it does crash, how do you know the CPU is crying for help? In terms of safety, though, I wuold much rather have the CPU downclock then to simply crash the computer. If it downclocks or throttles, then you are basically guaranteed that temperature will lower and that the CPU will be safe.

Also, Melty does have a point.

Matisaro, how exactly does the AXP protect itself if the HSF dies? Also, I never heard of this feature on AXP, even when looking around at AMD's site. Is it stated anywhere on AMD's site? Now, i know AXP has some sort of thermal protection, but never have I heard that it can protect itself against a dead HSF.


<P ID="edit"><FONT SIZE=-1><EM>Edited by Dark_Archonis on 05/28/02 09:53 PM.</EM></FONT></P>
 
The axp when used with amd specified thermal protection circutry on the motherboard can shut down the axp even in the event of a hsf removal, they have a video which proves this out on the net.

It came out after thg's video (which used a chip which fujitsu designed which did not follow amds spec on the motherboard which led to the frying).

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
I can't believe you guys didn't consider the fact that in many bios you can specify how much you CPU will throttle down when it overheats. If you tell it to throttle down 10% or lower on an overclocked 1.6a (most likely 600MHz or more) then of coarse its going to fry, and it'll go hard too.
I have a 1.6a with stock HSF @ 2.4 GHz and I get a maximum CPU temp of 53C, I have a 50% throttle down set in my bios if it overheats which will hopefully be enough, but honestly I'm not sure. Now I have a new Coolermaster Copper heat pipe cooler in the mail and I'll probably push for 2.6 GHz+. Guess I'll ramp the over heat throttle to around 75% or higher.
From the pictures it clearly looks to me like no over heat temp shutdown was set, and that the chip was set to hardly throttle in an overheat situation. The thermal pad on the bottom of the stock intel HSF is pretty crappy, and it was clearly used and heated onto the CPU. I wonder where all the other P4 overclockers are to testify on how hot their Northwoods get. OCed chips die easier and faster, why do you think overclocking isn't recomended 🙄

Gosh I'm such a nerd sometimes, but then again arn't we all. :smile:
 
<i>texas_techie says:</i>
you said the thermal protection was on the mobo only
Yes, the thermal monitoring is built-in, the actual control (i.e. shutoff) is on the motherboard. I'm not blaming AMD for that, which it seems people think I am. The motherboard makers should be following AMD's specs for thermal protection, or at least making more than a half-assed attempt. Kudos to Asus for being the first to really make an effort.

Whether or not it's AMD's responsibility to make something that will function on it's own and not have to involve the mobo makers, is another topic.

ZD Net: "The Athlon XP series have an enhanced hardware pre-fetch cache, an on-die thermal control diode and an extended version of the 3DNow!"
This is the only one that says it actively does something, as opposed to just monitoring, but I trust Zdnet as far as they can throw me.

<i>Matisaro says:</i>
OK, I would like some links to the restarting=toast claim, I have never heard that before.
You really think that if you start your machine with no HSF installed, you'll be ok? Grizely1 can tell you otherwise.

<i>Matisaro says:</i>
why the hell the mobo makers dont just use it is beyond me
I agree. But mobo makers are cheap, we saw that when everyone used the 686b instead of AMD's southbridge with the 761.

<i>Matisaro says:</i>
It came out after thg's video
But was made before THG's, proving THG's video even more worthless.

<i>Oni says:</i>
in many bios you can specify how much you CPU will throttle down when it overheats
I don't think I have that option, which motherboard do you have?

<font color=blue>Hi mom!</font color=blue><P ID="edit"><FONT SIZE=-1><EM>Edited by FatBurger on 05/30/02 09:06 AM.</EM></FONT></P>
 
I have the Epox 4BDA2+
Excelent board, but mine has a cmos reseting problem, its going back to newegg soon, I just need to decide if I want another and shoot for DDR400, or if I want the TH7-II RAID and some PC1066. I know the PC1066 would perform better, especially if I could get it to clock PC1200, BUT its expensive and I'd only be able to get 256 megs of RDRAM instead of 512 of DDR.

Gosh I'm such a nerd sometimes, but then again arn't we all. :smile:
 
You really think that if you start your machine with no HSF installed, you'll be ok? Grizely1 can tell you otherwise.

On amotherbiard using amds specified thermal protection circut YES you can start the mobo without a hsf and it will be fine.

Thats what were discussing, not general mobos, amd specified thermally protected motherboards.

:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
Is there any info on exactly how the AXP thermal technology works?

IMHO, I still believe that the P4's protection is better because it's on-chip. You don't have to worry about finding a mobo which supports the tech, you simply get a mobo for p4. For example, if you're selecting a mobo for your AXP which has the AMD specced tech form only a list of several, you might not like any of the mobos, or they may be so rare that it's impossbile to get in your area. I believe that AMD should put everything (thermal protection) on-chip. It might increase die size and cost a bit, but it improves the longetivity and reliability of the AXP, or whichever CPU has it. Plus, you won't be "forced" into getting a mobo from a list of few which actually supports the thermal tech, because (as I said before) you might not want any of those mobos. Imagine you find the perfect mobo for your AXP only to find out it doesn't support thermal protection.
 
IMHO, I still believe that the P4's protection is better because it's on-chip. You don't have to worry about finding a mobo which supports the tech, you simply get a mobo for p4. For example, if you're selecting a mobo for your AXP which has the AMD specced tech form only a list of several, you might not like any of the mobos, or they may be so rare that it's impossbile to get in your area. I believe that AMD should put everything (thermal protection) on-chip.

The socket a specification calls for the thermal protection, any mobo manufacturer not putting it on them motherboard isnt following the spec and they are at fault.


The p4's solution may be better cause it dosent rely on an external chip, but remember, the amd solution contains 2 parts, both are specified, and as for picking a motherboard with thermal protection and lack of choice, you must complain to the manufacturers, because their cheapness and lack of followthrough is the reason you have little choice when it comes to that feature.

Amd had an issue with the tbird, it had a tendancy to die without a hsf, amd did what it felt it could to assist the situation, I am postive an onboard thermal soultion would have caused socket incompatability, and that would have been an unacceptable step for amd, so they did what they could to protect the cpu(without ruining compatability or cost), and it works when properly implimented. AMD did more than enough imo to protect from the extremely rare occurance of total hsf removal.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
This argument may seem a bit sparse or not complete but, when I think about it, Intel has been long in the CPU business before AMD, and only later with the P3 that they began using some kind of thermal shut off techniques, until the P4 in 2000. AMD OTOH had begun this a year later with mobos now using it in 2002, and when you think about it proportionally to the time Intel has been around, IMO AMD has been faster at it, just that at current times Intel felt the need to do it, as MOST of its users are OEM buyers and those who just don't know what they're buying, so they fear any could toy with the inside. The small % of OCing just doesn't matter to Intel. Show me a link where they do state they want to support the OCers, and I'll be happy, but until then it's AMD who does so, and they never had to create a thermal protection since not many would have frying issues.

--
Meow
 
<i>Matisaro says:</i>
That would be throw*

Holy Jebus, how did I do that? :frown:

<i>Dark_Archonis says:</i>
if you're selecting a mobo for your AXP which has the AMD specced tech form only a list of several

There are currently none that use AMD's designed protection.

<i>Matisaro says:</i>
I am postive an onboard thermal soultion would have caused socket incompatability

I hadn't thought of it from that aspect. What part would cause the incompatability, though?

<i>Matisaro says:</i>
AMD did more than enough imo to protect from the extremely rare occurance of total hsf removal.

"more than enough"? If they did less, then it wouldn't work at all, and that would be enough?

<i>Eden says:</i>
Intel has been long in the CPU business before AMD

AMD has been around for what, 20-25 years or so? That seems like long enough for me. They've been making processors of some sort for most of that time, and CPUs specifically for about...10 years, maybe? Mat would know dates more than I would.

<i>Eden says:</i>
Show me a link where they do state they want to support the OCers

There aren't really any links to say that for AMD, either. Both companies say your warranty is void if you overclock, i.e. they don't support overclocking.

Intel does care about the enthusiast though. They had a gallery of modded cases at IDF, which surprised me a fair amount.

<font color=blue>Hi mom!</font color=blue>
 
This argument may seem a bit sparse or not complete but, when I think about it, Intel has been long in the CPU business before AMD, and only later with the P3 that they began using some kind of thermal shut off techniques, until the P4 in 2000. AMD OTOH had begun this a year later with mobos now using it in 2002, and when you think about it proportionally to the time Intel has been around, IMO AMD has been faster at it, just that at current times Intel felt the need to do it, as MOST of its users are OEM buyers and those who just don't know what they're buying, so they fear any could toy with the inside. The small % of OCing just doesn't matter to Intel. Show me a link where they do state they want to support the OCers, and I'll be happy, but until then it's AMD who does so, and they never had to create a thermal protection since not many would have frying issues.
you go girl! I totally agree with you! Intel didn't use "some kind of thermal shut off techniques" on the 8086, 186,286,386,486 and 586 damn it! I don't understand them I mean those chips were so freak'n hot they burnt my finger off!!!

btw, I upgraded my "eden post decipher system©"(SSE2 optimized) to ver3.0 fyi, and it takes your gobbydegoop posts and makes it much more readable.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
 
Yeah I also updated so that whatever you say against AMD, is in fact WITH AMD. Now you like AMD, to me, and you also value them a lot. Thanks Melty for supporting both companies!

--
Meow
 
That is why I said "proportionally" and I also stated my argument might be off a bit, but again from how I see it, Intel has been much longer in the business than AMD and only recently they started the use of thermal prot. As for AMD they have 2 years later added it, which is not enough to seperate the long time Intel has been in making CPUs. This is why I say that AMD isn't late, proportionally they are faster in implementing this.

--
Meow
 
I don't think I have that option, which motherboard do you have?
are you refering to P4's duty cycle? look under [Power Management Setup] default is 62.7% afaik it's not in the manual.

"<b>AMD/VIA!</b>...you are <i>still</i> the weakest link, good bye!"
 
Oh my god. Not the same issue again.

The socket a specification calls for the thermal protection, any mobo manufacturer not putting it on them motherboard isnt following the spec and they are at fault.

Yes, It's not AMD's fault at all. So if you want to upgrade you machine please buy a new MB and new processor which will take advantage of this thermal protection. And of course this will cost you little bit more but hey It's not AMD's fault that they are making you buy the new MB so you can protect your CPU.

All I am trying to say is if I have an old T-bird CPU and I just wanted to get the Palomino I wouldn't have to buy a new MB with the processor. That defeats the argument of AMD has a price/preformance advantage. Not if you add the cost of new MB.

And ya, If AMD really wanted it they could have done it without making a socket switch.

KG.

"640K ought to be enough for anybody." - Bill Gates.
 
Yes, It's not AMD's fault at all. So if you want to upgrade you machine please buy a new MB and new processor which will take advantage of this thermal protection. And of course this will cost you little bit more but hey It's not AMD's fault that they are making you buy the new MB so you can protect your CPU.

If you buy a new amd cpu and expect it to have thermal protection, then yes you need to have a compatable motherboard.

Or would you prefer amd change the socket every 6 months like some companies?



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 
And ya, If AMD really wanted it they could have done it without making a socket switch.

And you know this because your in the industry, or an engineer of some sort?

As for burgers question, the p4s method is complex and I am sure that it would raise comptability issues with existing motherboards, the cpu has to tell the motherboard its throttled dosent it?

If the p4 throttles and dosent inform the motherboard, how does the motherboard know to slow down the data transfer to the cpu, or to slow down vital time dependant functions to avoid data corruption and crashing, no I dont think the p4's thermal protection is completely motherboard dependant, but since intel forced a new socket on the marketplace they can ensure compatability with all motherboards.

A luxury amd does not have, and does not want imo.



:wink: The Cash Left In My Pocket,The BEST Benchmark :wink:
 

TRENDING THREADS