AMD Smokes...

[Pabster]

mikewarrior2 wrote:

"What are you talking about? The MP was tested in a Fujitsu-Siemens KT266 board that supports internal diode readings... Its one of a handful. Yeah, the Tyan's do, but there's no way to read them in windows."

Here's the problem. Regardless of whether the Fujitsu-Siemens board Tom used is capable of reading the Pally's thermal diode, I'm 100% certain it is incapable of asserting the Pally's internal clock throttle. Anotherwards, the Palomino requires a board capable of telling the CPU to throttle itself down -- the processor itself doesn't do this automagically. Simply being able to read the temp from the internal diode isn't good enough.

 

[LucJoe]

<< think if this....if you live somewhere cold like I do, then your AMD rig can become your portable space heater >>



Haha, sucks for me, living in an attic in summer heat. Only one small window a/c. If I don't turn off my two tbirds when not in use the place turns into an oven

 

[Mikewarrior2]

Pabster,

So this isn't AMD's fault... fault automatically goes to the motherboard?

I've seen P4s on Socket-thermistor ddr setups not fry themselves... this should be a CPU thing, since the fast majority of users are still going to be SOcket-thermistor setups.

THey will have no way of even reading the diode, much less having the mb throttle the cpu.



Mike

 

[Imperium97]

<<
Here's the problem. Regardless of whether the Fujitsu-Siemens board Tom used is capable of reading the Pally's thermal diode, I'm 100% certain it is incapable of asserting the Pally's internal clock throttle. Anotherwards, the Palomino requires a board capable of telling the CPU to throttle itself down -- the processor itself doesn't do this automagically. Simply being able to read the temp from the internal diode isn't good enough. >>



The real problem to me is that AMD didn't build the clockthrottling into the cpu. The p4's is built in so it doesn't matter what MB you use. That seems like a pretty smart way of doing it. Why on earth would you try to pass the job onto each individual motherboard manufacturer? Seems pretty stupid on AMD's part, not the fault of fujitsu-siemens.

 

[Pabster]

Imperium97 wrote:

"The real problem to me is that AMD didn't build the clockthrottling into the cpu. The p4's is built in so it doesn't matter what MB you use. That seems like a pretty smart way of doing it. Why on earth would you try to pass the job onto each individual motherboard manufacturer? Seems pretty stupid on AMD's part, not the fault of fujitsu-siemens."

While I tend to agree, my point is that Tom's test isn't accurate. If the board he used wasn't capable of monitoring the diode in quick intervals, realizing a temperature fault, and then addressing the Pally's clock throttle to kick in... Anotherwards, take the results with a grain of salt.

 

[Kingofcomputer]

not a big deal,
buy retail athlon,
if it gets burned, just do a rma.
why need to care or worry about if it doesn't have burn protection?

 

[CraigRT]

<< stability - and amd chip with a good motherboard and windows 2000 will give you tons of stability. >>



I agree 100%.

my T-Bird machine is 100% rock at 150 FSB and would make one hell of a sweet development machine...

either chip is just as stable as the other...

 

[Imperium97]

Are there any boards out there capable of allowing the AMD chip to throttle back before it burns up? If not, then it's a moot point in that AMD's clock throttling is as good as having no clock throttling at all. If nothing out there will react fast enough to save the chip than the ultimate conclusion that AMD's will go up in smoke without HSF contact still stands. I agree, take things with a grain of salt, but it still seems to be AMD's problem and if they want to make sure temp changes are read fast enough for clockthrottling to work, they should do it themselves.



<< why need to care or worry about if it doesn't have burn protection? >>


I'm sure a lot of people. Why go through the hassle of burning a chip up during installation and then having to get another one? Is it 100% certain AMD will replace your chip? Aside from that, I'd like to see a more realistic test of clockthrottling. Instead of seeing the entire HSF removed, i'd like to see the fan just stop working. A fan failure is more common place. I'm curious to see if and how long an athlon would last without active cooling.

 

[NesuD]

<< But heatsinks need a fan to do their job. The result would be the same it would just be a few minutes more. >>


Not if you use a decent mother board that will shut itself down at say a 70c reading or a temp moniter proggie like mbprobe that moniters fan rpms and will shut down if the rpms go to low. The fact that a fan fails doesn't automatically mean fried cpu. A decent heatsink should keep things in check long enough for the system to shut down. I modified my connector on my delta black label to moniter rpms even while running off the power supply before anyone offered those adapters or fans could be bought that way just for that reason. I expect that the odds of a properly installed heatsink simply falling off are astronomical and the fan failure thing is a non issue if proper care is taken in the configuration. Besides for cost of a P4 setup you could accidenly smoke a couple tbirds and still not be in the hole moneywise.

 

[majewski9]

All you people that constantly mention the heat factor with AMD are very misinformed! The P4 is not exactly cool running, but yes it does run cooler than a Tbird at clock speed. The Palomino core is cooler running than the Tbird core. The Palomino runs cooler than a similarily clocked P4 from what I hear. By the way the coolest running processor if you are paranoid about heat is the Via C3! The transmeta cruesoe also runs at a much cooler temp.

 

[Hugenstein]

Turn your tower case on its side ... that way if the heatsink clip breaks, gravity will hold the heatsink on top of the CPU ... then again your cupholders don't work to well in this configuration.

 

[Celeryman]

I also have a 1ghz t-bird slightly overclocked and an iwill kk266 and this is the most stable machine I have ever had. I just recently upgraded to W2k from W98se which was a very wise investment. I think W2k is a huge improvement over W98 as far as stability goes. And i love amd but have owned both amd and intel.

 

[Macro2]

It's beyond me how Tom could be so stupid as to publish such an unrealistic and damaging article concerning AMD...never has Tom published a more damaging article about AMD.
I hear Intel is now using that very same video in their presentation to corporate buyers.
If layoffs occur at Dresden...seems Tom's countrymen may have gotten a raw deal, perhaps in part to Tom..

Mac

 

[Texmaster]

<<

<< But heatsinks need a fan to do their job. The result would be the same it would just be a few minutes more. >>


Not if you use a decent mother board that will shut itself down at say a 70c reading or a temp moniter proggie like mbprobe that moniters fan rpms and will shut down if the rpms go to low. The fact that a fan fails doesn't automatically mean fried cpu. A decent heatsink should keep things in check long enough for the system to shut down. I modified my connector on my delta black label to moniter rpms even while running off the power supply before anyone offered those adapters or fans could be bought that way just for that reason. I expect that the odds of a properly installed heatsink simply falling off are astronomical and the fan failure thing is a non issue if proper care is taken in the configuration. Besides for cost of a P4 setup you could accidenly smoke a couple tbirds and still not be in the hole moneywise. >>

\

I dont know about the money thing but only a few montherboards have the technology you speak of. I'd rather have peace of mind.

 

[FatFrancis]

i think that a lot of people are mistaken as far as the price difference between amd and intel goes... on a clock for clock basis, they are very similar... within $15-20

of course, there is the ram discrepantcy... around $50 for 256 megs...

and if you get an asus or abit board, you are probably looking at similar costs for p4 and athlon boards...

so we'll say $70 difference... maybe a little more, but not much....

I really dont think this is enough to go crazy about... If someone wants an intel, let them have it...

and no, you cant buy 2 amd's for 1 pentium system

FF

 

[Texmaster]

<< All you people that constantly mention the heat factor with AMD are very misinformed! The P4 is not exactly cool running, but yes it does run cooler than a Tbird at clock speed. The Palomino core is cooler running than the Tbird core. The Palomino runs cooler than a similarily clocked P4 from what I hear. By the way the coolest running processor if you are paranoid about heat is the Via C3! The transmeta cruesoe also runs at a much cooler temp. >>



You obviously didnt see the video

 

[ST4RCUTTER]

I think what a lot of people are missing is that the P4 and the Athlon are made by different companies. They accomplish the same thing but go at it at in different ways. Intel has the capital and the connections to ensure their chips behave within narrow margins of error and with guaranteed safety stops. AMD chips are made by a competitor a fraction of the size of Intel. They make CPU's in the raw, with far less features because they can't afford to do it any other way. They charge far less for this fact. If that means I have meet them halfway to make everything play happy that's ok. I am a tweaker, and that's what this chip is good for! I see plenty of people complaining about their Athlons running hot, but I have yet to read about one dying because the fan failed.

 

[Imperium97]

<< It's beyond me how Tom could be so stupid as to publish such an unrealistic and damaging article concerning AMD...never has Tom published a more damaging article about AMD.
I hear Intel is now using that very same video in their presentation to corporate buyers.
If layoffs occur at Dresden...seems Tom's countrymen may have gotten a raw deal, perhaps in part to Tom..

Mac >>



What?
anywy, the video was showing the capabilities (or lack thereof for AMD) of clockthrottling. If you want to talk about how unlikely a heatsink would come off a chip, fine, but that wasn't the main thrust of the video. Of course I maintain that while heatsinks may not fall off once they're on, its quite easy and feasable that a T-bird would go up in smoke duting the installation of a heatsink. Everyone makes mistakes and if that heatsink isn't firmly on the cpu core then your thunderbird is fried. You won't know what happend until it's too late. That seems like a pretty legit concern to me.

If layoffs occur at dresden it's because AMD isn't making any money and that's not Tom's fault. You guys like how cheap your AMD processors are, but how long can AMD do that and stay in business? Didn't they just report a loss of 110 million in Q3? And now dresden is their only major means of chip production...and motorola owns half of it.

 

[MrHelpful]

You're probably a web designer from a different country, right?

 

[MrHelpful]

<< i think that a lot of people are mistaken as far as the price difference between amd and intel goes... on a clock for clock basis, they are very similar... within $15-20 >>


I think that's where most people go wrong. Any Thunderbird/Palomino will beat a similarly clocked P4.

 

[Mikewarrior2]

The Palomino is cooler overall than the p4, but....


The Major Reasons: Wattage per surface area is drastically higher on the Palomino. Cooling is much harder. The IHS is a huge key to why p4s "run" and operate at cooler CORE temps than palominos. Noticed I said core, not some crap @ss socket-thermistor reading. THe p4, internal diode versus internal diode reading, is much cooler than any comparable or even near comparable speed AMD core temp.



Mike

 

[ST4RCUTTER]

If layoffs occur at dresden it's because AMD isn't making any money and that's not Tom's fault. You guys like how cheap your AMD processors are, but how long can AMD do that and stay in business? Didn't they just report a loss of 110 million in Q3? And now dresden is their only major means of chip production...and motorola owns half of it.


Actually, Motorola doesn't own squat of it. They (AMD) licensed their copper interconnect and High Performance Logic Process technology for the Dresden fab. They have agreed to collaborate on future copper interconnect technology, but as far as I know Motorola has no claim to fab 30. If anyone has a claim to it it's the German government. The German Federal Government and the Saxon State Government support the project with approximately DM 800 million. I think AMD has been paying off the loan however. AMD sales stayed flat for this quarter which means they're selling a lot of processors. The only reason they've posted a loss is because the ASP's are in the dirt, for both AMD and Intel...

If you mean fabs 14 and 15 by saying "...now dresden is their only major means of chip production..." I would stress that those fabs were 15 years old and costing more money than they made. They were used for programmable logic chips based on huge processes (.25um+). An economy like this requires a slim down of wastefulness.

 

[Jejunum]

holy crap how many times do we hvae to see this link!!
lock this thread WE hAVE SEEN IT A BILLLIIOON TIMES

stuipd pro intel fanatics

 

[Imperium97]

My point was that motorola produces chips at dresden, right? That's still less maximum production for AMD. About the losses for AMD, who cares where they came from, point is, they came. For a company who's income for the past 12 months is 752 million, I'd be worried.

 

[Mikewarrior2]

Another REason why Palomino's need a working Internal DIode

Yes, I know it was a t-bird... but he got 2 seconds of life from it... What was wrong with it that caused it burn up? probably a slightly tilted heatsink... A Palomino would have reacted the exact same way... And for those who say socket-thermistors throttle back at 70C... they react too slowly to sharp temp increases. There's no way it'd be able to tell from startup, taht there's no heatsink and it should throttle back.

Bottom Line: The diode is a NECESSITY. Having internal diode reaidngs on motherboards is something amd should be REQUIRING of their motherboard manufacturers.



Mike