• 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.

Tbred over at AMDmb.forums

M

Mikewarrior2

Diamond Member
Hey everybody,

Someone over at the amdmb forums got a tbred 2200+. Setup on his 8k3a+(reading the internal diode of the tbred) and watercooling, he's getting shocking results.

http://www.amdmb.com/vb/showthread.php?s=&threadid=135142

He/She is running ~10C hotter with the tbred than wtih a palomino. So there are some real-world, independant temp readings. And not socket-thermistor, but internal diode. I had originally thought that tbreds would run slightly cooler with watercooling, but maybe not.....


Mike
 
Did these boards read xp internal diodes??? Maybe it just shows what you stated all along that the temps we were seeing with our amd chips were totally inaccurate because they were not measuring at the core....

Are we comparing apples and apples here??? Internal diode versus socket a thermistor
 
I e-mailed him regarding the 8k3a+, but i belive he was using 8k3a+ with both chips. will post his reply.

Even in the case he didn't use the 8k3a+ with the palomino, there's a few things to consider. Abit boards tend to read higher than the average mb, so the difference between int. diode and socket-thermistor is somewhat minimized. Secondly, socket-thermistors report closer to true die temp when there is no airflow interference. Another item that would somewhat minimize temperature disparigy from teh two sources.

And thirdly, since he was comparing idle temps, you're looking at sokcet-thermistors usually being able to give fairly good idle temps with most "non-asus compensation" mb's.

As far as socket-thermistor inaccuracy, i think the procooling.com, voidyourwarranty.net, and some diligent independant people at hardocp have confirmed over and over again that socket-thermistors are garbage for temp readings.


Mike
 
T-bred looks weak and not really anything to get excited over, Barton will probably be a different story. I think I would rather stick with the Palmino for now due to it's better cooling.
 
Daaang... 🙁 That power-to-surface-area ratio needs to change!

Ok, so... EPoX has boldly taken the initiative and put out a board and/or BIOS revision which gives real-time readings of the CPU diode. Excellent! Now if I only had some money... 🙁 Ah well, if I did, I should be saving it toward a Clawhammer anyway. 🙁 Thanks for the info Mikewarrior2!

 
Duvie,

"predicted' wattages, but usually these are accurate... Granted, the testing software usually does not max a core fully, but under teh same test situations, the Pally at higher voltage is hotter overall than the Tbred he's running.


Mechbgon,

totally agree with you. Kudos to Epox. They took the true first step, and hopefully others follow. Seeing the temp range on the Tbred with hte 8k3a+ between the amdmb forums and the hardocp "overclocking" test are fairly good indicators that these are really internal diode readings (readings go from 46 to 80C). That, and the repeated posts from Epox employees at various forums stating that the 8k3a+ does read and give temps from the internal diode.



Mike
 
I dont know why people say the Thoroughbred runs cooler. It does not. It runs awfully hot. Actually the hottest yet. Hotter than the Thunderbird.
It generates the least heat, but that heat has 60% less area to be cooled off of. Here some numbers.

Palomino

64.3 watts/128 sq. mm = .502 watts per square millimeter

Thoroughbred

56.4 watts/80 sq. mm = .705 watts per square millimeter

40% more heat to be dissipated per square millimeter.

Thunderbird 1400

64.7 watts/120 sq. mm = .539 watts per square millimeter

hope this cleared out some issues.
😎
 
> dont know why people say the Thoroughbred runs cooler.

If you are using the actual die dimensions, the energy density, and therefore
the temperature is higher.

Normally this is not the case when they reduce die size because they simultaneously
reduce the operating voltage, and the energy density goes down. For some reason
AMD did not do this. It seems strange. Smaller dimensions also reduce transistor
switching speed, and so increases top chip speed, but the Thbred does not appear
to be faster than the Palimino either, which is also strange. I wonder what AMD is
up to?
Keeping the voltage high should also keep current high, tending to also increase
switching speed.

A smaller die should still make the cost per chip lower once the yields get good. But
if AMD can't keep pace with the P4, Athlon is going to lose pricing power like they did
with the K6. Times are going to get very tough for AMD, and I wonder if Hammer can
save them.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top