Thoroughbred to Require 45 Amps

[68GTX]

What's up with this?

The following is from a story over at HardOCP. The following information from the AMD tech tour 2002 is primarily for system builders.

"The Thoroughbred core was touched on a bit, but not really much even though people are saying the launch is just around the corner. "Smaller, Cooler, Faster" was the message being broadcasted. It was discussed that the Barton (SOI core technology) would have 384K on-die L2 cache as well."

"Thoroughbred CPUs will initially stretch deep into Palomino MHz territory. Here is not only a list, but also required voltages for the TBred CPUs."

1.5V - 1700+, 1800+, 1900+
1.6V - 2000+, 2100+
1.65V - 2200+


"Also PSU power was heavily touched on for the system builder present. It was stressed that over wattage rating for PSUs was not important, but rather the 5V rail is what to pay attention to. A minimum of 45AMPs will be required on the 5V rail due to TBred power consumption. The rail amperage rating should be printed on pretty much any PSU you find."

Can that power requirement be correct? My 400 watt Antec power supplies only put out 40 Amps on the 5 volt rail.

 

[Viper GTS]

My TruePower 550 only puts out 40A as well, that's an insane requirement.

Viper GTS

 

[BD231]

The T-Bred is looking worse and worse every day, wtf did AMD do? They should have gotten similar results to the P3 Die shrink shouldn't they? Who know's, maybe all this garble about the T-Bred wont be half as bad once the processor is out. I'm skeptical though, when processors have insane requierments it usualy means problems.

 

[Derango]

Glad I decided to go for a P4 instead of waiting for the T-bred. Thats just crazy!

 

[NOX]

Wow, is this really true? Not even my Enermax EG365P-VE FCA has 40A, it only has 32A. In order for me to run this proc, I would need an 551W (same brand) which has 46A, even the 431W has only 44A.

 

[salman327]

I don't think it's true. Right now it's all pure speculation, especially since no reviewers have gotten their hands on a test model. AMD needs to hurry up and release it soon, I need a new proc.

 

[spanner]

I thinks thats bullcrap, no way a tbred would take that much off the 5v rail, I am pretty sure the power would be sucked from the 3v rail. Anway I suppose there is no way it will work on my 8K7a and my 431W enermax. Oh well, I am going hammer anyway.

 

[Adul]

i dont think it is really going to take that much. sounds like maybe it a requirement for 2 cpus

 

[ViRGE]

45amps I could believe; 45amps all from one rail, no. There has to be some sort of mis-understanding there, there's no way that any CPU could need 45amps all from the same rail; and this is before we calculate what the drain is for other devices that use the 5v rail. Don't worry about a thing until the reviewers have Tbred's in their hands, as anything before now is pure speculation.

 

[pillage2001]

THis is bad, my PSU only pulls out 22A. I don't get it, why such high current requirement? I guessed they messed with the transistor sizing too much that it cannot pull enough current to flow through it.

 

[Wolfsraider]

on the front page is:

Many of you are writing in asking if I have a typo in the above AMD Tech Tour story. The 5V rail amerperage has some of you upset which I have to fully agree with you on.

It was stressed that over wattage rating for PSUs was not important, but rather the 5V rail is what to pay attention to. A minimum of 45AMPs will be required on the 5V rail due to TBred power consumption.

There are no typos here and we have already sent mail to AMD in Austin asking them to verify the AMD remarks from Thursday night. Hardly any current PSUs carry 45 amps on the 5V rail and this does not seem to be correct. We will let you know as soon as we get verification.

so i guess we wait for confirmation

and i am not trying to be mean just providing an update >mike

 

[Comp625]

Wow...:Q

 

[Piano Man]

That has to be a typo. It would be suicide to have that requirement.

 

[Emo]

The Thoroughbred looks more and more like a stillborn cpu, doesn't it? According to AMD's own roadmap we should be seeing the Barton in the summer and hopefully it will be a Socket A-compatible cpu.

 

[ShadowDJ]

:Q
wow... there goes a possible upgrade!

 

[Didou]

Isn't the 45amps requirement only for the 2200+ ? The lower speed grades are supposed to work with current PSUs.

 

[neuralfx]

eh it's a little early .. I don't think that will be the case .. It would be suicide to an extent, how many OEM's would want to use T-breds if they have to also supply PSU's that will provide 45A on the 5v rail? .. I think it its a little minsinformation or maybe miscommunication ..
-neural

 

[68GTX]

I wonder if they are talking about this.... "AMD is warning that to support 2.0GHz operations, the onboard regulator must be capable of yielding 45 amps." from theinquirer.net

 

[Sid03]

wow. they must be having some serious problems with .13 micron.

with these kinds of obvious problems with a straight die shrink, i can't see how they can possibly get the hammer out on time.

 

[Modus]

This seems highly unlikely.

1) AMD is not stupid enough to destroy their stong PSU support by making a chip that requires *everyone* to find a new, insanely powerful PSU.

2) The Thoroughbred is a die *shrink*. When was the last time a die shrink of an existing processor *increased* its power consumption? Even acknowledging the higher clock speed, it doesn't make sense.

Modus

 

[robg1701]

AMDs PSU recommendation thing was always on the basis of cobined 3.3+5V power...as otehrs have said, i cant believ for a second that its from one rail.....way back when the KT7A etc were hitting, people were making a big deal about the 3 phase power, and the requirements for the next gen Palomino being up to 40Amps from the regulators..thatd be a combined thing as well.....so, im expecting AMD's response to pretty much say "45A was correct, but from 3.3 AND 5V rails.... "

 

[Barnaby W. Füi]

<< This seems highly unlikely.

1) AMD is not stupid enough to destroy their stong PSU support by making a chip that requires *everyone* to find a new, insanely powerful PSU.

2) The Thoroughbred is a die *shrink*. When was the last time a die shrink of an existing processor *increased* its power consumption? Even acknowledging the higher clock speed, it doesn't make sense.

Modus >>


i agree, it has to be a typo. 5v @ 45a is about 250 watts, you think you're gonna need 250w just to power the cpu?? give me a break. it's either a mistake, or AMD is going to have lots of trouble.....

 

[CTho9305]

<< i agree, it has to be a typo. 5v @ 45a is about 250 watts, you think you're gonna need 250w just to power the cpu?? give me a break. it's either a mistake, or AMD is going to have lots of trouble..... >>


he's back!

anyway, I htink 45A at vcore, not 5v. if it was at 5V, your mobo would explode

 

[WarCon]

In straight power consumption that would be 225 watts. No way that processor pulls that kind of power without a chernobyl like experience.

Maybe that reference is for a single phase solution though and not a 2 or 3 phase regulator.

Or maybe it was a misquote by a sales rep/ignorant tech who doesn't understand regulator circuits. I believe that my 1gig T-bird that I ran at 1.4gig with 2.05vcore had an estimated power usage of 100 watts. If that was solely dependant on the current capability of my Enermax I was only 17 amps shy on its rating. 100/2.05 = ~49amps.

JMTC

 

[BD231]

Someone just screwd up with the 5 volt rail thing. I can totaly see 45a when the current Palaminos are 40a. AMD is smarter than that.