High memory voltage on Phenom...

[gOtVoltage]

I'm happy with that...

What Tony says is absolutely true! Mem controller in K10 is way better compared to K8. My Ballistix is working flawlessly with 2.25V @1000MHz 4-4-4 timings. On K8 it needed 2.42V and it wasn't completely stable ...

If increasing NB voltage will mitigate risk of breaking CPU with over 2.2VDIMM then it's OK and most affected are AM2+ board owners (due to 1.1V NB power plane).

Thats good to know for us AM2 users....i cant wait for new revisions of stuff....i keep holding back longer and longer untill there is some thing stable...

Another thing isnt AMDgoing to surprize the Public today about blah blah something...?

[nemrod]

Another thing isnt AMDgoing to surprize the Public today about blah blah something...?

Good surprize:
http://www.newsobserver.com/1566/story/820114.html

[karbonkid]

Good surprize:
http://www.newsobserver.com/1566/story/820114.html

Mike Uhler, former CTO of MIPS, has joined the company as its first vice president of Accelerated Computing

MIPS: best. CPUs. ever.
I still mess around on SGIs when I get the chance
This is awesome news indeed...

[darkorb]

so whats the deal with K8 and high voltage? i dont wana break this puppy!

[nemrod]

MIPS: best. CPUs. ever.
I still mess around on SGIs when I get the chance
This is awesome news indeed...

Sure but AMD lovers should not hype always inquirer news... All time this a disappointing reality after. I'm sure that one of the big trend of the amd bashing is more than one year of stories of inquirer or fuad equivalent cleaming super powerful extraordinary cpu... Obvisously the reality is then very deceptive...

[KTE]

Lightman: are your Ballistx D9GMH? That's quite bad for those IC's, they normally can do 2.3/2.4V 1200 4-4-4-4. Mine are just binned versions of your chip (same SPD, but added EPP) and they do 1240 4-4-4-4 PL5 stock voltage since I bought them (2.2V).

Running more than that and you are rsiking it big time. My older D9 (crucial) with a 252CFM delta fan on them died in 3 days when I benched at 2.3V/2.35V for around 1 hr. One stick was fine and one was full of 8000+ errors after Test#1. And they were kept far colder than at stock, ambient was 4C and they were only slighty warmer than that.

[Face]

I think that's where it originally came from:

http://my.ocworkbench.com/bbs/showthread.php?%20threadid=69350

AMD Phenom gets killed easily with high vdimm set
--------------------------------------------------------------------------------
According to a reviewer who has been testing Phenom CPUs, he has killed 4 Phenom 9000 series CPU by applying just 2.3-2.5v to the vDIMM setting.

So, beware if you intend to o/c the Phenom, bumping up the voltage to 2.3v can easily kill the memory controller resulting in a dead CPU.

[KTE]

Can someone who believes in this try to explain it with some technically sound knowledge? It makes little sense to me, if I look to the physics involved unless we're missing something crucial, in which case I want to know what it is.

[RGone]

I think that's where it originally came from:

http://my.ocworkbench.com/bbs/showth...threadid=69350

That link works for me in describing the situation when running a PHENOM processor. No one knows all the whys nor will likely. When a warning is given, peeps can take the warning or do with their equipment as they please. Always cometh the one whom finds himself the exception to rally against anything not happening to them, happens all the time. However in reality you can likely save yourself some headaches by realizing that over 2.2Vmem on PHENOM processor stands a fair chance to damage the memory controller and since it seems PHENOM only that should be logic enough to know it ain't the boards doing the deed.

[Brother Esau]

Hey, Robert have you gotten you're DFI 790FX-M2R yet?

[AlabamaCajun]

Voltage leads, current follows. When you increase voltage, more current is pulled through. If the VDimm is higher than the IMC then the IMC has to sink more current to create the zeros on the Address and Data lines. The term "sink" I use to definite to pull down as from high about 2.1v (Logic 1) to low at around 0.25V. ( Logic 0 ) The source in this case is a bank of resistors feeding power into the A and D lines.

Lets say the resistors are 560 ohms using ohms load we take Volts / Ohms = Amps
2.0V / 560 = 3.57 Miliamps.
2.1V / 560 = 3.75 Miliamps.
2.2V / 560 = 3.93 Miliamps.
2.3v / 560 = 4.11 Miliamps.

You can see as the voltage is rising so if the amount of current the IMC has to pull to get a zero on the line.

[Tony]

Hey, Robert have you gotten you're DFI 790FX-M2R yet?

he has one yes....he will be posting on the club with it

[Timbosan]

This is interesting. if voltages over 2.3v are likely to immediately? (time frame?) kill a Phenom, wont running 2.1 - 2.3v be coming very close to that? How long will a Phenom be able to handle even 2.0v before it burns out? Surely if 2.3v can kill it in a matter of minutes or hours, wouldnt 2.0 - 2.3v kill it in less than a year of use?

[Brother Esau]

Tony......this is what I am referring too with Default Values AUTO Detect Being wrong in PCI-E with Voltage allocation when Cross Fire is in use or more than 2 PCI-E Devices

[bingo13]

This is interesting. if voltages over 2.3v are likely to immediately? (time frame?) kill a Phenom, wont running 2.1 - 2.3v be coming very close to that? How long will a Phenom be able to handle even 2.0v before it burns out? Surely if 2.3v can kill it in a matter of minutes or hours, wouldnt 2.0 - 2.3v kill it in less than a year of use?

I killed my 9600 in a matter of minutes at 2.38V real while running under load, now the 9500 took a couple of days but that bastard needed to die. The good news is that I can hit DDR2-1200 5-5-5-15 on the Gigabyte board (almost on the MSI board - 1180) with 2.2V. At the request of AMD I am not going to say what I just said so you never read what I did not say.

[darkorb]

can anyone link me to an article or somthing that explains this more in depth? i dont wana bust my cpu!

[KTE]

Thanks.

So bingo13, what happened to your dead CPU's, RMA?

[bingo13]

Thanks.

So bingo13, what happened to your dead CPU's, RMA?

Newegg is not too happy with me right now.....

[Jim Morbid]

Newegg is not too happy with me right now.....

Hahahaha, the perfect answer Bingo13

So what happens if I have a pair of Corsair TWIN2X2048-10000C5DF Dominator's which SPD at 2.4V? They should in theory fry my cpu as soon as I put some load on them without overclocking anything?

JM

[dinos22]

that's pretty sad news

AMD is basically building CPUs for stock machines

why even waste time in PR on forums like extreme systems is beyond me

[Brother Esau]

I get the same impression dinos AMD qualifies for the stock PC sector too bad I had very high hopes for all of this and personally as a customer I feel betrayed the way they blatantly lied to all of us me included.

[KTE]

why even waste time in PR on forums like extreme systems is beyond me

Are you talking about macci here?

You seem to be.

[Cheffy]

well, look this way, it might encourage low voltage pc2 8000 (+) ram to be released. and one would assume that would clock better on high voltage tolerant systems.

darkorb - you are fine. this relates to K10 and VERY OLD K8 - eg the first clawhammer cpu's you have an x2. x2 = new k8.

my understanding goes thusly:
on an amd K8/K10 system the v-ddr is getting applied to certain areas of the dram controler, because it is sat on 1 end of a bus running on vddr voltage! its not soemthing that can be avoided. chipsets that are connected to the ram for non imc cpu's are also exposed to it, but they tend to be running much older, more rhobust process techniques - eg i think the KT800's are .25, while the first K8 were .13. thats a big difference. and as we all know, process shrink, voltage shrink.

amd fixed the delicate imc in the end, im sure that the same will happen again. - and even back then, if you ran an old core, you just used micron/samsung rams at upto 2.8v, and avoided bh5 at >3v. it wasnt the end of the world.

[dinos22]

Are you talking about macci here?

You seem to be.

i am talking about AMD.

why would you make it personal......i enjoy reading anything the big man puts up as he glows in knowledge almost none of us here have......you can always learn something new from the guy

[KTE]

Sorry dinos I was asking you to clarify since when you said "why even waste time in PR on forums like extreme systems is beyond me" I couldn't understand who that AMD PR you're referring to is apart from one who is involved here.
I agree with you, having these new CPUs and even the X2s macci has been the best of help, really appreciate his contribution or no one would help us resolve anything of the many issues 'cus no one else really knew anything factually as good as he does because of his poistion. Even MSI/ASUS don't help no where near as much even when they know details, they all get too high headed and show an attitude.