Undervolting Experiences with Athlon64 X2

mrochester · Post by **mrochester** » Sun Sep 11, 2005 10:00 am

Hi there guys,

I have an almost totally silent system with an undervolted Athlon64 X2 3800+ processor. However, my Asus A8N-SLI Premium motherboard has a bug in the BIOS which prevents you from using a vcore of lower than 1.1volts. I've contacted Asus about it, and I hope they release a new BIOS soon to fix this. However, I'd like to know how low other members have been able to drop the voltage while leaving the processor at it's stock speed? Mine works perfectly at 2Ghz at 1.1volts, and I suspect it could go lower if I had a working BIOS!

Many thanks,

Michael.

eLekTRiK · Post by **eLekTRiK** » Sun Sep 11, 2005 11:29 am

I just set up an X2 4400+ on a DFI nF4 UT Ultra-D board. The processor stock VID is just 1.30V, and it also will only accept settings down to 1.10V.

I think the 1.10V is a constraint that the processor imposes (just like the FID locks on non-FX CPUs). At the very least, it is not just ASUS.

I am not sure why AMD would be doing this, except possibly to create more market share for mobile processors that have lower VIDs enabled.

mrochester · Post by **mrochester** » Sun Sep 11, 2005 12:09 pm

Well that plain stinks

I was looking forward to getting my 3800+ even lower. Trouble is, there is no X2 mobile alternative, so why are AMD imposing these limits? The reason I thought it a BIOS problem is that I also can't overvolt any higher than 1.35volts, and I've read of plenty of people going higher than this, even with the same motherboard and BIOS I have and an X2 processor.

swissguy77 · Post by **swissguy77** » Sun Sep 11, 2005 12:50 pm

Same problem with my setup: an X2 4400+ (E6-stepping) with an Asus A8N-E. At 2.2GHz it goes down to 1.15V (but I'm working with 1.20V, just to be save) At 2.0 Ghz I’m also able to work at 1.12V. My board always adds 0.02V. No voltage lower than 1.12 possible...

I hope this is fixable with a new bios, because I think these new X2 have an extraordinary undervolting capability. I’ve never heard of any “voltage-lock“ that AMD could build into the processor.

If the limit is in the processor then older core’s (e.g. CG-Stepping) should have a lower possible “lowest voltage”. If it’s just a safety feature in the bios, the different steppings should behave identical. (unless there is a specific lowest voltage for every stepping in the bios…)

So…

Who has an “old” CG- or a D-core in an NForce4 board that has a lower “lowest voltage” than 1.1V?

Who has an E-stepping core in a non-NForce4 board that has a lower “lowest voltage” than 1.1V?

bobov · Post by **bobov** » Sun Sep 11, 2005 7:07 pm

Hi, I can run my X2 3800 1.1v at 2.0Ghz prime95 24h stable, below 1.1v prime95 failed.

mrochester · Post by **mrochester** » Mon Sep 12, 2005 1:06 am

Hi, I can run my X2 3800 1.1v at 2.0Ghz prime95 24h stable, below 1.1v prime95 failed.

What motherboard are you using to be able to set the voltage lower than 1.1?

ronnylov · Post by **ronnylov** » Tue Sep 13, 2005 6:57 am

Anybody tried undervolting with ASRock 939Dual-Sata2 motherboard?
According to the specifications the voltage range is 0.80 - 1.55 V but the maximum voltage is automatically adjusted to 0.05 V above default voltage and I don't know if it does something similar with minimum voltage. I am thinking of maybe buying this motherboard and AMD X2 3800+.

Here's a review of this motherboard:
http://www.anandtech.com/mb/showdoc.aspx?i=2524&p=1

GHz · Post by **GHz** » Tue Sep 13, 2005 1:08 pm

I have the same problem. I have an Epox 9NPA+ (NF4 Ultra) and I can't get vcore below 1.1V. The BIOS allows for lower than 1.1V, but those settings still report a vcore of 1.1V, as if they were being ignored.

I'm running a 4200+ E4 2.2 GHz @ 1.25v, Prime95 stable.

spworley · Post by **spworley** » Tue Sep 20, 2005 4:44 pm

Slightly different regime, but another datapoint.. I'm very mildly overclocking my X2 4800+ from 2400 to 2520Mhz. It does this at stock 1.30 voltage. However, this voltage seems necessary for this speed, 1.275 volts will almost work, but dual Prime95 testing failed after 5-8 hours. 1.30 volts successfully ran 30 hours before I stopped.

I have not run underclocked voltage tests, those would be interesting. I am only overclocking since my machine is silent anyway. (The Scythe Ninja + Antec P180 + Seasonic + 7800GT combo is unbeatable! Only drive seek noise remains..)

Hifriday · Post by **Hifriday** » Tue Sep 20, 2005 6:49 pm

Possibly 1.1volts is the minimum VID limit for all rev E cores?
Although not X2s, I also could not lower vcore below 1.1v for the following rev E CPU/mbs even though settings did go lower.

A64 3500+ Rev E3 Venice - DFI NF4-D
Sempron 3000+ Rev E3 - MSI K8MMV
Sempron64 3000+ Rev E6 - DFI K8M800-MLVF

Mats · Post by **Mats** » Tue Sep 20, 2005 8:20 pm

Have you guys tried software like CrystalCPUID?

mrochester · Post by **mrochester** » Wed Sep 21, 2005 1:09 am

A friend tried lowering the voltage on his Venice 3000+ last night and that refused to go any lower than 1.1volts. He was using CrystalCPUID to do it.

dan · Post by **dan** » Wed Sep 21, 2005 5:31 am

what is the estimated power consumption of an x2 @ 2ghz and @ vcore 1.1v?

Schlotkins · Post by **Schlotkins** » Sat Sep 24, 2005 3:35 am

dan wrote:what is the estimated power consumption of an x2 @ 2ghz and @ vcore 1.1v?

I would love to know this too.

Chris

mikellpp · Post by **mikellpp** » Sat Sep 24, 2005 6:28 am

Looks like the x2-3800 is a real cool cpu.

http://www.lostcircuits.com/cpu/amd_x2-3800+/10.shtml

~El~Jefe~ · Post by **~El~Jefe~** » Mon Sep 26, 2005 7:55 am

Wow.

Yeah hm. I guess when I get my dual core I will just put it to 1.2 volts vs 1.3 stock. that's some incredible numbers there.

I wonder how the 4400 does vs the 4200

widman · Post by **widman** » Mon Sep 26, 2005 4:25 pm

The problem is not on motherboard but is on CPU itself. I have Clawhammer 4000+ on MSI Neo2 K8N, run on 0.9V 800Mhz perfectly. But when I plugged Venice 3000+ to this motherboard, it refuse to go below 1.1V. I think there is 'spec' to limit the min voltage.

widman · Post by **widman** » Mon Sep 26, 2005 4:35 pm

dan wrote:what is the estimated power consumption of an x2 @ 2ghz and @ vcore 1.1v?

based on this review
http://www.xbitlabs.com/articles/cpu/di ... 800_3.html

load ~65W idle 5.8W.

I doubt that low,