GA-K8NSNXP-939

[Lotdog]

Is there anything bad to say about it?

 

[anthrax]

Its loaded with every thing. My board goes up to 248MHz HTT..Thats not bad for a 939 MHz board.. The only issue is that the features might be a bit of an overkill................ The biggest advantage that it has over the MSI K8N Neo2 Platinium is that you can actually buy it unlike the MSI board.

 

[Lotdog]

it's loaded with everyhin in bad or good way?

 

[pantner]

Its great...a few small probs right now....

it doesn't read the CPU temp correctly, my 3800 apparently runs at 12C-18C....but this is just a BIOS issue, and should be fixed soon...GB are working on a F4 BIOS (the board comes with F2, and F3 is on the GB website) which will hopefully fix it...

And it doesn't like certain RAM...to get past 210FSB i have to clock down the mem to a 4:5 ratio....but i have cheap generic RAM....which i could get to 235FSB in my old motherboard...

one more thing....right now, it doesn't like having all 4 RAM slots used.....once again, prolly a BIOS issue and should be fixed soon...

once all those probs are fixed it'll be a perfect MB!!!!!!!

 

[Chronoshock]

Originally posted by: anthrax
Its loaded with every thing. My board goes up to 248MHz HTT..Thats not bad for a 939 MHz board.. The only issue is that the features might be a bit of an overkill................ The biggest advantage that it has over the MSI K8N Neo2 Platinium is that you can actually buy it unlike the MSI board.

The MSI board will be out within a week (see the other thread comparing the GA-K8NSNXP-939 and the MSI K8N Neo 2 Platinum)

Also, the GA board is ~200 dollars where, judging by the price at ZZF, the MSI will retail for ~150

 

[mickles]

In a week. That is what you/we are hoping.

 

[anthrax]

Actually, I am taking the the things i said above back..

I just got my Corsair XMS-3200XL running..

CPU @ 2200 MHz
HHT 200MHz
RAM voltage @ 1.7V

Its nothing but trouble. Prime 95 will fail in less than 5 mins.
Memtest however works fine... Reading around in the corsair forums & the overclocker.co.uk forums. there seem to be a HUGE issue with DDR3200 running in dual channel mode.

 

[tieros]

Yup, I did an RMA on my K8N-939 almost a month ago due to memory problems in dual channel mode.

I tried three different brands of 512MBx2 RAM, including Corsair 2-2-2-5 "WalletBuster" sticks and none of them were stable without dropping the mem clock to 166. They all worked perfectly in single channel mode, though.

Fortunately, Newegg was out of replacements, and issued a refund instead. It's been a long 3 weeks, but hopefully I'll have an MSI before week's end

 

[trevorlj]

There's a few of us out there that have the board running relatively well. The only problem I have is the CPU temp thing (mine reads 25-34C depending on load). I have the "wallet buster" Corsair chips running at optimum timings and 204 fsb rock solid. They move about 6050-6080 MB/s in Sandra.

It's a good board on paper but from browsing different forums it defnitely looks like it was pure luck that I got a board that works this well.

 

[tieros]

Congrats on getting a good one! It really is a nice board, with dual everything, and more accessories than you usually get on any two other motherboards combined.

I would definitely recommend it if they can fix the memory issues lots of people are seeing, especially if it's a simple BIOS revision.

 

[SIKLCELL]

This board totally sucks if you are going to overclock. Well I hope my MSI gets here tomorrow. My Gigabyte board is available CHEAP if anyone wants it.

 

[anthrax]

The conclusion is, Don't buy it unless they fix the issues..

 

[cluster550]

I bought this board because the MSI board wasn't out when I needed it. So far everything has been fine except for the CPU temperature (reporting too low), but I have yet to try any overclocking. My board came running the F3 BIOS which is the newest one. I will probably end up giving this one to my dad in a few months because I want to make use of eVGA's Step Up program if I can get PCI-X parts by early November.

 

[imported_Hornswoggler]

I have the board in question, and while it runs awesome in single-channel mode, when I put a second DDR DIMM in there, everything goes downhill... FAST!

System specs:
GIGABYTE GA-K8NSNXP-939 (BIOS F3)
Corsair TwinX-1024 (2x512) 3200XLPT (the 2-2-2-5 stuff, non-bling edition)
3500+
x800 Pro VIVO
blah, blah

I have tried messing with numerous DDR timings, followed Corsair's RAM GUY advice, and still cannot get this board to successfully run dual-channel memory. Most common problem was a STOP 24, ntfs.sys BSOD. Once booted, it could take 2 minutes, or it could take 15 minutes but it would always crash. I have also swapped DIMM's and either stick works great by itself.

Single channel, works flawless. Just something unsettling about paying $220 for (what you think is) the best motherboard you can get, and have it turn out to be a dud. I don't want to run in single-channel mode for the 2+ years until I buy a new system so I am going to try sending this mobo back, and I already have the MSI nforce board on order.

I have no idea how none of the reviews for this board had any problems getting dual channel to work. I know that I am not the only one having this kind of problem. I have heard others getting better results by using "lesser" RAM.

 

[tieros]

I had the exact same problem.

The only way I got it to even be stable with dual channel RAM was setting the memclock to 166MHz, which is a pretty crappy way of getting a 200+ dollar board to behave. I hope it's something they can fix with a BIOS patch, because it really was a nice board.

I've got 2 of the MSIs sitting here now, thanks to EXTREMELY bad customer service from Monarch. They'll never see another cent of my (or my company's) money.

 

[imported_Hornswoggler]

yeah, for $220 I expected something a lot more reliable. :/

 

[cbns]

Yes my mobo reads the CPU temps like a donkey as well. Think its 23C degrees when I start it up cold and go directly into the BIOS to see the temp. When running MBM5 I also get psycho under 20C readings, and no higher than 28C case readings. My computer case has a temperateur probe that I have sitting on the side of the CPU heatsink and I get readings around 40C from that.
So you think the new BIOS version will fix it? I definitely hope so!

Another problem I have is the keyboard. My USB keybaord dies when I enter windows setup to put on an OS. I had to get a PS/2 keyboard to do it, and when it did ONLY the enter key worked. Luckily that's all I needed. Then I tried upgrading to XP Professional. Apon restart I my USB mouse and keyboard were still working until it windows setup siad that some of my network drivers weren't digitally signed for XP and if I want to installthem anyway. I clicked yes and then some other message came up and i coudln't click continue! my I/O ports died again! ALL of them!
I restarted many times trying to fix it again, and on the 3rd restart I could no longer resume the xp pro upgrade. I got screwed over, badly. I had to reformat and fresh install xp home and I lost everything.
I emailed those gigabyte bastards and they respond once a month. I can't get any help!
HELP ME!

 

[Ballistyx]

Well FINALLY! I thought I was the only one having these sorts of issues.

Here's my story. Take it for what it's worth.

What I started out with two weeks ago:

Athlon64 3800+
1 Corsair TwinX1024-3200LLPT kit
2x512MB PC3200 GEiL ValueRAM (before you go "Unngh!" keep reading!).
2x74GB Raptors
2x120GB WD-SE HD's with PATA to SATA adapters.
9800Pro vid card
Water cooling.
550W/36A (on all rails) server grade PSU

Now, I CAN GET all my RAM to work, with one small exception. Whenever I fill all four DIMM sockets, the memory speed drops down to 166Mhz. It doesn't matter what order you put the RAM in. It stays at 166Mhz.

I posted about this on nforcershq's forums, and one other person had an issue with this too. The exception is that his RAM worked at DDR400 speeds, just not anywhere near stable.

So on my quest for answers, I contacted Gigabyte and AMD. AMD gave me two different answers. One was that DDR400 with four modules is dependant on how the motherboard manufacturer followed their guidelines. The other answer I got was that DDR400 is only qualified with two DIMMs installed. You can imagine that after spending $700 for a CPU, I wasn't thrilled with hearing this.

Gigabyte wouldn't give me a straight answer. They insisted that I try four matched DIMMs. While I can understand that, what I never got an answer for was that if it was a memory timing issue with mis-matched DIMMs, why is it dropping the memory speed instead? I was then told to talk to AMD about it. Basicaly I just went in a complete circle.

In a fit of frustration, I ended up buying a Corsair TwinX2048-3200 kit (two 1GB modules). NOW the machine is running great at DDR400 speeds, but I took a good hit on latency over it (I won't even go into the hit on my wallet).

I did an interesting experiment with the new memory though. I attempted to get to 3GB of system RAM, just to see if it'd work. My results were interesting to say the least.

With the TwinX2048 and TwinX1024 kit, THE MOTHERBOARD WOULD NOT COMPLETE THE MEMORY TEST AT POST. Heck, it wouldn't even start it! So, just for giggles, I tried my "inferior" GEiL ValueRAM.

Not only did it post and count the memory, it booted XP just fine. Once I got into the OS though, it all came apart. I'm chalking that up more to the memory controller going crazy from the different densities of RAM on the DIMMs.

I then checked the memory speed in the BIOS. Once again it dropped down to 166Mhz.

If anything, I learned that the memory controller on the CPU is downright fickle. It only likes certain types of RAM. That's not an issue usually, since most higher end systems are like this. That's why these vendors have a qualified memory vendor list. I can find one for the Asus A8V, but good luck trying to find one for the K8NSNXP-939!

What it also doesn't explain is why the memory speed keeps dropping down to DDR333 speeds with four DIMMs. Either the second answer AMD gave me is correct (and they're not publicizing this fact), or Gigabyte is doing something in the BIOS to try and compensate for the extra noise on the memory bus. What's the answer? Your guess is as good as mine! Neither company will fess up and take ownership of the problem!

That's why I'm about to ask for all of your help on this one...

After spending nearly $2000 this month on computer parts, I'm a little on the poor side, and I can't afford to buy any more of the "good stuff" (RAM). I can swing some cash and buy two more GEil PC3200/DDR400 ValueRAM DIMMs and try running them and the ones I have to see if I can reproduce the problem (with four identical DIMMs from the same manufacturer). In the meantime everybody who's experienced an issue with this board (or ANY Athlon 64 board with memory issues for that matter) with two or four DIMMs in their setup should write to your favorite review sites and ask them to perform a four DIMM test on any Athlon 64 board they review. Although the problems do manifest themselves with only two DIMMs, using four DIMMs really stresses a board, and it'll make any weaknesses readily appear.

I don't know about you guys, but I feel a little burned and lied to with the lack of information and the lack of support from Gigabyte (and to an extent, AMD). Maybe if review sites start testing for instabilities in this way, we can weed out the bad boards from the bunch, and make both companies more aware of the problem.

 

[imported_Hornswoggler]

I have some news to report.

I am running the Gigabyte GA-K8NSNXP-939 and it fails the Prime95 torture test (always within the first two minutes) with BOTH Corsair 3200XL and Crucial Ballistix PC4000.

With the Corsair, I have tried all sorts of memory timing combinations (some even received from the "RAM GUY" himself) with no success.

The Crucial I just installed a few minutes ago and let the motherboard auto-detect SPD timings. Windows boots and appears to be working right, but Prime95 gives me this awful rounding error (same error as with the Corsair 3200XL).

So... I shift my blame back to either the motherboard or the nforce 3 chipset. Fortunately, I have an MSI K8N Neo2 Platinum mobo here that I can pop-in and test with either the Corsair or the Crucial.

Hmm, the more I think about it, the K8 does have an integrated memory controller... maybe the CPU is the cause of all these problems? Especially if the VIA guys are having dual-channel issues.

This sucks.

 

[pantner]

Ok....you can run 4xRAM at DDR400, apparently you have to enable T2...

rather then going out and buying a new board (ie, the MSI) wait a little while and give gigabyte a chance to fix the problem...consider this

1.The Athlon 64s are VERY VERY new
2.THe Athlon 64-939 are even newer!
3.The Chipset is VERY new
4.The Motherboard is (yet again) VERY new
5.This motherboard is on BIOS F3....others are up to F11

Take all those into account, and probally may other variables, and you may be able to understand how this is possible...now im not saying that exstensive testing should've been done, but thats one of the reasons you are able to update the BIOS, and why games have patches...you can't expect GB to release a BIOS fix the day after somebody complains, it takes time to find the problem, find the cause of the problem and then fix it...i have already tried 3 BETA BIOSes myself froim GB....now, none of them helped with the temp issue (the first 2 gave me a "- -" reading for the temp in the BIOS the 3rd shows a temp, but its still not correct...)BUT ATLEAST THEY ARE TRYING!!! If they had 'left us all out in the cold' then i could understand why people are buying other boards...

 

[Ballistyx]

Pantner, normally I'd agree with you. Here' s why I flat-out don't agree.

1) Hook up a SATA drive to the nVidia PHY controller. Tell me, does the activity LED show any activity when you access the drive while it's attached to that controller? If your board is anything like the majority of boards out there, IT WON'T WORK. Sure the SATA controller works, you just can't tell when your drive is being accessed. Gigabyte said to me that it was by design. I think it was a QA error, or should I say a lack of QA. The LED works fine on the other controllers though.

2) When I asked Gigabyte why the temps were so off, you know what they asked me to do? Remove my waterblock and try a normal HSF to see if the temp changes. Now I ask you this. How would downgrading from a water cooling system to a plain-jane HSF affect THE MOTHERBOARD temp sensor? I know that watercooling can throw off the temp reading for the CPU, since the BIOS calculates it on the assumption that air is flowing over the area. If anything, it should be off by a few degrees, not 20 degrees. I tried to explain that to the "tech" (and I use that term VERY LOOSELY) from Gigabyte, but they won't respond to my e-mails.

3) Now here's my favorite.... EasyTune4! Tell me if you can get the program to work without errors on your machine. Whenever I try to run the version that comes ON THE INSTALL CD, it WILL work, but I get errors when I start it about the program not being able to find certain resources on the board. When I asked Gigabyte why it's not working correctly, you know what they told me? IT'S NOT SUPPORTED BY THIS MOTHERBOARD. If it's not supported, then WHY did they put it on the CD, and put the %^#(*&$ logo on the box for it?? Oddly enough, the version of EasyTune4 that came with my Gigabyte 7NNXP works fine, albeit you can't check the temps or the voltages. For that I use nVidia's ForceWare app. THAT works, unlike Gigabyte's utility.

Being new is not an excuse for all these QA bugs to have just conveniently slipped by in production. If you think about it, this board really isn't THAT NEW. It's based off of the K8NSNXP-940 board. I'll put money on it that Gigabyte didn't redesign the board from the ground up to support the socket 939 processors, but just "tweaked" the 940 version to work instead.

Yes, I know over time these bugs will be (or should I say THEY BETTER BE) worked out over time. Beleive it or not, I can cut Gigabyte some slack about the memory issues, IF THEY'D OWN UP TO THE PROBLEM. It's the stonewalling that's really irking me.

By the way.. I tried enabling T2 (as in physically changing it from AUTO to ENABLE in the bios). Same deal. DDR333 for me. I haven't tried recently with my new RAM though. I'll give it a whirl and let you know. I'm not going to hold my breath.

 

[Ballistyx]

Well, T2 didn't give me DDR400 with my TWINX2048-3200 kit and my two sticks of GEiL ValueRAM, but it did do one thing. It made it stable! I'm now running 3GB of physical RAM. Should come in handy someday. hehe I'm running Prime95 in the background right now, and all appears well. I'll fire up Memtest86 later and see if that works too.

It didn't resolve the issue with running the TWINX1024 alongside the TWINX2048 kit I have though. My system will work with the TWINX1024 kit's RAM and the GEiL RAM together, but not with the TwinX2048 kit. I think this board doesn't like the BH6 chips on those DIMMs.

For those who want to know, the TWINX1024-3200LLPT kit is comprised of two XMS 3205V1.1 CMX512-3200LLPT DIMMs. I'm still not thrilled that it's only running at DDR333 speeds, but somehow I find comfort in knowning that I have 3GB of physical RAM to mess around with.

 

[pantner]

ok, the Nvidia SATA thing i know nothing about...im using 2x200GB on the Silicon 3512 controller (very good!)

I have never tried running all four dimms before...thats why i said apparently in my post, (thats just wat i have heard)

i haven't had ANY problems at all with EasyTune4...i used it quite alot about a month ago (when i was trying to overclock, and didn't have much luck), and while it did hang once or twice, it never came up with those erros...

For the temp and asking for you to change your HSF, in my opinion is a good idea....if it does change then its sorta like a calibration error, and the board can be fixed with a BIOS flash.....if it doesn't change, then the temp probe is faulty....

 

[Ballistyx]

Originally posted by: pantner
ok, the Nvidia SATA thing i know nothing about...im using 2x200GB on the Silicon 3512 controller (very good!)

I have never tried running all four dimms before...thats why i said apparently in my post, (thats just wat i have heard)

i haven't had ANY problems at all with EasyTune4...i used it quite alot about a month ago (when i was trying to overclock, and didn't have much luck), and while it did hang once or twice, it never came up with those erros...

For the temp and asking for you to change your HSF, in my opinion is a good idea....if it does change then its sorta like a calibration error, and the board can be fixed with a BIOS flash.....if it doesn't change, then the temp probe is faulty....

I'm running the following for IDE devices:

Two 72GB WD Raptor HD's off of the nVidia PHY SATA Controller
Two 120GB WD Special Edition HD's off of the SilImage controller, using two PATA to SATA adapters. Set up as RAID1 (mirror)
One Lite-on DVD-ROM off of the Primary PATA IDE controller
One Plextor 8X DVD writer off of the Secondary PATA IDE controller

You should have seen me pull my hair out when I couldn't figure out why I could see activity off of my RAID mirror and DVD drives, but not off of the nvidia SATA controller. An e-mail from Gigabyte, and a post from another K8NSNXP-939 owner confirmed that it wasn't a broken board on my part.

What version of EasyTune4 are you running (go to C:\Program Files\Gigabyte\Gigabyte Windows Utility Manager\ET4, and then look at the file properties to find it)? I even tried the latest version downloadable from their site. Both of them give me errors. Do me a favor. Open up EasyTune4 and try to open the voltage/temp monitor (it's the uppermost red circle on the left of the "LCD" readout of the CPU speed). Using the version that came with my 7NNXP, I can go in and play with the HT speed, but I can't bring up the voltage or temp display. The one I downloaded is version B04.072101. The one I currently have installed is v1.0.0.1. I don't think the version numbers are correct though...

I don't beleive that replacing my waterblock for a HSF is a valid proposition for them to suggest. To me, it sounds like a "feel good" tactic that some of us in IT like to have end-users do (which for the record, I NEVER do to the users I help). To me, that answer signifies that they KNOW there's an issue, but they don't want to tell me what they think the problem is. This way, they stall for time while they fix the problem. Stuff like that goes up my butt. Palm, Inc is currently on my "crap" list for doing this to me this week with an Outlook 2k3 sync problem.

Here's a cut n' paste of the e-mail I sent to Gigabyte (along with their reply):

Answer - 95826
Answer : Hi,

1) hdd led is only for ide hdd and will not work for sata hdd

2) ET4 isn't supported by this motherboard.

3) please try a retail HSF and see if the temp. reads at the same temperature.


Thank you

*********************************

Question - 95826
From : xxxxxxxxxxxxxx
Sent : 2004/8/17 03:55
Question : Three questions really.

1) Whenever there is IDE activity on the nForce3 intergrated SATA controller, the IDE activity LED does not light. If I access the hard drives on the intergrated Silicon Image controller, the LED will light up. Is this normal? Is there a way to make the IDE LED show activity on the nVidia SATA controller?

2) EasyTune 4 will not operate correctly. I upgraded from a 7NNXP (which EasyTune 4 worked fine on). When I uninstalled the old version of EasyTune, and then installed the version of EasyTune for the K8NSNXP-939 that came on the support CD, the program shows error as it is starting up about not being able to read certain functions of the motherboard. The program will eventually start, but I cannot check the CPU temperature. Is there a way to correct this?

3) I noticed that the CPU temperature is incorrect by a large margin. Under 100% CPU utilization, in the thermal monitor in the BIOS, my CPU temperature shows as being about 57F / 14C. Although I am watercooling my PC, my room temperature is nowhere near this low. My guess is that the BIOS is mis-reading the thermal diode on the CPU. Is there a way to correct this?

Thank you for your time,


-------------------------------------------------------------------------------
Model Name : K8NSNXP-939
--------------------------
M/B Rev : 1.0
BIOS Ver : F3
Serial No. :
Purchase Dealer :
-------------------------------------------------------------------------------
VGA Brand : ATi Model : 9800 Pro
CPU Brand : AMD Model : Athlon 64 Speed : 3800+
Operation System : Win XP SP : 2
Memory Brand : Corsair Type : DDR
Memory Size : 2GB Speed : DDR400
Power Supply : 550 W

Server grade PSU. Capable of outputting 36A on all power rails.

I'm starting to wonder if I have a beta version of the F3 BIOS on my system. I just might re-flash my BIOS's to see if it allows me to use all four DIMMs at DDR400, and if it corrects the EasyTune issue. By the way Panter, please accept my apology if I sound irate with you. I'm really not upset with YOU. I'm just really irked by all of this. After spending $220 on the board, $700 on the processor, $480 for two 1GB sticks of RAM, and $500 for two HD's, I feel like I got burnt by Gigabyte and AMD. I can go out right now and buy an LGA775 based system that will work with all four DIMM slots occupied at DDR400. Granted, the intergrated controller on the A64 still works plenty fast at DDR333, but it's the point that I paid top dollar for this performance, and it won't run at that speed because of some quirk that neither Gigabyte will fess up to, and AMD gives me a conflicting answer on. I posted this question on AMD's forums, and I was told that it's up to the motherboard manufacturer to adhere to AMD's specs for it to work. Who's right? Who's wrong? Tune in next time as the microprocessor turns! :brokenheart: :laugh:

Here's a cut n' paste about running four DIMMs in the board. I love the brush off I get in the end (reverse chronological order)

Answer - 97782
Answer : Hi,

For more information please contact AMD on the reason for more on the memory is downclock to 333 when there are 4 sticks are installed.

Thank you

*****************************

Question - 97782
From : xxxxxxxxxxxxxx
Sent : 2004/8/21 07:01
Question : Hello,

I am aware that the memory controller is intergrated into the processor.

I have tried using two dimms only (both brands, and also mixing the brands). With only two, it will work in DDR400 mode. When all four DIMM slots are populated, the memory runs at DDR333.

As of this moment, I do not have four sticks of the same brand of memory. Tomorrow I was going to purchase two more sticks of Corsair memory, but I do not beleive that this will resolve the issue.


******************************

Answer - 97382
Answer : Hi,

Please note that the memory controller is integrated within the processor. With 2 sticks does it run properly, or it only applies to 4 sticks of memory?
Also have you tried using all 4 sticks of the same brand instead of 2 different manufacture.

Thank you
Question - 97382
From : xxxxxxxxxxxxxxxxxxx
Sent : 2004/8/20 11:03
Question : I have all four DIMM slots populated on my K8NSNXP-939. When populating either bank, the RAM will run at 200Mhz (DDR400). If I populate all four DIMM slots, the RAM will only run at 166Mhz (DDR333).

If I change the RAM Mhz limit in the BIOS from AUTO to 200Mhz, the motherboard will not POST.

My RAM consists of the following:

2 512MB Corsair TwinX1024 PC3200 DIMM modules.
2 512MB GEIL PC3200 DIMM modules.

It does not matter which DIMM banks the RAM modules are placed into, or if I manually set the timings for the RAM in the BIOS. I can mix one Corsair and one GEIL DIMM in the same bank, and the motherboard will work at DDR400 speed, as long as only TWO DIMMs are inserted into the DIMM sockets at any given time. Having all four DIMM modules populated forces the memory bus into 166Mhz operation.

Is this a bug/limitation in the BIOS, a limitation because I am not running RAM from the same manufacturer in all the DIMM slots, or is this by design?

Thank you,

xxxxxxxxxxxxxxxxxx


-------------------------------------------------------------------------------
Model Name : K8NSNXP-939
--------------------------
M/B Rev : 1.0
BIOS Ver : F3
Serial No. :
Purchase Dealer :
-------------------------------------------------------------------------------
VGA Brand : ATi Model : 9800 Pro
CPU Brand : AMD Model : Athlon64 Speed : 3800+
Operation System : Win XP SP : 2
Memory Brand : Corsair Type : DDR
Memory Size : 2GB Speed : DDR400
Power Supply : 550 W

Unfortunately, I don't have an e-mail record of the conversation I had with an AMD tech support rep on the phone. Her name was Lisa, and she's the one who told me that it's SUPPOSED to run only at DDR333 with all four DIMMs populated.

 

[Ballistyx]

The more and more I see, the more I'm inclined to beleive that the whole DDR400/DDR333 with four DIMM's is something that has to do with the BIOS. I'm thinking Gigabyte goofed big time, and is looking for an answer themselves.

I just bumped into something unexpected a few minutes ago. In the BIOS I bumped up the RAM voltage by .1, because I wanted to try overclocking the CPU clock to 205Mhz. That should bring it to 2.7v correct? Well, it's only going to 2.6v. Correct me if I'm wrong, but don't the majority of DIMM's out there run at 2.65v? If this is the case, then the motherboard is undervolting the RAM, which could be part of the reason for the instabilty some people are experiencing.

This could be easily fixed in a BIOS update too.