The OFFICIAL MSI K8N Neo2-939 tips and tricks guide thread

[sideeffect]

you can get the 1.41 and also a lot of other bios files from here link

I did try the 1.41 it does boot but i found HTT wasnt as good as with the 1.8 bios. But worth a go It has optimisations for certain ram i think so maybe it works better for some configurations.

Im pretty sure the 1.41 bios shows the correct -12 volt voltage line as well cant remember but 1 of them does.

NFM the 1.9 mod has the temperature issue for me so it wouldnt boot bios setup. It wouldnt boot windows either. Not sure why but its kinda killed that bios chip :P

I have a spare though so its no big deal and ill hotflash the other chip back so no harm. I also have my trusty nforce2 lanparty mobo that can hotflash neo 2 bios files to.

EDIT IMPORTANT - That link also has the 1.52 and x bios dont try that with a venice or san diego it really doesnt work and the bios wont even boot. I think it might be using a earlier romsip than the 1.36 bios and so it doesnt have support for new cpus. Like all the bios files information on changes is sparce. Just dont try it.

[Lestat]

the one thing i saw with the 1.9 when it came to booting is that APIC is disabled and if you had APIC enabled when you loaded windows your windows install will not boot now if APIC is disabled so be sure to check that. i also had to no booting to any media at all for the first few times i tried to boot after flashing but then it went away. :\ no matter what i try to boot do cdrom floppy or hard drive it acts like it was gunna but then nothing happened. (i doubt this is APIC when it comes to a cdrom or floppy or USB related but its possible)
i would find it hard to see how it killed your chip. it also sounds like you have had to do this before ? lol sure your chip isnt bad ?

I also saw some people over at the german MSI forum where murdock gives out his bios say that they can not do 250mhz overclock now either. i replied and said the same thing.

what in the world is MSI doing with these bios's there is no way they are actually TESTING them. I've worked in a test lab and believe me you have to be blind def and umb not to see these issues and even then you would have to be bound and gagged not to notice booting issues.
who do they have coding and testing these ? bring the 1.8 guys back atleast that bios was good.

the whole 1.9 series has been one scary bios lol that not booting issue really had me going for a while that the rom chip or board was toast.


EDIT

hhmm seems like the chaps over at pcper have found out that spread spectrum is still enabled according to cpu-z even if you disable it.
http://forums.pcper.com/showthread.p...47#post3551447

im sure NFM can fix that though. maybe lol.

[sideeffect]

Yes i was being a bit over dramatic when i said dead. It was just ill.

I cant enter 1.9 bios setup files that show the temps in the bios because well ask msi. Something to do with cold temperature maybe.

So if apic was disabled by default that would be why windows wouldnt load with defaults too i guess. I dont have a floppy so if i cant get to windows and use winflash then the bios is stuck the way it is.

Well i hotflashed it back to 1.8 1t mod bios now and going to use that again for the time being.


Edit - there isnt a reading for spread spectrum in cpu-z. I dont think it does stay on when u disable it the guy in that thread was saying it came back on in bios by itself. Doesnt for me.

[Sonic9]

It's interesting to use 233 / 250 divider for testing our ram chips ... without cpu's factor limitations imo ...

[Lestat]

Sonic

what do 233 and 250 make your ram freq ?

250 was 100mhz i believe. i may have been 133 but i was overclocked when i tried it so i dont know.

spread spectrum never came on for me either. it stayed off. what is odd is that once you go over 210mhz (i think its 210 right ? ) its off perminantly you can not change it. so whatever this guy was doing... hhmm ?

[arnemetis]

I have tried nfm's rev2 mod of 1.9 official bios, and i must say it despises my setup. ANY changes, even so much as disabling a sata port or changing just the boot order results in me NOT being able to load windows, ONLY with freshly cmoc reset settings can i boot into windows. I was really looking forward to that 216 divider, but I guess I will settle on 1.8 forever now. by the way, if my system is 100% stable at 250x11, with ram 1:1, but fails superpi/prime after several minutes at 275x10, do you think it just hates the 10x multi or the cpu's memory controller is just garbage? btw 216 divider with 250x11 wouldnt even POST. this is g.skill pc4400 btw, ran 280mhz rock solid on my crappy clocking cbbid. I have since thrown the cbbid back i nto see if the ram somehow degraded, but it still runs sweet at 280mhz. so this cpu is just garbage. any way to incorporate the 216 divider into the 1.8 bios??

[TMM]

Sounds like your CPU just has a crappy memory controller, arnemetis Can you Prime with your memory at 275 on lower multis?

Btw, my CPU won't go over 240HTT on a multi of 9, but does 10x260 - so it could just be like that with your CPU also

[Hoot]

I was the person reporting that CPU-Z said Spread Spectrum was showing still enabled in CPU-Z. That was wrong. For whatever reason, I saw Sideband Addressing and my mind registered Spread Spectrum.

My foray into the NF3 Ultra chipset is just about done. I also believe there are issues with that chipset and the San Diego core, that can not be made right by simple bios changes. It pains me because I have a large chunk of change tied up in my AGP video card, which I love dearly. I'm not entirely convinced that the NF4 chipset is the "manna from heaven" either, but we are being forced to migrate to PCI-E whether we want to/need to, or not. The tail is definitely wagging the dog on this one. I will continue to play with 1.90 and it's modded derivatives. NFM, can you disable temperature shutdown in yours, like you did in 1.9b8? I run my thermal diode pins to an external reader. I tried using MODBIN6 to do this, but it doesn't save my changes. Perhaps an NTFS issue. I'm not sure.

Thanks,

Hoot

[arnemetis]

still, point remains that any changes to the bios resulted in me not being able to load windows, even when not overclocking, just disabling a lan port or something simple such as that. THIS was my primary concern, I had already established my cpu was garbage, was just hoping maybe 1.8 didnt really support san diego, and maybe my cpu didnt like high ht but would run the mem faster with 216 divider, and 1.9 would like it better. as far as I can tell, 1.9 is completely unusable. I will just have to wait on 2.0 Once again, thank you nfm for your hard work.

[Lestat]

Hoot.

i can agree as ive stated and had people shoot me down and some agree that the nforce 3 is out of its league when running a sna diego and even the venice is plagued with memory and voltage issues.
the nforce 4 is a good chip but makers are going into the overclocking realm so hard that they are releasing really crappy bios's that arent fully tested i think.
the DFI boards really do kick but but wow when you have bios after bios after bios that are just garbage or fixes one thing and breaks 2 others... well then... what can one do.
I agree im not ready to go to pci-e as there is no viable performance increase by doing so. But alas. here i sit today pricing parts and selling my neo2 and my ram and video so i can get something in the pci-e flavor.
but remember if we werent speed freaks and tech in our heart of hearts we wouldnt have the need for speed and we probably would still be happy with our machines from 5 years ago...
its that pc lovers double edged sword poking you.

[Garrett]

true that Lestat.

Well I've flashed the damn bios anyway, the light version (mem timings unhidden) and it works fine for me

Had to set it up all over again because of clearing cmos of course, but no pain here... now I can finally run 166 divider and manually alter the trc and trfc timings so that my UTT ram doesn't crap out at this 166 divider :thumbsup:

But on the other side SLI looks kind of nice too LOL I don't know what I want

The newer cards probably will not come in AGP flavor so one day we'll have to go PCI-X... ah well, until then

[sideeffect]

Not shooting you down just disagreeing

I dont see a difference like for example Hoot has his san diego running at 2700 mhz at 1.52 volts. If You read the san diego thread thats about normal on any motherboard. Mine is at 3ghz at 1.6 volts which is also good.

The nforce 4 wont help anymore with cpu speed it might help with ram speeds and htt though but not by a great deal. The limits seem to be the memory controllers on the cpus rather than the motherboard.

If i was dead set on moving away from nforce 3 i wouldnt get nforce 4 because its not any better. The ATI xpress might be worth a look though.

Dont forget that pci express is being upgraded soon anyway and so revision 1 boards could become obsolete.

[Lestat]

the whole argument FOR dfi is its 4-phase power units which helps beyond anything else.
voltage stability bleeds into almost everything else... heat.. overclocking, performance. and just overall stability. the board has proven that already.
plus the nforce 4 does perform better.

in general its not the nforce 4 its the architecture.. the new motherboard layouts the chip design the capacitors and so on that are on the new motherboards.

and pci-e isnt going to be upgraded as soon as you might think.
pci-e was a viable upgrade since AGP had been around since the mid 90's. But to force an upgrad of a component thats 1 year old that would make current video cards and other devices obsolete. I just cant see the hardware designers forcing such an imense change on pc society after thousands and thousands JUST upgraded or are currently upgrading.
I think that if they do so that there could be some legal ramifications that could send some serious shockwaves thru the pc industry.
At the time PCI-E was released they had already designed and were testing the next level. All of this didnt happen just within this last year. they had the architexture designed and testing long before that.
One could argue that the hardware makers knew this and only released this as a reason to boost income, and thus creating havoc in the industry as everyone spent thier money on PCI-E 1 and now they go and shove PCI-E 2 at us ?
there could be legal issues with that if a person got a good lawyer. and one that would be awarded with amount sof moneyt hat no one has seen before.

I for one dont see any reason to not make it backwards compatable and if they have 1/2 a brain they will do it.

i want to upgrade but dont NEED to. i need a upgraded video card as this 9800 XT isnt cutting it for my gaming and certain other needs.
so that is first if i can get a good deal on a 6800 series or x800 series then ill stay with AGP and not look back. but if an overall deal on a good PCI-E mobo and video arrises then i will gladly take that also.

[sideeffect]

Yes they will will for sure make it backwards compatible for the video cards but only new motherboards will support pci express revision 2. Sort of like the move from AGP 4x to 8x.

Pci express revision 1 doesnt perform any better than AGP 8x the only use i see for it is SLI.

I dont see how the 4 phase power solution will help when your voltage rails are already spot on. Yes it is a step forward but its main implications are directed more towards a SLI setup as well.

I just got a great deal on a 6800 GT in the uk. Upgraded from a 9800 at xt speeds. Its so much better for gaming would really recommend it.

[Lestat]

yeah i sold my 6800 GT and my Gskill LA a few months ago so i could get this 2005FPW. was it worth it. i havent decided lol.

4 phase power helps alot. it really does. again the dfi boards compared to a similar board without 4phase shows how good it is. its not the end all omg 4phase. its just that much extra electrical support that the chip and mobo and ram needs.

[corruption]

i'm still trying to figure out why 1.55v +10% doesnt give me 1.7v its only 1.65 or so.(im basically going thru every vcore setting to see what they show and what the meter shows)

If your cpu's default voltage is 1.4v then you WILL NOT BE ABLE TO USE ANYTHING ABOVE 1.45V FOR VID. That's why 1.55v * 10% will give you the same voltage as 1.45v * 10%. I have had no problems with any of the 1.9 beta bios's either. I'm beginning to think that in your case there is either a user "issue" or you have a flakey board. I've added, and removed, the 4th phase to my board and have had no results worth mentioning....neither positive or negative. Are you sure that you're measuring CPU voltage from the right location? (The correct mosfet, and the correct leg on the mosfet.) Are you grounding the negative lead on the case?

I have had positive results from 4 rigs that I've built around the K8N Neo2 Platinum.....none of my builds have had any problems that you seem to be experiencing aside from innacurate readings from core center, mbm, speedfan, or any other "monitoring software." Unless you have a bad board, getting an nForce4 board will not help your overclock. If you're looking for the added features that the nForce4 chipset brought to the table, then that's a different story. I feel badly that you seem to be having so many problems with your motherboard, and I hope that they can be resolved somehow in the near future....good luck to you.

[Lestat]

If your cpu's default voltage is 1.4v then you WILL NOT BE ABLE TO USE ANYTHING ABOVE 1.45V FOR VID. That's why 1.55v * 10% will give you the same voltage as 1.45v * 10%. I have had no problems with any of the 1.9 beta bios's either. I'm beginning to think that in your case there is either a user "issue" or you have a flakey board. I've added, and removed, the 4th phase to my board and have had no results worth mentioning....neither positive or negative. Are you sure that you're measuring CPU voltage from the right location? (The correct mosfet, and the correct leg on the mosfet.) Are you grounding the negative lead on the case?

I have had positive results from 4 rigs that I've built around the K8N Neo2 Platinum.....none of my builds have had any problems that you seem to be experiencing aside from innacurate readings from core center, mbm, speedfan, or any other "monitoring software." Unless you have a bad board, getting an nForce4 board will not help your overclock. If you're looking for the added features that the nForce4 chipset brought to the table, then that's a different story. I feel badly that you seem to be having so many problems with your motherboard, and I hope that they can be resolved somehow in the near future....good luck to you.

wouldnt you suppose that considering that sideeffect gave me pictures on the previous page that im doing the right spot
the wrong leg ? please dude, i leg shows 1.xx and the other shows well... a whole helluva lot more than that. same with vdimm 2.85 in the bios and 3.x on the mobo... the 2nd leg reports over 7volts.
ground ? no i have a wire connected to a 5 foot pole i have shoved in the ground.... yeah guy the thing is grounded properly. if its not grounded your not gunna get a reading with a dig. meter.

issues with a flakey board sure.. but you might wanna go searching around on the results for the Ultra-D AGP version. the EXACT same thing.... ..... massive overvolting on the motherboard when metering it.

im not saying there is anything wrong with this board either stability wise. But this board is severely lacking and has been since the day it was made. There was only 1 build date of this board that came out that has superior strengths to the other boards and its a highly sought after version of the neo2, ill have to check with breaker,, he knows the one cuz he has one.

you speak as if im the only one here having this issue and yet im the only one who has actually reported metering the motherboard. maybe sideeffect did i dont recall.
look around this is a highly known issue

and 1.45 + 10 is 1.6 1.50 + 8.3% = 1.625 the system reports it as 1.58~1.6


what seens to be really funny is the fact that core center can adjust the voltages allt he way from defualt to 1.7 so how does it do that when the bios cant ? or is flakey at best.
well other than direct interaction with the chip at a software level.

well anyways there is no user level errors you can be darned sure of that.


I would dare you to prove that you removed the 4 phase off that motherboard and still had a full functional board. you can't simply remove a phase of electricity on an item like a motherboard and still have a functional board. just like attempting to run a 4phase or 3phase motor on non 4 or 3 phase electricity... just aint happening guys not without changing 1 wire inside the motor. which i suppose could be possible with this board but again im not aware of any mot removing the 4 phase as you would be dumb to do so, the loss of that 4th phase ,,ehh yuck,, why even think about it. your not gunna see it unless your into some heavy overclocking. thats where the 4th phase shows its colors.
unless your referring to something your changing in the bios or a jumper, which the board is 4 phase always ther eis no on off switch for it.
Unless there is and i just havent paid enough attention to those dfi boards to hear about it.
im sure you didnt do this as an experiement so if you care to provide me with some proof or reading materials on how to do this with this board and still make the board functional id find it interesting reading. (yes i mean that seriously i would indeed find it interesting to read about)

again i encourage you to research the neo2's the flakey voltages are a well known issue with the venice's and even the boards themselves... you not knowing this indicates to me you been building systems but not knowing about the hardware your using. or maybe you just deal more with dfi than the neo2.(as i did at one time)

[Lestat]

ok confusion and mass histeria has set in.


in my curiosity of overclocking and hardware stressing i pulled one of my sticks of ram. which i rarely do simply becuase it serves no purpose in overclocking to run 1 stick of ram. What does that prove of your hardware but its pure weakness in not being able to handle 2 sticks of ram, mainly 512meg sticks.

So in my testing of ram this morning i pulled 1 out ... and everything is fine i squeezed a few more mhz out of it as i thought i would but then decided id see how the voltages looked with only 1 stick of ram in.... and this is what i found.
Voltages are tested with M.O.M. (meter on motherboard)

Vcore - 1.68 .03v over what i have set in the bios
vdimm - 2.89 .04v over what i have set in the bios.
Those 2 readings are what i should be reading at any given point in time no matter what.

with 2 sticks of 512 in dual channel.

vcore - 1.70+ when bios is set to 1.6. And when the bios is set to 1.7v it reads over 2.0v!!!
vdimm 3.0v when bios is set to 2.85v

so now the question arises WTF!! now aside from the immediate answer of nothing more than a fooked up motherboard what explanation is there ?
i realise the memory controller is ondie so there for dual channel configuration may cause the system to possibly boost the voltage on the vcore slightly but not the whacko voltages im reading same for vdimm.

anyone have any generalized theories or thoughts ?

(i can not elaborate on 2 256 sticks as i dont have any)

[harpyboy]

hhmm.. tat's pretty weird as many would aledi toasted their CPU with 2.0V...

i saw the pic on page 77 on vcore measurement... how about vdimm ? lestat can u tell me where to test the vdimm ? sorry but i was MIA for some time due to exams.... i will measure the voltages after someone told me where to for the vdimm..... oh ya.. where did u ground ur meter ?

[harpyboy]

oh nvm..i found....

[Lestat]

lol i was gunna say the link for vdimm is right under it.

well it may say v but no way on this iearth the pcu is actually getting 2v. The temps show that they dont go above 33 idle if it was 2v well... the chip would have fried and the heat would def be more. lol. same with my ram.. i would be overclocking the ram higher if it was actually getting 3v.

the regulators are doing some wierd stuff with 2 sticks of ram

i tried running 2 sticks in single channel but oh my guess what.. this crazy board wont let me!! it just beeps at me. GGRRR man o man talk about anger management.
i'll have to set up an rma or something this board is really whacky.

[Jhya]

lol i was gunna say the link for vdimm is right under it.

well it may say v but no way on this iearth the pcu is actually getting 2v. The temps show that they dont go above 33 idle if it was 2v well... the chip would have fried and the heat would def be more. lol. same with my ram.. i would be overclocking the ram higher if it was actually getting 3v.

the regulators are doing some wierd stuff with 2 sticks of ram

i tried running 2 sticks in single channel but oh my guess what.. this crazy board wont let me!! it just beeps at me. GGRRR man o man talk about anger management.
i'll have to set up an rma or something this board is really whacky.

That's true about the voltage. Perhaps the regulators are messed up. For me as long as the motherboard is 100% stable, I dont care what the voltages are. AND the NEO2 is a very stable motherboard if configured correctly!

[Lestat]

omg i jsut saw my typing wow i gotta get help for this i cant stand it anymore. i have to type almost like a snail to make sure my mind and my fingers work together.
my aplogies for the spelling guys. its something i cant controll.


yeah i agree its stable just the voltages freak me out.
but like i mentioned the votlages running above 2v every single piece of software i try reports pretty much the same voltages which are under 1.7 or 1.65 or whatever i set it to.
my heat is the final indicator if voltages to me.. the heat isnt what it would be at those insane voltages and again,,, the chip would have been dead by now at 2v on water cooling.

[Hoot]

hhmm.. tat's pretty weird as many would aledi toasted their CPU with 2.0V...

i saw the pic on page 77 on vcore measurement... how about vdimm ? lestat can u tell me where to test the vdimm ? sorry but i was MIA for some time due to exams.... i will measure the voltages after someone told me where to for the vdimm..... oh ya.. where did u ground ur meter ?

With all due respect, that is not the best place to read the Vcore, though it is convenient. It would be more accurate to read it on the load side of the toroidal inductors (donuts with wire wrapped around them), AKA Chokes. If you look in the picture on Pg 77, below those inductors is the array of electrolytic capacitors that filter the Vcore. You will see one of them is missing. MSI probably saved $.50 omitting that capacitor. Anyway, where it belongs are two plated through barrels from the Vcore and ground planes. Those two pads are the best compromise between accessability and accuracy to take the reading. There is a half-shaded-in circle silk-screened on the motherboard where the cap belongs. The pad in the shaded-in half it positive and the other pad is negative. As for Vdimm measurement, again you want to be as close to the load as possible. Again, MSI omitted a capacitor in that circuit. Up near the top of the board, to the right of the fan header, you will see a missing electrolytic capacitor that filters the Vdimm. Take your Vdimm reading from the two pads where the cap belongs. Again, the shaded-in pad is positive. FWIW, I tried adding a low ESR electrolytic capacitor to each of the two missing ones and it didn't increase stability, so don't bother pondering that one. MSI already figured that out.

Hoot

[4rory]

What about another choke, 2 more fets and the missing caps on the vcore side, is that like a missing phase or something?