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

Sideeffect -

Thank you for the picks. ok now im gunna sound a friggin retard. but what setting do i switch the tester to. meaning what setting on the Dial do i use ? again i just have never had the need to learn about voltage testers thats why i dont know.
and WHEN do i test this. i mean do i test it when its running prime so i know what the voltage is under load ? and again under idle conditions ?

my point about the X2 and NEO2 is that the neo2 is how old? i mena look, it has enough issues already running the winchesters and newcastles. (most were worked out yes)

But the motherboard is old ok. the X2 is new. its not an old technology it s anew chip from the ground up.

the power demands, the memory controllers, the .... everything about it . just look into a new mobo preferably nforce 4 to make this chip run properly.
seriously look at how certain chips run better overclock better and all around are better on let say the DFI boards with the 4-phase power.

am i wrong ?


12.5v into the 12v line messing up the 12v regulators ? oh come on now i know the posssibility is there but please. The chance of that happening when SOOOOO many psu's come out of the box running over 12v is rediculous. the 12v regulators will handle well into 13v i imagine. i mena sure it COULD happen. but ive seen psu's running 12.3v right out of the box. and reports of ones running higher.
i dont have the balls to push the 12v line higher and no need to do so. The reason i made it 12.5v was to make sure i got 12v on everything and not 11.xx i wanted a nice 12v and 12.5 makes everything run at a reported 12.1v

sideeffect thank you again for the pics i grabbed one of the guys who knows this stuff. not sure what his name is here. and he showed me how to test properly

Vcore 1.69 even though everything else says 1.6 EXCEPT corecenter. Corecenter is SET to 1.7v in the little side window but it REPORTS 1.6 in the main screen

are the chips undervolting themselves.. i mean what do i trust ?

so by testing it on the board thats what the cpu is actually getting right. damn damn damn i dont like that.


vdimm is 3.01 how ? i have no idea since its 2.85 in the bios and i dont even hve the 3.3v rail raised.

that vcore really freaks me out. i mean all my overclocking thoughts just went out the window. if i have to run 1.7v to get to 2.65 then im not happy at all.... not at all,


ok wait wait wait i just realised something. my bios is set to 1.4v + 10% which is only 1.54v

how is that possible ? i mean wow im gunna toast this chip lol and my temps sure arent reflecting 1.7v

im scared lol :stick: ... ... ... :explode2:


Edit again

Raising the Bios to what SHOULD be 1.57v vcore -> 1.425 + 10% reselts in PC health core center and cpu-z showing roughly 1.625(roughly) and M.O.M. (Meter On the Motherboard) shows 1.75v
so i seriously gotta ask which do i believe ?
do the venice chips have the ability to actually undervolt themselves ?

Quote:

---

Originally Posted by Tyrvidar

I can not for the life of me figure out how to get the hidden bios up. I've done the whole shift +f3, then alt +f2.

I am running 1.8 offical

---

I'm not surprised that you can't get the hidden options to show up. It's Shift+F2 and Alt+F3. LOL
:toast:

Quote:

---

Originally Posted by Lestat

ok wait wait wait i just realised something. my bios is set to 1.4v + 10% which is only 1.54v

how is that possible ? i mean wow im gunna toast this chip lol and my temps sure arent reflecting 1.7v

im scared lol :stick: ... ... ... :explode2:


Edit again

Raising the Bios to what SHOULD be 1.57v vcore -> 1.425 + 10% reselts in PC health core center and cpu-z showing roughly 1.625(roughly) and M.O.M. (Meter On the Motherboard) shows 1.75v
so i seriously gotta ask which do i believe ?
do the venice chips have the ability to actually undervolt themselves ?

---

I never trust software to report voltages properly. This board has always had "issues" when it comes to reporting voltages and temperatures. No software that I've used has ever reported either correctly. I use core center to get a ballpark reading on the temperature, and I use a multimeter to test CPU voltage. (Just make sure that you're reading from the right place and use a digital multimeter.....the old-school analog ones suck where precision is concerned.)

lol yeah i got a digital one and im using Sideeffects pictures to guide my metering location.

im mean seriously fellas.
How can a person trust a product that has a bios setting of 1.4v +10% and yet you Meter the Mobo and it goes 1.75v. basically if thats the case i might as well kiss this cpu goodbye if im actually running 1.75v.

im thinking of calling either MSI or AMD and talking to their people about this issue.
That really bothers me that th emeter shows 1.75 v and yet the Bios setting is for 1.57 and cpre center reports 1.6v
I mena if i were to be open minded i could simply average the 3 and get 1.65v. that would be fair assumption.

personally i belive im ready to RMA the board. voltage issues such as this, if infact its true that the cpu is getting 1.75 v then this motherboard voltage regulation is severly fubar'd. not to mention the fact that the vdimm shows 3.0v instead of 2.85 which i have it set to in the bios.

The pictures you see here are jsut some playing around im doing trying to see the limits of ram/cpu/mobo
the ram is running 133mhz just to lay that out right now.
the 2 key pictures to look at are the 2nd and 3rd CPU-2726 and cpu-2749
notice the voltage differences. There is more testing to b edone but the difference in voltages are needed because of the cpu miltplier. one is 8 one is 9.
9x multiplier i can not reach near as high as 8x multi. which seems odd to me since the 3000's seem to be the highest overclockers. and they have a 9x multi. correct ?

anywho. those are the voltages i go with regardless of what the meter says. I also go with them as my temps reflect that that appears to be the correct voltages also. If i were infact running 1.75v my temps would be much higher than what i am seeing. which is 30-33 idle and 43-47 load. depending on how warm my room is. those temps reflect roughly 1.65v also.
The first picture is where i am at right now btw.

http://valid.x86-secret.com/show_oc.php?id=28215

355 HTT 133mhz ram 1.65ish lol i never know what im running now lol.

suicide change with clockgen. prime failed and windows was still ok until i opened Trillian... then it rebooted lol.

1.9 final bios is out guys. 1.9 final thread

Some mod bios files are there to. Seems to be the same as the 1.98 beta bios but now usb 2.0 is working again and supposedly the 100; 133; 1 166; 200; 233;250 all work now.

183 and 150 are not available still though. Im going to test it now and see what its like.

SWEET! 1.90 TIME TO TEST FOR 3GHZ!!!! is it better to run speed spectrum on or off?

time to test 3ghz and yet you ask if you should turn spread spectrum off ?

no offense guy but you better stick with a few hundred mhz and not clear to 3 mhz if you dont know if spread spectrum on or off.


and why put 233 and 250mhz in the ram divider this mobo cant even support it. ?

iot only supports 400mhz ram and that kiddies is a hardware restriction. not something that you can just change with a bios add-on. this is a fact of all nforce 3 chipsets.

the 233 and 250 did not work in the beta even though when you posted it showed ddr500 when you got into windows the ram was only running 100mhz. not 200mhz or 225 or 250 so.. why they added it ? who knows. MSI screwed up yet again by not adding the 183 divider... uhg when are they gunna learn.

Sideeffect do you concur with what i just said about the 400mhz limitation of the hardware. ?

Well speed spectrum is sapouse to reduce crosstalk I read, this sounds good for a motherboard running at very high freqency, and maybe out of it's engineered specs.

Also notice how when you bump up the htt a big the option for speed specturm goes disabled? At least it does on my board, so I figured ok msi is doing some thinking here for us.

I never tried disabling it then upping the htt freq. but I just read somewhere here on the fourms someone running it disabled.

I don't think there is a 400mhz limitaion on the ram, the divider is part of the memory controller on the cpu, It's something amd put there before they decided if they were gona go ddr2 or ddr500 for example. Amd does not offically support those dividers, all new cores have those dividers.

the nforce 3 chipset can only run ddr400
DDR2 was out "officially" after the amd64's so AMD had no way of knowing that ddr2 would come around for sure. There was never any plan to allow the amd64's to use ddr2. possibly in the future but not now. and those divider speeds dont work that im aware of becuase again the nforce 3 chipset can only use up to ddr400
Intel 865PE was the last chipset to officially use anything higher than ddr400 that i know of.

If it works super but realise that running your ram divider at DDR500 does not automatically overclock your cpu that simply lets the ram run that high the HTT will still be at 200. If its the same way that Intel handled ram higher than ddr400.
This in turn severely limits your overclocking as you have already knocked off any where from 450-600mhz of overclock cuz your ram is already running at ddr500 but the htt is still only 200.

the 1.98beta really sucked balls as it killed my HTT cap. as my pictures show i was able to reach 355 htt. and run stable with 330. (350 would be stable but i need to lower the cpu multiplier i simply had went to far with the cpu and it crashed)
the beta didnt allow me to go close to 300... so be aware of that and test it someone.
i dont wanna be MSI's guinnea pig AGAIN.

Spread spectrum actually is disabled once you go above 210mhz.
and no it doesnt offer any type of stability when you overclock. only at stock speeds.
its just hte opposite. it makes overclocking worse.
you always leave it disabled.

lestat maybe it depends on the board..... i do not hav such issues. my HTT still max out the same... i can still run mine at 306x9 very stable. although i chose 275 now for 1:1

in my case. the 1.9b8 is jus as good as the 1.8 i was using all the way b4.

and yes i do agree the 233 250 divider is useless... especially for users like me who runs a 3200+ with multiplier 10 max. i need at least 275HTT to max out my CPU.. but tat will gives ram speed more than DDR633... but im not surprise that those divider will work in the final version of 1.9.... as said since the beginning even the 1:1 ratio is actually running off a divider from the cpu... if everything is 'divided'... it makes to difference to divide a larger or smaller amount. and now that higher speed ram is getting cheaper... mobo makers begins to see that this is a nice features to implement now.

are you saying you can run 275 with spreadspectrum ON ? then your a first. lol.
or did you simply mean that 1.98beta or 1.8 regular it still overclocks the same ?

for me its not. the 1.98b bios is really limiting my htt for some reason. and the venice voltage issue is still there.

all i can say is the neo2 isnt designed to run the Venice San diego and X2. its never gunna run them properly and as you may already know the nforce4 boards, primarily the DFI and i believe the aBIT Premium or was it the ASUS Premium ( dont matter to me i hate both mobo makers. i've had nothing but absolute crap mobo's from them) that allow these chips to really run properly. although the X2 still has issues. The sandiego issues where very quickly fixed with temp and 1T mod's.
the venice is a different animal. the voltages this chip uses are whacky and so are the memory dividers.

i didnt wanna upgrade yet but if this chip is gunna run properly i am gunna have to. more money i dont have and a neo2 and 9800 XT i have that i have to try to sell or trade for a nforce4 and pci-e vid card.

anyways.. ill keep tabs on things with this mobo to see if some hacked bios comes around to address the venice issues but i dont think it will happen.
the beloved neo2 has had its time in the spotlight. and still does for older chips. but not these newer ones unfortunately.

New rev.2 BIOS is up based on official release

@Lestat
Interesting vcore results. I used to run my CBBID, winnie 2.6GHZ@ ~1.6 repored by CPU-Z. Now I see performance degregadion. I suppose my mobo used to overvolt my CPU severly. I was at 2.5GHz around last month. Last I got BSOD at 2.48GHZ. Today I got BSOD at 2.2GHZ. Now i run default vcore and im doing 2.4GHZ fine, but can't get 2.6GHZ though :( Neo2 weakened my CPU by overvolting :slapass: I got very pissed today, look what i pumped through winnie
http://www.xtremesystems.org/forums/...id=34482&stc=1
^^^ OWNED That was on thermaltake venus12

I agree the neo2 had a good run, but new technology always wins.

holy crap dude your running a 3200 winnie at 2.6ghz @ 1.9v omg dude that chip is toast ? cascade dont even run 1.9v lol.

my whole issue is with the fact the the setting i use in the bios is one thing.
Core center, and cpu-z says another. and when i use a volt meter on the mobo its another different reading.
and frankly i dont know which to trust .. ? the way the cpu is acting and the way the temps are showing, i think core center and cpu-z are correct.
also the fact that the meter is reporting my vdimm as 3.0v when the bios is set to 2.85... frankly i dont believe its 3.0v simply cuz of the lackluster overclocking with this ram. i know for a fact that with 3v this ram will do more than 3-4-4-10 260mhz. my last kit of this ram did way better.
And lastly the only way for me to actually reach 1.7v is by using core center to raise the vcore. and then i have to set Core center to 1.75v to make it READ 1.7v
setting the bios to 1.55 +10% only gives me about 1.625 in windows and thats wrong. it should be a stright 1.7v

i think personally as i said earlier that i need to rma this board. but the damn S/N is on the botton GGRRR and i cant tear it out just for that.. not right now at least.i hate msi for putting the s/n sticker on the bottom.


also any results yet from the 1.9 testing ?
everything there ? anything missing ? are things as stable and overclockable as 1.8 ?

NFM please do us a favor


make your memory hacks or special setting use something OTHER THAN FULL SCREEN LOGO! PLEASE!!!!!

that full screen logo is pretty dumb by itself let alone when you force us to leave it on so your tweak is on. use something else please please PLEASE!!

Actually memory testing cannot be disabled, I believe it's harcoded into the bios. But when you enable full screen logo it seems that memory testing does not take place. Maybe it does but logo hides it :) lol

No matter what you do memory tweaks always be apllied by default

Quote:

---

Originally Posted by Lestat

NFM please do us a favor


make your memory hacks or special setting use something OTHER THAN FULL SCREEN LOGO! PLEASE!!!!!

that full screen logo is pretty dumb by itself let alone when you force us to leave it on so your tweak is on. use something else please please PLEASE!!

---

do it urself then.....

i hav nothing against wat NFM had done... he has help a lot.. and even if he even force a full screen logo fixed at ON.... no one should even complain a bit..... really....

lestat: if u hav so much to complain about Neo2 ... maybe is time u change out ur mobo ? for problems like we r facing and for 'hero' like NFM... it is all part of the fun in playing with hardwares.... if u r looking for perfection.... maybe u should get a DELL system then......

Quote:

---

Originally Posted by harpyboy

do it urself then.....

i hav nothing against wat NFM had done... he has help a lot.. and even if he even force a full screen logo fixed at ON.... no one should even complain a bit..... really....

lestat: if u hav so much to complain about Neo2 ... maybe is time u change out ur mobo ? for problems like we r facing and for 'hero' like NFM... it is all part of the fun in playing with hardwares.... if u r looking for perfection.... maybe u should get a DELL system then......

---

:lol: :woot:

WOW what a jacked up bios.!!

i cant even overclock any where near where i was .. dang

what is msi doing to these bios's.

blah im going back to 1.8mod.


and yes the full screen logo was rather silly to force that one. do you know anyone who uses it ? i dont. (but that is also the point of using that for the hack cuz no one uses it,, i realise that.) the point of removing the full screen image is cuz it blocks the post screen where there are messages and it shows your ram speed and timings.
now you have to quickly tap TAB. if you catch it in time.

All i said was use another part of the bios for your trick. the full screen logo is something people want OFF not on...

and i said please lol.

So does the new official 1.9 bios still overvolt without ppl knowing?

according to the meter on the motherboard it does.
I'm back on 1.8mod. the 1.9 i couldnt even boot at 250mhz. with 1.8 im back at 300+ HTT. ggrrr msi...

i dont know guys is it really the bios thats overvolting ? or is it a combination of this new cpu and the motherboard itself. ?

Not sure if some of you understood but the full screen logo has nothing to do with the modified memory timings in the bios. They are set by default. The full screen logo just stops the memory testing which as you might notice takes longer on the 1.9 bios than it did on 1.8 bios even when u set quick boot.

I still havnt checked my vcore with a multimeter as my setup would involve me removing the power supply because its vapochill and then i have to worry about the clamshell moving etc and insultation. But i will check reading soon.

The vddr reading is 0.5 above bios setting so if i set 2.75 in bios i get 2.8 volts reading. The reading is very constant. The 1.9 bios seems about the same as 1.8 bios for me as far as memory overclockability goes. I dont need a htt above 275 so high htt doesnt bother me.

San diego is still stable on the neo 2 at 3ghz no problems. Moving to a nforce 4 wouldnt help me get any higher imo.

The 233 and 250 dividers well I dont see a reason why they wont work eventually when msi implements them properly. They dont effect the cpu only the ram and it could be usefull for people wanting to keep the cpu in spec. They are useless for me however and i would have much prefered 183 divider.

well what the devil ?

where what when and how... gawl dang it i do apologise to you fellas and NFM most of all. i absolutely misread the readme. i read memory timings (the extra settings) and not memory test like it says. so yes i was wrong and gladly and humbley admit it.

my vdimm is ehhh roughly .10 to .15 higher than the bios. 2.85 in the bios produces a rock solid 3.0v on the mobo. but i know the ram isnt getting 3v cuz of the lackluster overclocking i am getting with the ram.
again i dont feel this to be a bios issue and sideeffect you can probably concur that its a hardware issue. Unless osmeone can tell me that there is a bios that isnt reporting overvolting when you meter the motherboard.

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)

i actually have to set core center to 1.75v to make it report 1.7v.

i'm back on the 1.8 bios fellas. im not sure what happened with the 1.9 but wow that bios is a total bomb.. i really lost all my overclocking.. i didnt go into to much testing i just set it to what 250mhz should have booted like the 1.8 did and it wouldnt even post. i then reset the bios since you ahve to do that now since MSI took away the "Press Insert Key for Safe Mode" which again i cant believe they did that. total carnage when you make 1 little mistake in the bios and you have to reset the jumper.
i loosened the ram a little and it still wouldnt boot. so no thanks ill stay with 1.8

one comment/question though.

i have seen a few fellas running the 1.41mod bios and having really good luck overclocking and such.
Is there something in the 1.4bios thats different ? i dont have that particular bios and cant find it yet. i think im gunna pm one of the guys and see if he can email it to me. unless one of you gent's have a link for it ?
im willing to try anything but the actual hardware overvolting is really disturbing me. i dont wanna fry this chip or the mobo...