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

[Jacke2k]

Wich bios are working best and wich do you think I need for my setup ?

[4rory]

1.8mod rev.3 I would think,

[harpyboy]

i used 1.8 rev.2.. which only pin-fixed the trc and trcf...

someone enlighten me... what actually does it mean by 200 3.9us ? or 200 7.8 ?
im using 3.9us for any divider i choose.... but wat r the technical explaination for those values ???

NFM.. write a tutorial would ya ? for the modding of ram timing... everyone will pray down to u........

[CD 1986]

I think of Tref values as being refresh rates for the RAM. So the lower the number, (eg. 3.9us lower than 7.8us) the faster the capacitors get re-charged. If they are re-charged too quickly, or too slowly, performance and stability will be affected.

Most 512Mb sticks have a default refresh rate of 7.8us, and so when using a 1:1 ratio, using the default refresh rate of 200 7.8us, (shows as 200 3.9us in A64 Tweaker) should lead to greatest stability at higher clocks. If using a divider though, other rates can be used. 133 7.8us (again, 133 3.9us in A64 Tweaker) seems to be good on the Neo2 for a lot of people when using a divider. :)

Guide to tweaking bios:

http://overclockers.com/tips905/

Tictac's Mem Timings source code:

http://www.xtremesystems.org/forums/showpost.php?p=917052&postcount=384

Plus, here is an attachment of the code used by digitalfrost (hope thats ok with him to post it!) to tweak the bios...so you can get a feel for it, and you can use Tictac's source code to add in what you want. :)

[Lestat]

although i havent tried NFM 1.96 beta bios i did try murdocks(and yes i got the 7v on the -12v line also) and have higher 3.3v rails causes the system to refuse to boot to anything.
it will post fine and go into the bios fine but when its tiime to boot to something, nothing happens.

knocking the 3.3v rail back down to 3.3v allowed the bios to boot to media properly.

sounds like some sort of overvolting protectiong that needs to be removed.
i raised my 3.3 rail to 3.55v to test how it helped my ram voltages past 2.85v with the 1.96 bios i cant.


im sticking with 180r3 until a better one is released with more options and allows us to raise the vdimm past 2.85.(i believe although cant confirm the hardware lock is 3.0v just like the ASUS 754 boards did, they lock the vdimm at like 2.9v i believe and unless you hardware mod it whatever you put into the bios dont matter. at least past 2.9v i mean.

[Lestat]

Link please. :)

http://www.msi-forum.de/thread.php?threadid=17206&threadview=0&hilight=&hilightuser=0&page=1

ive posted in that forum to him reporting issues before but no one responded to me and frnakly the crap mod's he is putting out now i could care less.
TicTac and NFM are doing a better job.

Murdock's last bios have some horrible issues with USB i lost nearly all my USB devices using his last 1.8mod but going back to NFM's or the murdock biuos prior was ok,,,

avoid the 195 and 196 bios's unless you have a sandiego or venice... and even then ... you need a new board if your running a X2 anywyas this NEO2 isnt gunna cut it, and if you think it will your lost.
same with the venice frankly this board has had its time in the spotlight

[nfm]

I never modded any new v1.9x beta bioses.
@Lestat, sorry to hear problems with 3.3v rail :( btw what's your cpu, you probably can't clock more than 255 HTT because on-die mem controller craps out. It happens when you try to pass certaib HTT speed, you bump volts, go crazy loosening timings, 2T etc. and nothin, that's problem with mem controller. I'm just begginner in bios mooding so I hope i will get a hell lot better in modding bioses :up: and about the tref: 200MHz@15.6us is good, default 200Mhz@7.8us is good also, 166Mhz@7.8us (166Mhz@1.95*us in A64 Tweaker) seems best at least to me.

[nfm]

somethin for desert :p: http://www.theinquirer.net/?article=24353 orginal http://forum.msi.com.tw/index.php?topic=82787.0 time to get hands on msi :hrhr: , show some support to our brothers

EDIT: I got my hands on MSI, already send buch of emails about Neo2 :clap:

[sideeffect]

Just thought id say about the modified bios files im hosting.

need.for.mhz, CD1986 or the website owner take no responsiblity for any damage that the modified bios files may cause to your system. The bios files have been tested and showed to be working fine but we cannot guarantee your success.

Flashing a bios is risky and can cause errors and if your not able to except the consequences dont mess with your pc.
:rolleyes:

[jeandobre]

undefined
I've been reading this column for awhile. I have a 3000 1.8 and have been playing with the settings in Core Center. I can't figure out how to increase the multiplier past nine, I see several people running higher. Can someone tell me or point me in the right direction to achieve this? I have the latest bios from MSI, but, I'm missing something and am asking for your expertise.

Jeandobre
Noob Boob

AMD 3000 939
MSI Plat running at 232 x 9 (runs around 2087)
1 GB Corsair Value Mem (2x512)
2 IDE hard drives

[sideeffect]

The cpus are locked for higher than default multipliers. Its worked out by 200 x the number to make the cpu speed so 200 x 9 = 1800mhz which is the clock of the 3000+.

So with a 3000+ you can use any of the lower multipliers but none higher than 9. The people that use higher multipliers have higher rated cpus like a 3200 has 10x 3500 11 and 3800 has 12 x 200 to make their operating frequencies.

San diegos use different performance ratings than venice so a 3700 like mine has a 11 times multi same as the 3500 venice.

[nfm]

@sideeffect

Tmod is on our ftp, grab it :) and then you can remove from there if you want

[4rory]

I remeber reading some where that you can get a 10 multi on a +3000 buy using 9.5 multi, it was some kind of bug or something, Anyone read something like this befor?

[sideeffect]

@sideeffect

Tmod is on our ftp, grab it :) and then you can remove from there if you want

Thanks nfm, cd1986 remembered he did send u that after all :) we were thinking there was a version u could use actually in windows xp and not dos. But ill plug the floppy in and use dos.

But 1 stick of my ballistix has failed. Think it has been flakey for sometime now so im sending that back and getting some gskill instead. Sick of messing with average ram.

[sideeffect]

I remeber reading some where that you can get a 10 multi on a +3000 buy using 9.5 multi, it was some kind of bug or something, Anyone read something like this befor?

Give it a try jeandobre and report back. Wonder if it will work or not i have doupts :p if it does work maybe it would work with all athlon 64s.

[4rory]

i've tried it didnt work with my 3200 but i read someone say "theres no point in getting a 3200 if your just getting it for the multi, just get a 3000 and use 9.5 multi it will up it to 10" ... thats all i remeber. dont remeber what the tread was about.

[CD 1986]

CD1986 remembered he did send u that after all
What a dummy eh! ;) All sideeffect's fault really! lol :p:

[jeandobre]

thanks for the info. It's the chip, not the board.

[jeandobre]

the multiplier goes up to nine only, I have read that if you put the multiplier up to .5 that it will default itself up, however, that's not an option I have.

I was hoping to bump my speed up, it seems I'm pretty much maxed out around the 230's. I guess I shouldn't complain, though, at least everything runs.

Thanks for your help.

Jeandobre
Noob Boob

[4rory]

try higher then 230 try to skips a few sometimes it gets hung up on a few bad mhz. Worth a shot.

[alexgontijo]

Hi guys,

I took a new CPU a few days ago, 3200+ Venice 0519, it got rock solid till 2.7 GHz and 1.575v.

2.750 I tested with 1.6v but was kind of unstable.

So I wonder: there's anything I can try to get it like 2.8 GHz?? Is it safe to run it on 1.65v daily? Worth trying it? How far Venice vcore can go safe?

Thanks a lot any answer,

AleX

[harpyboy]

Hi guys,

I took a new CPU a few days ago, 3200+ Venice 0519, it got rock solid till 2.7 GHz and 1.575v.

2.750 I tested with 1.6v but was kind of unstable.

So I wonder: there's anything I can try to get it like 2.8 GHz?? Is it safe to run it on 1.65v daily? Worth trying it? How far Venice vcore can go safe?

Thanks a lot any answer,

AleX

im running my at 1.6-1.62V at 2.75Ghz now.. 24/7 machine. but im on WC. temp is about 33C idle and 44C load...

i say 1.7V for temporary benchmark... not for daily use... tat's ofcoz if u can get more juice out of 1.7V..... anything more will be dangerous even tho it isn't heated up

[Lestat]

NFM -

by YOUR 196 bios i simply meant the one thats on the website hosting your other bios's

im trying to make a deal on a venice as we speak but with newegg showing $190 for a retail Venice 3200+ im about ready to break out the credit card lol.

[sideeffect]

NFM -

by YOUR 196 bios i simply meant the one thats on the website hosting your other bios's

im trying to make a deal on a venice as we speak but with newegg showing $190 for a retail Venice 3200+ im about ready to break out the credit card lol.

Thats not NFM's bios its cd1986's bios. Its not in nfm's section its in cd1986's section and the readme says its made by cd1986 and not nfm.

[alexgontijo]

im running my at 1.6-1.62V at 2.75Ghz now.. 24/7 machine. but im on WC. temp is about 33C idle and 44C load...

i say 1.7V for temporary benchmark... not for daily use... tat's ofcoz if u can get more juice out of 1.7V..... anything more will be dangerous even tho it isn't heated up

Thanks Harpy,

I got it stable at 2.750 (275x10) exactly as yours, with 1.62v. But as it got hotter, I don't think the only 50 MHz worth the 0.5v raise. I'll use it in 2.7 GHz! My iddle is at 40C and full load at 55C.

I'm taking an OCZ 3200 Platinum Rev2 to run it on 1:1, 270x10, I think it will get even better. My Corsair 3200XL is running at 166 divisor, in 225 MHz 2-3-3-6, and doing 6650 in the SANDRA. My SuperPI is at 32 seconds. How can I see the following digits on it? (Like 32.150s..)

I'll post results latter with the OCZ.

AleX

[sideeffect]

I added the super pi versions to the website for you alexgontijo. There are 3 version there, default, milisecond version and sse3 version. Have fun :)

superpi (http://www.swjka.com/k8nneo2bios)

[Jacke2k]

My crapy cpu dose 2500mhz whit 1.53v and nothing more. Dumping in more volt dosent help.

[brschmid]

well, got some Corsair 4400c25

guess what i can do with this ram? boot up at 250x10 1:1 1.4vcore 2.7vdimm 2.5-3-3-8 timings.
so that must mean my OCZ ram was bad.

so far this ram is working great!

[Sonic9]

I wait your stats with OCZ alexgontijo ! beacause I have Corsair 3200XL too and I can't run it over 260mhz rock stable 2.5-4-3-6 (nfm bios tweaks) ...

Goodbye.

[alexgontijo]

I added the super pi versions to the website for you alexgontijo. There are 3 version there, default, milisecond version and sse3 version. Have fun :)
superpi (http://www.swjka.com/k8nneo2bios)

Thanks a lot!!

I shall have the OCZ with me in the next days, I'll post the results for sure!

Today I tested my CPU at 300x9 and the Corsair at 245 MHz 2.5-3-3-6. It's doing fine, 7K in Sandra, a 5% more than running at 225 MHz 2-3-3-6. I recommend it!!

AleX

[kasius]

Who was say official bios would be in end of June?? We are in July, and I don´t see any bios :slapass:

[sideeffect]

well there is now a 1.97 beta bios

1.97 beta (http://216.239.39.104/translate_c?hl=en&u=http://www.msi-forum.de/thread.php%3Fthreadid%3D17656&prev=/search%3Fq%3D1.8mod%2Bk8n%2Bneo%2B2%2Bbios%26hl%3D en%26lr%3D%26sa%3DG)

[raydelee]

so many new bios which is good for UTT 2-2-2-5 at 250fsb
thx

[4rory]

What does the new 1.97 bios have that the others didnt?

[harpyboy]

i want to know too... anyone tested it out...


damn.... my RMA replacement raptor came to me DOA !!! gossshhh.. another few days on ubuntu livecd..... im starting to miss my OCing life... *sobbb

[corruption]

What does the new 1.97 bios have that the others didnt?

I haven't noticed anything except some ram timing changes from 1.96......nothing noticeable :shrug:

[4rory]

Dam that must suck, Hope the next ones fine, Just wanted to ask if anymore modded bioses are coming out, maybe the one with the 5.5ns hack that you turn on with logo setting in bios? I kinda need this since its all i need to be stable :D

[floodo1]

hey guys i have som basic mushkin ram. i bought some back in nov of last year, and then i bought some last week. they want to run at diff timings....2.5-3-3-7 (old stick) and the other wants to run 3-4-4-8 (new stick) by SPD. the trip is that if i set the timings to 3-4-4-8 only the new rams works. the old ram wont run at the slower timings. but if i set the timings to 2.5-3-3-7 then they both work. this makes the only condition where the old stick will work is 2.5-3-3-7 or auto. new stick seems to work at those setting or by its spd.

so anyway if i set timings to auto (or i set them to the 2.5-3-3-7) no matter what combination i use both sticks at i get ddr333/166mhz ram speed. if i use each stick individually its fine. doesnt matter if i use dual channel or single channel, if they are both in there its ddr333

any suggestions?

[harpyboy]

i hope the new BIOS makes it in time for the 3800+ X2 :)

[kasius]

have any new modded bios by Murdok VID + 18% in bios and it is work??

thanks

[alexgontijo]

any suggestions?


Try to set new vdimm values, try like 2.6v, 2.65v, 2.7v and 2.75v. Use the software: Memtestx86 to verify if the memory is error-free. Download it, make a boot disk and run it like 4 loops. If it's error-free ok, if you got errors, try the next voltage option.

And set the 2.5-3-3-7 manually, don't use SPD. If you got no boot, remember to clear the CMOS - no panic. :)

If you get only errors, I suggest take a new pair of memory, like the dual-channel ready ones.

Let us know,

AleX

[nfm]

Try v1.98 official BIOS people, official at least on korean msi :p: . Link in my sig

EDIT: Modded 19b8 BIOS. Need someone to test it with bios savior or extra chip. Don't risk it!

[jcollett69]

hey guys i have som basic mushkin ram. i bought some back in nov of last year, and then i bought some last week. they want to run at diff timings....2.5-3-3-7 (old stick) and the other wants to run 3-4-4-8 (new stick) by SPD. the trip is that if i set the timings to 3-4-4-8 only the new rams works. the old ram wont run at the slower timings. but if i set the timings to 2.5-3-3-7 then they both work. this makes the only condition where the old stick will work is 2.5-3-3-7 or auto. new stick seems to work at those setting or by its spd.

so anyway if i set timings to auto (or i set them to the 2.5-3-3-7) no matter what combination i use both sticks at i get ddr333/166mhz ram speed. if i use each stick individually its fine. doesnt matter if i use dual channel or single channel, if they are both in there its ddr333

any suggestions?

A few thing about RAM. Most "manufacturers" do not make the chips themselves. Rather, the purchase the chips and PCB on the open market, test them to certain timings and speed, and sell them as their own. Any particular line of RAM could have different types of chips in them depending what was available on the market at the time. Sounds line you may have RAM that was produced from Winbond dies, the older set most likely being BH and the newer most likely CH. BH is well known for not booting into CAS 3 but will do 2.5 or 2.0 with proper voltage. Get that vDIMM up into the 3.2 to 3.3 range and you may find that your RAM can do 2-2-2-5 at fairly high speeds (~250MHz). I know the MSI board will not do that voltage. A vDIMM hack or OCZ DDR booster would correct that issue for you.
As far as having four double-sided DIMMs populating your motherboard's RAM slots...this will always cause your BIOS to revert to DDR333 speeds. Search these boards for much of this knowledge being repeated over and over ad nauseum. OK, I'm done here.

Hoping for that new BIOS :shrug:

[sideeffect]

Try v1.98 official BIOS people, official at least on korean msi :p: . Link in my sig

EDIT: Modded 19b8 BIOS. Need someone to test it with bios savior or extra chip. Don't risk it!

Ok mate im gonna test it for you now.

[Garrett]

Try v1.98 official BIOS people, official at least on korean msi :p: . Link in my sig

EDIT: Modded 19b8 BIOS. Need someone to test it with bios savior or extra chip. Don't risk it!I downloaded the official bios from the korea site too... http://www.msi-korea.co.kr/program/support/support_dld_detail.php?part=1&serials=66&uid=K8N%20Neo2%20%ED%94%8C%EB%9E%98%ED%8B%B0%EB%84 %98&name=%EC%86%8C%EC%BC%93%20939%20(AMD%20K8)&chip_id=51

but you modded it?

What did you mod anyway? :D

[CD 1986]

What did you mod anyway? :D

Changelog Rev.1 for 1.98 modded bios:

- Read Write Queue Bypass = 16x
- Bypass Max = 7x
- Idle Cycle Limit = 256 clks
- Max Async Latency = 8ns
- Write Recovery Time (twr) to 3clks
- All hidden options are now available

They are some memory tweaks that help with stability when OCing. Thanks for the latest bios NFM, i'll try it out once sideeffect has tested it. :)

[sideeffect]

Sorry im taking so long the bios does work the timings do set but im having the cant enter bios issue still. Its nothing to do with the NFM bios its a flaw in the 1.9's that effects some people.

CD1986 is quickly modding a no temperature version for me so i will be able to enter bios setup then. But the bios flashed fine and the modded timings are ok.

[Garrett]

This bios shows the trc and trfc values so you can modify them at your own will, right?

Like this I mean:
http://home.megapass.co.kr/~rhgkswhd/bios1.jpg

[sideeffect]

yes

[CD 1986]

Works well NFM, all timings read correctly in A64 Tweaker! Any chance you might be able to enable 183 divider? Thanks :)

[sideeffect]

Ok the bios works fine now ill add the cd1986 no temp mod bios to the ftp for anyone else who has the issue.

Just had a little mess around with it the 216 divider now works. Couldnt test the 250 as my main ram is being rma'd and im using pc2700 ram atm. The 3.5 Ht multiplier options has gone, it didnt work in the previous betas but it was there, i guess they couldnt get it working.

Nfms timings set but nfm have you tried to add the 183 divider into the bios? does it have the options in bios. Also twtr (Write to read delay time) sets to 1 with 166 divider still. Please force it to 2 in rev 2 :)

Good job with the bios though :clap:

[nfm]

Rev2 will be out shortly. 183 divider is there but it is not activated. It's like with extra 13%, 15%, 18.3% VID option, they are there but thre's nothing you can do about it :( . Yeah I will remove temps from BIOS so people don't be pissed if something does not work.

[nfm]

Ok, Rev.2 is up. Testing needed.
Changelog Rev.2:

- twtr = 2 clks
- All temps hidden, to see them hold "Shift + F2, hold it still and with your
second hand press "ALT + F3"

[CD 1986]

Suggestion for Rev3 NFM, change what is loaded when you press "load optimal defaults" to some more OC friendly settings. I've found it a great way to save some time once you've cleared the CMOS. Some of the default settings used from the factory are no good for OCing, and we have to keep changing them everytime we reset the bios! :rolleyes:

[sideeffect]

yes good idea cd1986 im all for that. :slap:

[sideeffect]

Some ideas for defaults

Ati/nvidia speed up = disabled
boot up logo = disabled
small logo = disabled
S.M.A.R.T = Enabled

Tested the bios and it works. Flashed ok the temperatures are hidden and i cant unhide them with alt shift thing. i see the little flash u get when it supposed to show hidden options but they dont appear.

Twtr still defaults to 1 in 166 divider also. The rest of the timings set ok though

maybe you set read to write instead of write to read :slapass: :p:

; Write to read delay(dth)
twtr_data equ 0FFFFFFFEh ; Write to read delay (1bit)
twtr_1 equ 000000001h ; 1 clock
twtr_2 equ 000000002h ; 2 clock

[nfm]

yeah i probably messed up with that one, so many values :confused: :slapass:

Ati/nvidia speed up = disabled
boot up logo = disabled
small logo = disabled
S.M.A.R.T = Enabled

will do that

also i will do:

spread spectrum = disabled
cool n quiet = disabled
agp fast write = disabled

EDIT: Rev.3 is up

[sideeffect]

twtr still defaults to 1 in dividers. Maybe it doesnt work or maybe the code is wrong. Oh well not a big problem really.

The bios is really good the default settings work nicely and i can enter bios setup which is always a bonus :) I cant get the temperatures to show but that might be my problom alone as i have the issue.

Someone else try and unhide the temperatures.

Good work NFM going to use this bios from now on.

[nfm]

Well i tried 1.98 and i have to say it is still in beta phase. Don't know why did they post this bios on korean msi. I'm switching back to 1.8.

[harpyboy]

Well i tried 1.98 and i have to say it is still in beta phase. Don't know why did they post this bios on korean msi. I'm switching back to 1.8.


how bad is it ?

[sideeffect]

how bad is it ?

Its not bad its just not official. Somethings dont work like after more testing the 216 233 and 250 dividers dont work. The northbridge fan control isnt included. There are no .5 cpu multipliers.

The bios seems pretty much stable now though and my overclock seems ok with it as with the 1.8.

I dont think there is much point upgrading to the bios unless u have a san diego or a x2 processor.

[Jhya]

Its not bad its just not official. Somethings dont work like after more testing the 216 233 and 250 dividers dont work. The northbridge fan control isnt included. There are no .5 cpu multipliers.

The bios seems pretty much stable now though and my overclock seems ok with it as with the 1.8.

I dont think there is much point upgrading to the bios unless u have a san diego or a x2 processor.


I lose my USB 2.0 capability. It only has USB 1.1???

[sideeffect]

Yes usb 2.0 support isnt added yet so not i good idea to try this bios if you need usb 2 :D

[4rory]

But can I get the same overclock i'm getting on 1.8mod cuz befor 2.9 wasnt stable on 1.95, I need the memory options my ram is almost stable and it would be a killer 24/7 setup, 2.9ghz with 7500+mb/s :banana: , but I need better memory options to finish it off.

[Garrett]

But can I get the same overclock i'm getting on 1.8mod cuz befor 2.9 wasnt stable on 1.95, I need the memory options my ram is almost stable and it would be a killer 24/7 setup, 2.9ghz with 7500+mb/s :banana: , but I need better memory options to finish it off.What you could do is that you let A64 tweaker automatically load the timings every time windows starts...
(I do it this way, always start up with the '100' divider and let A64 tweaker do the rest)

[4rory]

lol that causes instand crash, I can't even change more then one setting in a64 tweaker without instant freeze.

[Lestat]

A64 tweaker is a windows killer waiting to happen.

for you to have someone start ram at 100mhz and then A64 tweaker INSTANTLY kicks the ram in tha arse and forces it to make in instant jump up to insane overclocked speads WHILE in windows is foolish to say the least.

many people wont be able to do this as it will simply lock their system.
i dont recommend anyone try this unless they have 100% stability with their overclock via the bios and booting into windows.

you could potential damage your systemand corrupt windows beyond a point of repair if the system crashed.


I NEEEEEED someone to test something for me

1.9 bios any version most importantly the supposed actualy bios from the MSI site and also NFM's modded version.
You MUST have adjustable rails on your PSU in order to try this.

1)
adjust your 12v+ rail above 12v maybe 12.5v
2)
adjust your 3.3+v rail past 3.3v 3.4 3.5 3.6

now boot the system.
Are you having the same issue i am having ith the 1.9 bios that the system will NOT boot to any media. no matter what you do it posts fine it goes thru the NVRaid detection fine and whatever else you may have, and then it might read the floppy like its gunna start but nothing happens.
Same for a CD like the XP boot cd. it says "press any key to boot from cdrom" then either nothing happens OR it says "checking system" like the xp cd does when you boot to it but again nothing happens.
same thing with the Hard Drive.
It will flash the HD light but nothing happens.
You can press F8 and choose safe mode but nothing happens.
If you have linux it MAY boot to the linux kernel argument options screen but nothing will happen after that.

the only way to fix this is to reduce all voltages down to damn near the exact voltage they are said to be,, ie; 12v 3.3v 5v

even after doing that i am having SEVERE issue trying to boot to a floippy to reflash the bios. roughly 1 out of 20 reboots will then boot ot a floppy where i can reflash with the 1.8 bios.

Also,

can anyone else chime in on stability or other issues with the 1.9 bios. ?
and has anyone of you modding gent's even TRIED to allow the bios to go above 2.85v on the vdimm ?
the hardwarelock on these boards is what ? 3.0v if i remember correctly and frankly gents... I NEED 3.0V this OCZ EL Gold Rev2 PC4000 cant even go past 260mhz STABLE ( and i enphasise stable) without more than 2.85v
this is good ram and will reach 1:1 ratios of 290mhz if you get a good kit.
My last kit did 311mhz STABLE on a 9/10 divider but i had to have a minimum of 3.0v to do it.

also the 183mhz mem divider. we all know thats simply the 1/2 cpu multiplier setting. thats why its in a reserved state.
Simply remove all 1/2 cpu multi's and any instance or mention of it in the bios and then remove the reserved setting in the ram divider section and one would think it would allow the ram to run at the 183mhz divider
yes ? or has MSI made it more complicated than that and its more of a hardware issue than a bios issue.

[mapel110]

http://www.swjka.com/neo2/view_thread.php?fid=3&tid=13
with this Bios, i have only 3000 MB/s Readspeed @ everest. With 1.8 Bios i have about 6000 MB/s. WTF?

[harpyboy]

hoohhhhh... tat's pretty weird... anyone has this issue ?

seems like dual channel is gone

[floodo1]

A few thing about RAM. Most "manufacturers" do not make the chips themselves. Rather, the purchase the chips and PCB on the open market, test them to certain timings and speed, and sell them as their own. Any particular line of RAM could have different types of chips in them depending what was available on the market at the time. Sounds line you may have RAM that was produced from Winbond dies, the older set most likely being BH and the newer most likely CH. BH is well known for not booting into CAS 3 but will do 2.5 or 2.0 with proper voltage. Get that vDIMM up into the 3.2 to 3.3 range and you may find that your RAM can do 2-2-2-5 at fairly high speeds (~250MHz). I know the MSI board will not do that voltage. A vDIMM hack or OCZ DDR booster would correct that issue for you.
As far as having four double-sided DIMMs populating your motherboard's RAM slots...this will always cause your BIOS to revert to DDR333 speeds. Search these boards for much of this knowledge being repeated over and over ad nauseum. OK, I'm done here.

Hoping for that new BIOS :shrug:

ok first up, i have 2 double sided sticks (2x512).
second up the chips appear to be one revision different. i dont recall the first letter but the two are: xb and xc for the chips. not sure what letter replaces the "x" but the 2nd letter for one is "b" and the second letter for the other is "c".

so idk. they are def diff revisions just dont understand why 2x doublesided is dropping it to 166? esp since each individually does 200. and it seems wierd that it happens in all combinations of using the 2 sticks together.

i have set the timings manually as i mentioned in my original post. basically the ONLY problem is the 166 drop from 200. timings wise i can run the faster timings with both sticks in at the same time no prob. so yeah why is it dropping back to 166?

[mapel110]

look at this. :O

/edit
Wahhhhhhhh, PC1600. What the F

[Lestat]

Go into A64 Tweaker and see what Max Async Latency says. and Tref
Controller mode also. And clock frequency
128bit should be dual channel and 64bit single channel (i believe thats correct, the single channel part i mean)
The first 3 things i just mentioned if they arent set right will just absolutely kill your ram thruput.
basically if you have a screen shot of what your settings were before. go into it now and compare them to see whats going on.
also CPU-Z see if its showing the proper settings. CPU-Z will tell you if its Dual or Single channel on the Memory tab.
and yes it does sound like Dual channel is gone.
but also set the ram chips like every other mobo in the world.
Slots 0 and 2 dual channel and slots 0 and 1 and 2 and 3 are single channel.
in other words first slot and 3rd slots. just for the heck of it try it. see what happens.
again reflash the bios again with the same rom file and see what happens.
im telling you guys this 1.9 bios is FUX'd BIG TIME!!
whom ever is programming it at MSI is a bunch of blooming retards.
Also does this same thing happen with the actual 1.9 bios from the korean site ? or IS that the korean site bios ?

[ColdCollisions]

I think you've got your wires crossed?
1&2 or 3+4 = Dual channel.

Msi (http://www.msi.com.tw/program/products/mainboard/mbd/pro_mbd_detail.php?UID=607)

[Vega]

Can someone please!!...tell me the best bios for my setup??
I have problem to run my memorys in 250fsb (2-2-2-8) 1:1.....I now they handle 258fsb (1:1) 3.3V on a DFI board!? :confused:

Now I use v1.8.

Thanks..

[mapel110]

I think you've got your wires crossed?
1&2 or 3+4 = Dual channel.

Msi (http://www.msi.com.tw/program/products/mainboard/mbd/pro_mbd_detail.php?UID=607)
dual channel is activ, but my RAM is adjusted as pc1600. I have switched in the Bios to Auto Mode and now I have 5600 MB/s (PC3200 speed again). :)

But Speed was better with default 1.8 Bios and same timings. :(

[ColdCollisions]

@mapel110 ...sorry mate I was actually talking about this post quoted below, wrongly stating which RAM Slots were Dual Channel. Highlighted in red.



Go into A64 Tweaker and see what Max Async Latency says. and Tref
Controller mode also. And clock frequency
128bit should be dual channel and 64bit single channel (i believe thats correct, the single channel part i mean)
The first 3 things i just mentioned if they arent set right will just absolutely kill your ram thruput.
basically if you have a screen shot of what your settings were before. go into it now and compare them to see whats going on.
also CPU-Z see if its showing the proper settings. CPU-Z will tell you if its Dual or Single channel on the Memory tab.
and yes it does sound like Dual channel is gone.
but also set the ram chips like every other mobo in the world.
Slots 0 and 2 dual channel and slots 0 and 1 and 2 and 3 are single channel.
in other words first slot and 3rd slots. just for the heck of it try it. see what happens.
again reflash the bios again with the same rom file and see what happens.
im telling you guys this 1.9 bios is FUX'd BIG TIME!!
whom ever is programming it at MSI is a bunch of blooming retards.
Also does this same thing happen with the actual 1.9 bios from the korean site ? or IS that the korean site bios ?

[sideeffect]

Can someone please!!...tell me the best bios for my setup??
I have problem to run my memorys in 250fsb (2-2-2-8) 1:1.....I now they handle 258fsb (1:1) 3.3V on a DFI board!? :confused:

Now I use v1.8.

Thanks..

1.8 is probably your best bet. Try NFM's rev 3 1.8 bios. And just because one board can run memory at a speed doesnt mean much nowadays esspecially when dfi relax timings to allow for high fsb doesnt mean the memory is running faster.

If it was the 754 DFI board it was single channel also and not dual channel like the msi.

Good luck anyway getting stable but the 1.9 isnt really ready yet to be as good as the 1.8.

[harpyboy]

yup... or 1.36mod if u need the extra vcore.... but 1.8 works so damn nice for me .....

do u guys think there will be a 500:400 divider for the new BIOS for the rev.E ??



alrite guys... my SECOND RMA RAPTOR WD HDD came DOA again !!! ... well not really DOA... but right after 100% format.... dead....... i called WD complainted a hell lot and they gonna send me another by tomoro and prepaid shipping for the last hdd back..... looking at the 3 dead raptor on my desk... how sad can it goes....... :'(

[Lestat]

@mapel110 ...sorry mate I was actually talking about this post quoted below, wrongly stating which RAM Slots were Dual Channel. Highlighted in red.


please read what i said coldcollision

i said if MSI would make their boards like everyone else in the world were slots 0/2 and 1/3 are dual channel
i did NOT say that msi boards are slots 0/2 cuz they are 0/1 for dual channel.

[ColdCollisions]

I did read what you said, that's why I posted?
:rolleyes: GO check it yourself:fact: . Now this is well off topic so I'm not going to post anymore.

[Tyrvidar]

I'm just now treading into 64 territory. Been overclocking XPs since they came out. So, I've bought a K8N neo2, Venice 3000+, and 2x512mb OCZ EL Platinum Revision 2. Im new to HTT and doing heavy overclocking on the FSB and ram.

I have read the guide Xtreme provides to overclocking. I have also done more digging on the potential of the K8N Neo2. I would like to squeeze about 2.4-2.6 ghz out of this.

Suggestions? My system should be here in a day or 2. What kind of settings should I test out. Help a amd64 o/c newb out.

[Gen0c1de]

Hey guys, i am new to the forum. I heard there is a cool for sale section, but you have to have 100 posts. So I am gona start here I guess.

See you in the future.

[harpyboy]

welcome tyrvidar...

anything we tell u now is pointless... u can figure everything out when u actually touch ur new rig :) ... and then u will realise there r a lot of issues..... tat is when u come here and ask again.... :D

[Tyrvidar]

haha I know man... Must you destroy my hopes ahhh.

[Lestat]

I did read what you said, that's why I posted?
:rolleyes: GO check it yourself:fact: . Now this is well off topic so I'm not going to post anymore.


once again i say it with more conviction.

read what i said and know what i said. cuz you didnt read it properly. i did NOT say 0/2 were dual channel i SAID 0/2 SHOULD be dual channel like every other board in the world but MSI has to be different and make things complicated by making slots 0/1 and 2/3 dual channel.

0123 call it whatever you want slots 1 2 3 4 or a b c d cuz technically they are called slot 0 1 2 3 and have been since god knows when.


Anyways has anyone seen what i mentoined above with the voltage issues ?
and why on earth does the -12v report at only 7v ? this really needs to be addressed cuz i have a feeling this too is causing issues even though the hardware may see the true -12v if the bios is reporting only -7v then it can and will halt the system from properly functioning.

i wont use 1.9 any further that bios scares the hell out of me. it took me nearly all of one evening to get the dang thing to even boot to a floppy when i tried the modded version. i thought for sure the mobo was dead.
Im still wondering if one of the modders can chime in on whether raising the vdimm voltage is even possible or if they have even tried it.

same thing with the 183mhz ram divider.
remove all 1/2 cpu multipliers and that should enable or allow you to enable the 183mhz. or simply removing the *reserved* value from that particular hex line in the bios might do it. but i have a feeling more digging in the never ending hex values for the root of that *reserved* value is needed to fully enable the 183mhz. i mean the 1/2 cpu multi is just running 183mhz ram. but since its catagorized as a cpu option is it possible that the system pukes on certain cpu's instead of it running properly when using the actual setting under a ram setting ?
im really grasping at straws there but hey sometimes the right question drives a programmer to think and look at things differently which can result in finding new and interesting things or tricks.

[floodo1]

ok first up, i have 2 double sided sticks (2x512).
second up the chips appear to be one revision different. i dont recall the first letter but the two are: xb and xc for the chips. not sure what letter replaces the "x" but the 2nd letter for one is "b" and the second letter for the other is "c".

so idk. they are def diff revisions just dont understand why 2x doublesided is dropping it to 166? esp since each individually does 200. and it seems wierd that it happens in all combinations of using the 2 sticks together.

i have set the timings manually as i mentioned in my original post. basically the ONLY problem is the 166 drop from 200. timings wise i can run the faster timings with both sticks in at the same time no prob. so yeah why is it dropping back to 166?


no help? sorry for such a basic question but im just running this board basic style until xmas when i can finally afford some real ram :(

[sideeffect]

Lestat all the .5 cpu mutipliers have been removed from the 1.9 bios but the 183 divider option is still not there. It cant be modded in by NFM hes already looked at it. So until MSI decide to be nice to us and add it themselves i think were screwed.

Cant remember which bios it was but 1 of them showed me a accurate -12 volt line. Think it was 1.41 mod. All the others with the option added show the wrong value :p annoying. But the issue that really annoys me is the cpu temperature in the bios it reads 268 degrees and hardlocks the bios when i try to view it. It even stops the 1.9 bios from entering setup. Thats why it was modded by nfm and cd1986 to remove the temp readings from the bios.

Once in windows it shows the temperature with core center or speedfan/mbm without a problem.

Ive been using 1.98 beta bios since it came out without a problem. Its never crashed for me and the extra configurability is worth it for me. But i dont see much point in people who arent using san diegos or x2 cpus switching from 1.8 atm.

[STEvil]

Slots 0 and 2 dual channel and slots 0 and 1 and 2 and 3 are single channel.
in other words first slot and 3rd slots. just for the heck of it try it. see what happens.

actually that doesnt come out to say a whole lot.. I think you left out a sentance?


ok first up, i have 2 double sided sticks (2x512).
second up the chips appear to be one revision different. i dont recall the first letter but the two are: xb and xc for the chips. not sure what letter replaces the "x" but the 2nd letter for one is "b" and the second letter for the other is "c".

so idk. they are def diff revisions just dont understand why 2x doublesided is dropping it to 166? esp since each individually does 200. and it seems wierd that it happens in all combinations of using the 2 sticks together.

i have set the timings manually as i mentioned in my original post. basically the ONLY problem is the 166 drop from 200. timings wise i can run the faster timings with both sticks in at the same time no prob. so yeah why is it dropping back to 166?

xb and xc? Its kind of hard to answer since we dont know the manufacturer (hyinx, samsung, winbond, etc) and you havent given us any other info off the chips to indicate model numbers or speed ratings. Do you have the ram in the two slots closest the CPU or two farthest from it (dual channel on the Neo2) or are they in 1 and 3 or 2 and 4?

[harpyboy]

guys im back in business.... hdd came back working fine ;) hohohoho

who else except sideeffect had tried the 1.98 BIOS ??


also, i manage to get hold of OCZ EL plat rev2.. the one with TCC5.. but it seems that it max out around 275Mhz 2.5-4-4-8... almost exactly the same as my Patriot XBLK TCCD.... so... does the problem lies with the Venice mem controller and the Neo2 BIOS ?? or it just coincident that both of these ram max out around there ?

[harpyboy]

i had tried the 1.98 NFM modded... very nice.... is a keeper... thx NFM again

for some reasons it allows my OCZ to clock up to 285Mhz.. alittle higher.... but still not good ennough for me...

[STEvil]

its likely a combination of both.

I have a 0444 3000+ winchester and the max I can do is 285 2.5-4-3-x-1T or 310 2.5-4-3-x-2T with 2.8v on my 2x512 OCZ PC4800 Platinum (TCC5), but this also requires a modded bios since the original didnt hold trc and trfc constant when using dividers or at different FSB's.

[harpyboy]

its likely a combination of both.

I have a 0444 3000+ winchester and the max I can do is 285 2.5-4-3-x-1T or 310 2.5-4-3-x-2T with 2.8v on my 2x512 OCZ PC4800 Platinum (TCC5), but this also requires a modded bios since the original didnt hold trc and trfc constant when using dividers or at different FSB's.

2T really hurts on benchmarks... but i really like to hear from you whether it does have any real life effects ... like on games and daily apps ?

im still playing around with it... hope i can get more out

[Tyrvidar]

I have everything hooked up ~ time to start testing, 210 so far and going up.

Edit- Blahh im such a 64 o/cing noob. Blue screens at 215.

Updated to official 1.8 bios.

Venice 3000+ OCZ plat Rev 2 3200... where do I start at haha. I turned HTT to 4x.

[4rory]

use htt 3x, and lossen you ram to 2.5-3-3-7, and 166 divider then push the cpu to test for max bootable speed, then test for max stable cpu speed, then report back and we will give you the next steps.

[Tyrvidar]

Update~ 255 and climbing. Thanks, will update in a minute will see how far this goes before I have to up the voltage on the cpu. U = roxor. I had to pop my 64 cherry eventualy this is good stuff.

Update~ maxs out at 270 what should I do from here, up the cpu voltage like any old XP O/c? I want to squeeze about 2.6 out of this puppy.

[STEvil]

2T really hurts on benchmarks... but i really like to hear from you whether it does have any real life effects ... like on games and daily apps ?

im still playing around with it... hope i can get more out

There are a lot of 1T/2T 1gb/2gb threads in the memory section of the forums.

[uniacid]

Hey guys I'm back, although I still don't have my new cpu yet since AMD is out of stock of 3700 sd's and so I have a question, they offered me a 4000 but it would be either a clawhammer or some other core and not a sd :/

so what do you guys think? maybe I should get the 4000, sell it and just buy another 3700 sd or wait till they get more sd's in stock

[corruption]

who else except sideeffect had tried the 1.98 BIOS ??
I've been running the 1.9B8 bios since it was first released. No problems at all so far. I've had a setback though....my TDX CPU block decided to puke on and drown my 9800pro on Saturday night. I've tried to get it working again, but it's fubar.....soooo, I have a new BBA x800xl in my board now. :banana: :woot: As soon as I get new gaskets for the TDX I'll resume with my OC tweaking.

[Tyrvidar]

Seem to be getting alot of static in my headphones from the onboard audio. I miss my soundstorm on my abit board nforce2 already.

[Lestat]

anyone recommend the best bios for Venice's ? My 3200 arrives this afternoon so i want to flash the board and be ready for it.

any particular bios shown to perform better than others ?

i figured NFM 1.8Rev3 would be good but didnt know for sure.

[Lestat]

ok im running 1.8rev3 and for some reason the voltages are whack.

My 3200 Venice is locked at 1.45v i've raised it to 1.5 and it doesnt go to 1.5 it is reported as 1.45 by every app i can find in windows.
even changing it with clockgen doesnt work.

anyone seen this before ? i cant tell if its a bios issue or a venice vcore lock.

LBBE 0519 APAW 3200+

[arnemetis]

the board undervolts...usually up to .1v for me, its .1 less than whatever you set, so i set 1.55+10% to get 1.62v.

[sideeffect]

ok im running 1.8rev3 and for some reason the voltages are whack.

My 3200 Venice is locked at 1.45v i've raised it to 1.5 and it doesnt go to 1.5 it is reported as 1.45 by every app i can find in windows.
even changing it with clockgen doesnt work.

anyone seen this before ? i cant tell if its a bios issue or a venice vcore lock.

LBBE 0519 APAW 3200+


Yes thats normal 1.45 is the maximum the vid will allow for venice san diego chips. To raise the voltage anymore u use the pecentages. Maximum setting is 1.45 + 10 percent which = 1.63 volts or so.

Try 1.45 + 5 percent that makes about 1.52 volts and is a good place to overclock from on air cooling.

[floodo1]

xb and xc? Its kind of hard to answer since we dont know the manufacturer (hyinx, samsung, winbond, etc) and you havent given us any other info off the chips to indicate model numbers or speed ratings. Do you have the ram in the two slots closest the CPU or two farthest from it (dual channel on the Neo2) or are they in 1 and 3 or 2 and 4?[/QUOTE]

ok i'll take the ram out later tonite and give you all the info i can from it.

doesnt matter if i run it in single or dual channel. atm i have it on the 2 slots closest to proc. but i can run any combination of using both sticks and it drops back to 166.

let me write everythign down and get back....thanks for the help so far :)

[arnemetis]

Ok just got done trying the 1.98 bios modded by nfm.doesnt like my mobo/cpu at all, I want to make it clear I am not bashing nfm with the following at all. currently on 1.8 rev3 by nfm, with 1t fix. 250x11 perfectly stable, ram 1:1. 275x10 with ram 1:1=no good, this was what I was trying to fix. now results with 1.98. Basically at stock(with default craptastic timings), it would post, but not load windows. Had to remove my pci ata controller card to get into windows. 250x11 = post but no boot, even without pci card. 275x10 = no post, 1.8 would get into windows, just fail superpi and prime after a while. so my experience with this bios is that its garbage, I await the final, and may even revert to 1.8 if these problems persist, I despise getting down on the floor and changing the cmos jumper.

[4rory]

@Tyrvidar ok so 270*9 is 2430mhz ok, so bump the vcore depending on ur cooling to about 1.55 to 1.65 for the best of air cooling xp-94s and such.

Keep your memory very low and laxed to keep it out of the question, and the htt mulit to keep it at 3.

A fast way to get the max cpu mhz i find is to set cpu to 2300 lets say and 1.65 volts prime look at heat, judge if its to much heat, anything hihger then 55c is to high, then determine ur max vcore then keep bumping up fsb with clockgen and doing superpi 1 mill till it doesnt go anyhigher then back off 10fsb and start priming. :D then once u get the max primeable, start to lower vcore and prime :D.

[N00B]

Did any one tried some mushkin redline 3500?
With a OCZ DDR BOOSTER?

TKS!

[Tyrvidar]

I appreciate your Help 4rory alot. I will hold off on testing this to the limits and in the meantime will let it burn in. Will look into one of the xp94s and such over the next month. Stock cooler can only do so much hehe.

Thanks again ;). Some noob o/cer will see this thread and use your post you helped me with.

[angra]

I recently upgraded my CPU to a manchester (X2 4200+), and no BIOS I've found yet lets me boot all the way into windows with both the CPU and my SATA drives being recognized properly.

1.8 : dual CPU is not recognized

all 1.9 betas: drive on SATA port #3 works fine, but drive on port #4 shows up as "new hardware", and windows wants to install a driver for it(?!), and the partitions aren't available.

anyone else run into a similar problem ever?

[sideeffect]

I recently upgraded my CPU to a manchester (X2 4200+), and no BIOS I've found yet lets me boot all the way into windows with both the CPU and my SATA drives being recognized properly.

1.8 : dual CPU is not recognized

all 1.9 betas: drive on SATA port #3 works fine, but drive on port #4 shows up as "new hardware", and windows wants to install a driver for it(?!), and the partitions aren't available.

anyone else run into a similar problem ever?

have you tried using xp?

[angra]

have you tried using xp?

Not yet - I'd really rather not rebuild my OS right now; I have a ton of apps installed, etc., and it's only been six months or so since my last rebuild.

Both SATA ports seem to work fine under 1.8, so I strongly suspect a BIOS issue of some sort. It's just odd.

though I do recognize that XP is a lot more likely to have up to date support and attention paid to it, so maybe I will end up going that way.

fortunately, the drive in question is used only for nightly backups so I can live without it for a while at least.

next test will be to switch 3&4 and see if the problem stays with the drive or the port.

[sideeffect]

Msi has been working on the sata ports trying to sort out the locking issue so its possible that the bios has messed it up :)

From what ive read though the 1.98 beta bios locks port 1 and 2 so pehaps its worth testing it out and see if you can boot that way but probably not best to plug it in at 280 htt. :stick:

[angra]

Yeah I noticed comments along those lines in this threat, I think that's worth a try, too, and I will.

Since the drive itself is purely a backup drive, I can afford to take more risks than normal in the data-corruption area.

[4rory]

@Tyrvidar
Glad I could help.

[4rory]

With the 1.36mod bios I can get volts up to +18.3% with venice my max effective vid is 1.450v correct? does this mean that if I apply +18.3% I get 1.71vcore? and then the board undervolts also? but 0.1volt right? that means max I can get is 1.61vcore this seems a bit low, I just want to know if I should try this bios and try the high volts to see my max overclock, I'm on air and am willing to apply a bit over 1.65 for a nice screeny, already got a 3131mhz on a +3200. That was with 1.8mod bios at 10% i think cpu-z reports close to 1.65 volts.

[sideeffect]

No the vcore reading is accurate. So 1.45 + 18 percent will give you about 1.7 volts. The chip doesnt undervolt it just wont accept a vid higher than 1.45. The percentages take it higher than that.

The newcastles would do 1.55 + 18 percent as their maximum because they had higher voltage defaults to begin with.

The vid is controlled by the cpu not the motherboard.

[4rory]

Ok so my venice with all voltage settings maxed on 1.36mod bios will give me a safeish testing vcore? 1.450+ 18.3% will be ok for venice?

[harpyboy]

Ok so my venice with all voltage settings maxed on 1.36mod bios will give me a safeish testing vcore? 1.450+ 18.3% will be ok for venice?

jus for benching.... i wouldn't recomend running it 24/7 with it...

1.8 BIOS allows u to run 1.7v as well... using corecenter

[Lestat]

this whole NEO2 + Venice voltage issue is a joke

MSI needs to fix this.

the voltages are off from what you have them set in the bios usuall .25v
plus most of the settings dont work
i had the bios set to 1.7v and it was only like 1.6 in windows.
or rather Core center said it was 1.7 but reported it as 1.6
totally screwed up

this needs to be addressed and quick before people start frying chips.

i can never tell what im setting the voltaged too unless i reboot and go back into pc health. and then when i get into windows sometimes its even different.
I cant even get 1.475 to 1.575v its a huge dead zone up to 1.45 its ok and from 1.6 and up its ok. thats a lovely hole in the voltage spectrum.

i am almost done with this board. i think this venice will be my last amd64 for a long time unless i win the lottery.

i do believe im gunna have to get me an Ultra-d msi just cant made this mobo with too many issues and now the venice's voltage issue;s yeah...

The only bios i have not tried is the 1.36mod becuase its too old for this venice and i dont feel it would be wise going back to such an old bios.

[STEvil]

considering doing a droop mod to my board..

[corruption]

I cant even get 1.475 to 1.575v its a huge dead zone up to 1.45 its ok and from 1.6 and up its ok. thats a lovely hole in the voltage spectrum.

I can see what you're saying about a "voltage hole," but I haven't had any issues at all yet. I know that if you set the VID to anything higher than 1.45v with a Winchester or a Venice, it will only see 1.45v. (i.e. 1.55v VID = 1.45v VID) The 1.37mod or 1.36mod bios, the one that supports 18% over-voltage will give you finer steps than any other bios, but I really don't see that much of a point unless you want over 1.6v for constant use. :shrug:

If you don't have time to calculate what each VID and %overvolt is, here's a chart for you. ;) Just find the voltage that you want to try and use the selected VID and %....

1.325v * 3.3% = 1.3687v
1.325v * 5% = 1.3912v
1.35v * 3.3% = 1.3945v

1.35v * 5% = 1.4175v
1.375v * 3.3% = 1.4203v
1.325v * 8.3% = 1.4349v
1.375v * 5% = 1.4437v

1.4v * 3.3% = 1.4462v
1.325v * 10% = 1.4575v
1.35v * 8.3% = 1.4620v

1.4v * 5% = 1.47v
1.425v * 3.3% = 1.4720v
1.35v * 10% = 1.4850v

1.425v * 5% = 1.4962v
1.45v * 3.3% = 1.4978v
1.375v * 8.3% = 1.4891v

1.375v * 10% = 1.5125v
1.4v * 8.3% = 1.5162v
1.45v * 5% = 1.5225v

1.4v * 10% = 1.54v
1.425v * 8.3% = 1.5432v
1.425v * 10% = 1.5675v

1.45v * 8.3% = 1.5703v
1.45v * 10% = 1.595v


I have it up to 18.3% on a sheet of loose-leaf, but I've spent too long typing this up as it is....:P And I know that the board won't be as precise with voltage as my calculations are...it will however give you an idea concerning which direction the voltage will want to swing. (i.e. 1.57v will want to swing up and 1.53 will want to drop a little.)

Hope this helps somewhat..... :rolleyes:

(I tried to line it up nicely, but it's getting too late to even worry....5:15am, no sleep as of yet...)

[sideeffect]

Hi guys its not msi's fault that the winchester venice and san diego cpus wont go above 1.45 vid. Any setting above 1.45 vid will be ignored by the cpu.

Its because newcastle had a stock voltage higher than the 0.9 nm cpus. Neo 2 was built before 0.9nm cpus came out and they need to leave the settings up to 1.55 for the newcastle cpus.

The vcore is about right. 0.25 difference isnt a lot if thats true all motherboards are never spot on because they use cheap sensors.

There is a thread about venice cpus and voltage and they said lots have been frying with 1.75 volts. They didnt recommend more then 1.65 for 24 / 7 usage
and even then keep the chip cool.

I only use 1.65 with my vapochill.

[sideeffect]

just bought a 6800 gt :banana: :banana: :banana: :banana: :banana:


sale on at www.overclockers.co.uk

Leadtek WinFast GeForce 6800 GT 256MB = 165 pounds inc vat

[Lestat]

http://msi.designlab.info/thread.php?threadid=17656

Murdock has some mod's on the 1.97 and 1.98 bios. these are new. they also have a new version of AWD Flash
Murdock also says that they are a new bios layout and such. basically he says its a new code or new sequence layout of the bios.
a guy actually spoke english in there and they replied in eglish.. must be nice i have posted in english and no one replies. ggrr
his issue does not seem to be related to the bios but more so a mod he did to the temp sensor.

has anyone tried either of those 1.97 or 1.98 bios ?


Sideeffect -

i wont run mine above 1.65 i mean be realistic the default voltages are only 1.38 or 1.4 thats a HUGE voltage jump.
mine is water cooled and runs idle 28-32c(depending if im running stock volts or more) and during Prime95 45c MAX.
that is even running 1.65v

the thing is i cant get the bios to give me an accurate setting. basically i have that hole in the vdimm from 1.45 all the way thru 1.6

and .25v different i didnt mean .25 i actually meant .025 but after during more experimenting i found that th ehigher vcore you use in the bios the more the vcore is wrong in corecenter and cpu-z

basically corecenter will report 1.7v for what you have it set,. but its real reading or realtime reading says 1.6
so there is a difference of .025 all the way to .1v thats a major screw up. and its the bios that can control this even though yes many people report this issue but how mnay report this issue with the DFI boards ? or asus or gigabyte ?

And Clockgen will not set the vcore or the multiplier. its like it doesnt work with the venice or something.

Corecenter will adjust the vcore properly BUT dont tell me or anyone else to use it simply cuz thats silly the bios should be handling the voltage settings porperly.

I told people when the venice came out to go buy a NEW board . this damn NEO2 has enough issues running 2 year old cpu's let alone a brand new model.
I guess i should have followed my own advice.
damn pci-express i aint upgrading to something that doesnt give me more performance. not to mention the insane cost facotr of doing so.
but thats the only way im going to get a NEW mobo that ould support this venice properly i suppose. unless of course ALL mobo's are having these voltage issue's.

[sideeffect]

Lestat are you using the most recent version of corecenter because corecenter, speedfan, systool, mbm, cpuz and the bios all show the same vcore for me. There is a slight fluctuation to the reading but that will be the sensor.

As has been discusssed lots before the sensors on motherboards are cheap and show variations that dont exist when you use a digital multimeter.

[sideeffect]

The fix for entering bios setup was fixed by me and cd1986 before murdok also. The latest Nfm bios has the fix. Funny that he ignored my post when i first posted the issue weeks ago. :p:

[sideeffect]

I just read your pictures in msi user forum. Strange i dont get those issues at all maybe its the venice ?

Anything over 1.45 doesnt have an effect maximum setting for my san diego is 1.45 +10 percent which gives me 1.63 volts.

The setting is different in corecenter than the reading but thats because of the newcastle venice thing. Thats why newcastles can use 1.7 volts instead of 1.6 volts with the same bios.

Also maybe its the bios i havnt used murdoks version try using nfm's 1.98 mod and see if its the same.

[4rory]

@ Curroption lol you should of listed the voltage settings from about 1.4 to the max we don't really need to know about vid settings under 1.4vcore.

Also has anyone noticed that under load the vcore goes up? quite a bit also, clockgen says something about startup vid... and max vid... seems like under load vid jumps up another 50mvs.

[corruption]

@ Curroption lol you should of listed the voltage settings from about 1.4 to the max we don't really need to know about vid settings under 1.4vcore.

Also has anyone noticed that under load the vcore goes up? quite a bit also, clockgen says something about startup vid... and max vid... seems like under load vid jumps up another 50mvs.

It was really late and I was a zombie when I listed the settings. The lower VID settings can be useful when using over 10% overvolts.

I've noticed that vcore jumps a little bit when the cpu is under load. I don't trust Core Center or any other software for reporting voltage. If you want to be sure what voltage you're putting through your core, then use a digital multimeter....that's what I do.

[angra]

All the 1.9 beta and mod BIOSes seem to be giving me really bizarre temp readings. I've seen some mention here and there of this here and there, but i'm not sure if people are having these issues specifically with the 1.9+. Could be a CPU issue? Dunno, I only get it on overclocking attempts - but I haven't done a lot of non-OC attempts yet either, to isolate.

I get negative temps, and sometimes crazy high (like 115C+). water lines seem warm but not hot, and the chip runs when I get these readings. don't know what to make of this. any ideas?

here is a screen

[4rory]

1.9 still in bios, no other bios offically supports x2, I'm sure if your contact is fine that ur perfectly safe just a software issue,

@Corruption where do I test with dmm on the boad for vcore?

[sideeffect]

All the 1.9 beta and mod BIOSes seem to be giving me really bizarre temp readings. I've seen some mention here and there of this here and there, but i'm not sure if people are having these issues specifically with the 1.9+. Could be a CPU issue? Dunno, I only get it on overclocking attempts - but I haven't done a lot of non-OC attempts yet either, to isolate.

I get negative temps, and sometimes crazy high (like 115C+). water lines seem warm but not hot, and the chip runs when I get these readings. don't know what to make of this. any ideas?

here is a screen

weird. You get these readings with other programs as well like mbm or speedfan ?

[angra]

haven't tried mbm or speedfan yet. Will do so if it would be helpful for diagnostic.

everest home and corecenter report the same readings.

I am becoming tempted to try and embed a thermistor between my cpu and waterblock to see if I can tell what the actual temps are, but I am not so sure it's a good idea to try something like that - afraid uneven pressure might risk damage to CPU...

[4rory]

The 2 pins on the cpu that transmit core temps are on the very edge of the cpu, it's pretty easy to attach wires to them, put the wires in the socket and install cpu, but you have to use very thin wires, like the ones on thermistor, could try cutting off the sensor and using the wire. This way you get core reading, maybe it's the sensor on the core thats bad, then you could probably rma it.

[sideeffect]

But you say its only with 1.9 version bios's? 1,8 gives constant temp readings?

There is definitely some sort of issue with the sensors and the 1.9 bios as i cant get any readings in the bios at all i have to use a modded bios to remove the options from the bios or the bios setup wont even load.

All the windows based temperature monitors give me constant results though.

Wizzard has just added a neo 2 profile to systool which gives the most accurate results from the sensors try and use that see what happens. systool (http://www.techpowerup.com/systool/)

[angra]

here it is under 1.8.

I am not 100% positive it doesnt happen under 1.8 yet. I had a strange unstable boot when I rebooted after flashing back to 1.8. I will run @ stock speeds and 1.8 for a while to see if I get any odd behavior.

screen under 1.8 (http://gtf.org/dan/1.8.png)

will try systools shortly.

the attachment is speedfan, running under 1.9b8 (leftover attachment from aborted post earlier)

[angra]

well it did it under 1.8 (http://gtf.org/dan/hot-1.8.png)

I guess all signs point to bad CPU at this point, huh? I'll try swapping my venice back in to see if if that clears things up.

[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

[bobbobson]

Hye guys, im purching one of these in two days and a 3000+ vience. Ill report back with my problems!

- Michael.

[uniacid]

hey guys I have a question, since I rma'd my cpu to AMD I was wondering how a 4000+ clawhammer would be (since they're out of stock of any sd's) also what about a 4000+ San Diego? I heard the lower multipliers were locked?

I'm thinking of just getting the 4000+ clawhammer and either using it for awhile or selling it and just buying another 3700SD

[sideeffect]

The lower multis arent locked on the 4000 sandy because ive seen posts with people using lower multipliers.

I wouldnt personally bother with a clawhammer because it wont clock as well unless your using a cascade or something:p

cant you get a refund and buy a 3700+ sandy else where?

[uniacid]

AMD doesn't do refunds, also I was wondering about a 4400+ dualcore, how well have those overclocked?

what I really think I'll do is just sell the 4000+ clawhammer from AMD since it'd be new in box and just buy a 3700SD from newegg again

[sideeffect]

well it did it under 1.8 (http://gtf.org/dan/hot-1.8.png)

I guess all signs point to bad CPU at this point, huh? I'll try swapping my venice back in to see if if that clears things up.


All your temperatures are messed up though not just the cpu. Id suspect the motherboard sensor rather than the cpu. I think it could be that the bios messes up the sensor rather than breaks it as i have had issues sometimes when i flash a bios.

To cure it i reflash using winflash clear cmos take out power lead for at least 30 seconds boot into windows and flash again if it hasnt cured it.

[sideeffect]

AMD doesn't do refunds, also I was wondering about a 4400+ dualcore, how well have those overclocked?

what I really think I'll do is just sell the 4000+ clawhammer from AMD since it'd be new in box and just buy a 3700SD from newegg again

The dual cores seem to clock really well. Similar to the san diegos in clock rate. If you can afford it.

Ask for 2 x 3200 venice cpus find the best clocker and sell the other maybe :p:

[Lestat]

I am astonished the you fellas are even trying an X2 or San diego with this motherboard. the new mobo's are the only ones that are gunna work.
This board is never gunna work properly with the Venice san diego or X2. no bios can fix the issues as they are hardware related not software. A bios can only mask them or work around them to a certain degree.

that being said i have a simply question.

Clockgen allows you to run the 183mhz memory which is the 9/10 divider that DFI uses.
We all know that that value in the bios is labelled as reserved. so why if clockgen allows it hasnt anyone actually enabled it via a bios mod ? I've seen the bios hex. and it says *reserved* yet the ability to make it 183 is there.
now somehting funny is that the 1.98beta mod bios(im not sure if the unmodded beta has it) actualy has 213 233 and 250 mhz in the dram freq, lol thats funny cuz its worthless. The mobo doesnt support preset ram over DDR400. yes you can raise the freq via HTT but not force the ram to run at 250mhz if its ddr500.
mine is ddr500 and i set it to 250mhz. POST says DDR500 but once you get into windows and or memtest its 100mhz. actually anything over 200 defaults the memory back to 100mhz.

I hate the way this venice is treating the voltages im setting in the bios i never know what its set to unless i set it reboot and go into PC Health and look there.

im not brave enough just yet to go attaching sensor wires onto the chip lol but does anyone have a picture of this ? id like to atleast see it.

also same for checking vcore and vdimm with a meter. im not much for using a meter as i never have had the need although i do have one. can anyone provide a link to a "how to" for doing this ?

i sure hope there is a way via a bios mod to address the venice issues i really cant afford to upgrade my mobo at this time. :\

[angra]

I am astonished the you fellas are even trying an X2 or San diego with this motherboard. the new mobo's are the only ones that are gunna work.


which board are you using/would you use with a X2, Lestat?

I don't know of too many mobo that have come out since the X2 did. I can see what you are saying, but it's not immediately obvious what mobos "should" work with x2 and which shouldn't.

my neo2 has been great with the venice.

[sideeffect]

I am astonished the you fellas are even trying an X2 or San diego with this motherboard. the new mobo's are the only ones that are gunna work.
This board is never gunna work properly with the Venice san diego or X2. no bios can fix the issues as they are hardware related not software. A bios can only mask them or work around them to a certain degree.

I have a san diego running at 3ghz on my neo 2. Pc is prime stable pc never crashes in what way does it not work?

Lestat maybe you messed up your vcore regulators when you were pumping 12.5 volts into the 12 volt line.

http://www.swjka.com/k8nneo2bios//vcore.jpg

http://www.swjka.com/k8nneo2bios//Vddr.jpg

[4rory]

Just tested vcore with dmm, wow I was tense lol 1.6 set in corecenter, 1.450 +5% in bios cpu-z shows 1.524 idle to 1.544 Load dmm shows 1.560 idle. Whats safe for venice again on air 1.65? 1.75 kills?

[urbanfox]

You guys know the earliest revision bios that will run venice?

[harpyboy]

You guys know the earliest revision bios that will run venice?

i had tried all the way back 1.36 it works.... any earlier i dunno

[Lestat]

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

[Lestat]

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 ?

[corruption]

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:

[corruption]

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.)

[Lestat]

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.

[Lestat]

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.

[sideeffect]

1.9 final bios is out guys. 1.9 final thread (http://216.239.39.104/translate_c?hl=en&u=http://www.msi-forum.de/thread.php%3Fthreadid%3D17878%26sid%3D8206bc6b3dca 99223f9dfec7e7d3cb51&prev=/search%3Fq%3D1.8mod%2Bk8n%2Bneo%2B2%2Bbios%26hl%3D en%26lr%3D%26sa%3DG)

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.

[4rory]

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

[Lestat]

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. ?

[4rory]

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.

[Lestat]

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.

[harpyboy]

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.

[Lestat]

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.

[nfm]

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/attachment.php?attachmentid=34482&stc=1
^^^ OWNED That was on thermaltake venus12

[4rory]

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

[Lestat]

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 ?

[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!!

[nfm]

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

[harpyboy]

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......

[FlyingHamster]

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:

[Lestat]

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.

[P_1]

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

[Lestat]

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. ?

[sideeffect]

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.

[Lestat]

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...

[sideeffect]

you can get the 1.41 and also a lot of other bios files from here link (http://www.lejabeach.com/MSIK8N/k8nneo2.html)

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.php?p=3551447#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 :D

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

[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. :shrug: 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. :shrug: 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 :banana::banana::banana::banana::banana: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. :rolleyes: 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?

[STEvil]

the 4th phase, yes.

[4rory]

4 phases is better then 3 right?:D But I guess It not really that much of differnce, well at least for normal users, is this why msi doesn't use it? How much better is it?

[sideeffect]

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. :rolleyes: 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


Hoot any chance you can get a picture and point arrows to these spots to make it clearer?

[Garrett]

Hoot any chance you can get a picture and point arrows to these spots to make it clearer?
I second that question :D (thanks for pointing it out sideeffect :thumbsup: )

[Hoot]

Hoot any chance you can get a picture and point arrows to these spots to make it clearer?

Unfortunately, I'm at work today, but I'll be happy to when I get home. Perhaps I can rework other pictures available online and post them.

Hoot

[nfm]

Many people requested no temp BIOSes. So here they are:

Link (http://www.swjka.com/k8nneo2bios/index.php?dir=./need.for.mhz/v1.9%20Modded%20BIOS(s))

[sideeffect]

thanks NFM :)

[flick]

Thanks NFM. I can't speak for everybody since I'm new to these forums, but I'm sure the rest of us really appreciate your time and effort as well as putting up with all our requests. Thanks again! :toast: :toast: :toast: :toast:

[FlyingHamster]

hey guys, is the 1.9 bios good for winchesters as well? I have 2x256mb gskill and a 3500+ winchester, and im still using the 1.36b bios. should I give 1.9 a try?

[Lestat]

Hamster yes the 1.9 should work fine for the winchster also but never forget. If it aint broke dont fix it.
the 1.8 should actually be more appropriate for your winchester as reading across many forums people are showing similar issues as what are stated here. not many issue but its enough to cause most not to use it. But then again people with san diego's and X2's i think are seeing things better for them while the rest of us ae seeing issues.

if your 1.36 is working good and overclocks are good then there isnt a reason to update.
1 thing i'll note is that for me atleast when i moved to 1.8 my HTT is able to reach 370mhz but with 1.9 i cant even hit 250. and with others 300mhz is an issue. so if you havent tried 1.8mod try it see what you think.

hoot if possible please throw us some picks i for one am interested in anything that gives me more accuracy.

*note of interest*
i put in my 2nd stick of ram and my vdimm instantly shot up to 1.96v instead of the 1.55v it was. (it actually showed 1.57v on the board)
vdimm went back up to 3v also. where it was showing 2.89 (bios is 2.85)

so something about 2 dimms really messes up this mobo's voltage regulation.

[corruption]

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 :banana::banana::banana::banana::banana: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)


I'm sorry that you're still uptight about your board....you should take a laxative and relax a little. I asked how you were using the meter because your voltage readings are wonky. You can use almost anything as a "ground"...."ground" is relative. If you have ever done a 7v mod for a DC fan in your case, you'd know that you replace the ground wire with the 5v line. The 5v line becomes the "ground" for the fan's circuit. The "ground" is nothing but a reference point....in the same manner, you can "boost" the 12v to the fan by using the -3.3v line as a ground. This would result in about 15v to the fan. Not all powersupplies "like" this though and often end up malfunctioning.

As for providing documentation on adding and removing components to complete the 4th phase, I can only say that I'm not going to be doing that again any time soon. I've experimented and have found that the differences are too small to be useful for me. If you're using phase change and are pushing over 1.75v through your chip, then the 4th phase might be useful to stabilize voltages a bit. I am only watercooling and I refuse to push over 1.65v through my chip for 24/7 use. The 4th phase will help to remove "ripple" from the voltage supplied to the chip/board. If you're not experiencing voltage fluctuations, I wouldn't recommend attempting to add it. (Some others online have reported the same as I have regarding the 4th phase on the K8N Neo2 Platinum.)If you are intersted in a few mods that require a bit of soldering, you may be interested in this page: http://www.hardwaredome.nl/reviews/view/18/ I find that the cap mod helps the most for stabilizing fluctuating voltages.

As for not knowing about hardware, I can only laugh and point in your general direction. :ROTF: :slap: It's not like a technician with an engineering degree knows anything about computers or discrete technology....right? As I've said before, I hope that you resolve your MB issue. :toast:

[Hoot]

Hoot any chance you can get a picture and point arrows to these spots to make it clearer?

Sorry to take so long. My dog ate some of the neighbors grass that had been treated with 2-4-D and I spent the evening at the animal hospital, the cost of which would have bought me my choice of NF4 boards. :rolleyes:

Here is the Vcore metering point:

http://www.frontiernet.net/~hill195/ne2p/vcore.jpg

And here is the Vdimm metering point:

http://www.frontiernet.net/~hill195/ne2p/vdimm.jpg

Hoot

[Lestat]

i like those mods but Cap mods really on serve their purpose on extreme overclocking.
but wouldnt hurt on low or midrange overclocking. Any extra punch that the system takes for power will be backed up by more cap's. Think of it as when you have anice car stereo and when the bass hits your headlights dimm. throw a 1 farad cap on your positive line to your amp and no more problems!. (of course a bigger alternator helps also). same thing here more power stored for use when the system needs it.
The vdimm mod where you run 1 wire from the 3.3v line to the vdimm line is a quick and painless mod for people who have ram with a sweet spot around 3.3-3.5v. Me personally i get nervous thinking about running my 3.3v line on my 600w Powerstream upwards of anything above 3.5v.
Personally if i were to do it i would run it to the 5v line and put a adjustable resistor on it so i could adjust it as high as needed. but then again i wouldnt want to turn the dial the wrong way when i first power up the system... ouch! you'd be sending up some serious smoke signals.
the htt mod is interesting for sure as MSI didnt give us HTT voltage adjustments. ggrrr. but alas he was right higher htt multipliers dont give any more performance. infact from 2.5x to 4x ive never seen anything increase or decrease.


Again if you do come across some materials on the 4 phase removal on the dfi board or discussions about it let me know, it sounds like an interesting debate. Just something that sparks an interest in my mind and i wanna do some learning also.
but again its been shown that the DFI boards "power" in overclocking and such comes from the 4 phase power. but then agian the boards themselves are some impressive work with or without that 4th phase.

My neo 2 and me have no need to take any type laxative. My gripes are not with stability, my system hasnt run better. its with the "how's" and "why's" and "what do i do now" when it comes to this voltage going thru the roof when i put in 2 dimms.

There has to be a pin on the back of the CPU socket (rear of the mobo) that gives a 100% reading of voltages. seriously there has to be the voltage is fed TO the cpu via one of those pins so taking a reading from it ... thats the best spot.

Hoot do you know or seen anyone do this ? or know where i might find out ?
One thing that would be cool would be a schematic or several pictures showing exact locations on the mobo of all known testing spots.

[Lestat]

Sorry to take so long. My dog ate some of the neighbors grass that had been treated with 2-4-D and I spent the evening at the animal hospital, the cost of which would have bought me my choice of NF4 boards. :rolleyes:

Here is the Vcore metering point:

http://www.frontiernet.net/~hill195/ne2p/vcore.jpg

And here is the Vdimm metering point:

http://www.frontiernet.net/~hill195/ne2p/vdimm.jpg

Hoot



Many thanks Hoot hope the puppy feels better.!
Although that vcore check point is only 1/2 inch from the other one readings arent going to be much different but yes its closer to the CPU so a more accurate reading should take place.
the vdimm one i would assume would give a way better reading since it is right up by the dimms.



Edit

just checked and voltages are exactly the same so they are no different.

Vcore 1.94
vdimm 3.56v that is impossible....( the other leg is over 7v ). the bios is set to 2.85 and frankly i can tell you from experience that this ram wont even BOOT at anything over 3.2. at least my last set of this same ram wouldnt. Again if this particular ram was volted that high then i would be getting better overclocks this this on the ram.

[4rory]

Dude, I think your dmm might be going wacky, I have one there the has gone wacky and reports off voltages like that, I checked those spots and the other spots in the pervious picks, and I get less then 50mv from what bios reads vcore as.

Your chip would be dead by now, test ur dmm on something like ur house outlet, should be 110-120.

[Lestat]

now see the whole thing is that with 1 dimm in the dimm slots the voltages are exactly spot on. and i dont mean just vdimm i mean vcore and vdimm. but since its an ondie memcontroller adding the 2nd dimm ... well... i could see how it might adjust voltages to comensate.
the voltages only are reported fooked up like this when i use 2 sticks of ram.

I wish i still had my newcastle to test things with.

[Hoot]

Many thanks Hoot hope the puppy feels better.!
Although that vcore check point is only 1/2 inch from the other one readings arent going to be much different but yes its closer to the CPU so a more accurate reading should take place.
the vdimm one i would assume would give a way better reading since it is right up by the dimms.



Edit

just checked and voltages are exactly the same so they are no different...snip



Was that while running Prime95 Torture Large FFT's or BurnK7 ???

I see a little voltage drop across the toroids. Remember, there is literally 7 or 8 inches of wire wound around each of them and even though it's fairly large gauge, it does offer a little voltage drop when a lot of current is passing through it.

Were you getting those odd voltages with both your positive and negative probes on the capacitor pads, or did you cheat and put the negative on say, chassis ground and only use the positive pad? If so, you can't do that as you add ground loop currents to the measurement, which can add up to as much as .1V error. :)

Hoot

[STEvil]

throw a 1 farad cap on your positive line to your amp and no more problems!. (of course a bigger alternator helps also). same thing here more power stored for use when the system needs it.
Sorry, but car audio caps are bandaids mostly.


There has to be a pin on the back of the CPU socket (rear of the mobo) that gives a 100% reading of voltages. seriously there has to be the voltage is fed TO the cpu via one of those pins so taking a reading from it ... thats the best spot.
You might get something off one of the SMD capacitors under the metal backplate, but other than that there are only the legs of the capacitors near the vcore mosfets and the torroids.

Best to use 2+ wires if you can for the 3.3=vdimm mod too.

[corruption]

Again if you do come across some materials on the 4 phase removal on the dfi board or discussions about it let me know, it sounds like an interesting debate. Just something that sparks an interest in my mind and i wanna do some learning also.
but again its been shown that the DFI boards "power" in overclocking and such comes from the 4 phase power. but then agian the boards themselves are some impressive work with or without that 4th phase.

Sorry if I wasn't clear, I added the 4th phase to the K8N Neo2 Platinum and then removed it. I wouldn't dream of removing the 4ht phase from a DFI board. I simply meant that after adding the 4th phase to the Neo2, that I saw next to no difference. So I've stuck with the cap mod until I want to upgrade to a PCI-E board. I found that the cap mod on the site that I linked to worked quite well. If your board is out of warrantee and you want to attempt a mod, the cap mod is the one that you should try first. It helps with fluctuating voltages especially when you're pumping 1.7v+ through your CPU. ;) (I run what's in my sig 24/7 though.)


My gripes are not with stability, my system hasnt run better. its with the "how's" and "why's" and "what do i do now" when it comes to this voltage going thru the roof when i put in 2 dimms.

I've never seen the voltage jump like that when 2 DIMMS were installed, but anything is possible. If I were you, I would write to MSI and try to get the board replaced. :shrug: Do you know where your chipset was made, and week/year? (Taiwan or Korea?) I know that some of the early Taiwan chips seemed to have issues with high HTT's. I've heard that the Korea chipsets run at a slightly higher voltage than most of the early Taiwan's which could help obtaining high HTT's, but I haven't verified this for myself yet. I have a Korea 0450 and have had no problems hitting over 300HTT with any of the bios's that I've tried. *knocks on wood* (1.36mod through 1.9, including betas) I actually found that 1.37mod bios gave me higher voltage fluctuations and temps, but that's with software reporting.....and you konw how "accurate" software monitoring utilities can be. :rotf: :zombie: I'm actually impressed that the board has lasted this long in today's market. It is a really good board, but 64bit technology has matured, and there are better options available if you're buying into the 64bit market for the first time.


My neo 2 and me have no need to take any type laxative.

Sorry if I appeared to be grumpy......I haven't had much sleep lately. I've been trying to ressurect my 9800pro that recently "drowned" when my cpu block threw up on it. :stick: I think I'm going to have a new R350 keychain pretty soon though, as things aren't looking good. :brick: On the bright side, I'm happy with the x800xl. :p: :woot:

[d1c1ple]

Anyone have issues with the 1.90 bios? I finally got mine flashed and the bios menu was totally different, and after a short amount of time the CPU Fan would spin down, i reverted back to 1.80

[d1c1ple]

I was attempting the 1.90 Release, not a modded version

[Hoot]

I was attempting the 1.90 Release, not a modded version

In the "Plain Vanilla" 1.90, Cool-n-Quiet is enabled. Set it to off and see if that makes your fan run full speed all the time. I disable all the temperature related features out of necessity since my thermal diode is routed off the motherboard, so I'm not speaking from personal experience, just a hunch.

Hoot

[d1c1ple]

thx - i'll try again and see

[uniacid]

whats going on guys, just thought I'd say I received an FX53 as a replacement for my 3700 SD since AMD was out of stock of any SD's, its a clawhammer and I believe this is what it said for the model: CAAXC 0433MPMW I took pictures but they didn't come out as clear as I wanted, anyways I'm burning it in with p95 right now and was wondering if there was any suggestions as far as burning in goes? thanks.

[sideeffect]

Try Sen's Burn-In Technique

Link (http://www.ocforums.com/showthread.php?t=382751&highlight=Sententials+Burn)

Good luck with that chip i think you got a good deal really as long as you can keep the thing cool.

[Ares323]

First of all I want to say hello to everyone on this forum.

All this beta bioses are just nerve breaking things. For me vers. 1.8 works pretty good but any other 1.9 related bios is crap. I tried all the NFM 1.9 bioses an I get only errors and no post :nono: With 1.9 "official" I can get my sistem in WIN but only at default speeds not even 1 Mhz more, that`s great heh ?
My sistem: MSI Neo2 Platinum 54 G, A-Data DDR600, Venice 3800+, Asetek dual radi kit Antarctica CPU block, X800XT PE Sapphire watercooled, 2x Maxtor 120 GB Raid 0 on Sata 3 and 4.

My CPU goes up to 2.95 Ghz max benchable frequency at 1.7 volts but my memory stops at 280 @ 2.5-3-3-8. Temp readings on this mobo are just crazy, I`ve reached 60 degrees on Asetek :stick: that`s just wrong.
I hope that MSI will get the final bios up and ready soon.

[sideeffect]

Hi Ares323

That was the final bios of 1.9 :)

Yes its not to good. Im still using 1.8 rev3 with 1t mod as i think its the best bios made so far.

The temps seem ok for me on my vapochill they scale quite well with the vapochill sensor so im pretty sure its pretty acurate on some boards at least.

Dont know why NFM's 1.9 mods wouldnt work for you but it sounds like a memory setting that is messing with you as thats all he changed in the bios.

[Ares323]

The same bios I`m using also. 1.8 Rev 3, I only hoped I can get this 3800+ at 3 Ghz but it seems that I`m out of luck. I tried all the possible settings on my memory but with no luck, I tried also 1.8 volts Vcore but that was a no go. This chip doeas 2.75 Ghz at default Vcore but one step higher I must pump the Vcore over 1.55 volts. I am convinced that my chip can do 3.0 Ghz on a DFI mobo.

[Hoot]

...snip...I am convinced that my chip can do 3.0 Ghz on a DFI mobo.

If you're talking the DFI NF3 Ultra-D, going to that from the Neo2 Platinum is akin to "Jumping from the frying pan into the fire." I just tried that route. It is as screwy as the Neo2. Sure there are a few lucky souls who get great results and the rest of us, like lemmings, give ourselves high blood pressure chasing their good fortune. A test of a good motherboard is not how well a few people get it to overclock, with a few processors and a few memory types. The test is how well the majority of people get it to overclock. Zoom out, read other forums, the NF3 Ultra chipset is doomed because it does not represent the upcoming revenue stream for these manufacturers. It's a bothersome legacy that the manufacturers throw their excess talent at, when they have a little money left in the budget at the end of the month. I'm certainly not going to throw more good money after bad. I'll enjoy my setup for what it's worth and start saving money for when a real solution comes out. Considering how much money that represents, I'll probably have it about the same time that real solution comes out. I'm feeling pretty tired of whipping this horse and it has put me in a cynical mood. That doesn't mean "Not Objective" nor "Trying just to stir the pot", what I said is heart-felt and based upon 30 years experience in the industry. Over that much time, you get a sense in your gut for things. You just heard my "gut feeling".

Hoot

[sideeffect]

The same bios I`m using also. 1.8 Rev 3, I only hoped I can get this 3800+ at 3 Ghz but it seems that I`m out of luck. I tried all the possible settings on my memory but with no luck, I tried also 1.8 volts Vcore but that was a no go. This chip doeas 2.75 Ghz at default Vcore but one step higher I must pump the Vcore over 1.55 volts. I am convinced that my chip can do 3.0 Ghz on a DFI mobo.


If your memory does 280 then cant you use the 11x multi and do 273x11 ? I also had a lot of luck with the 10.5 multi seemed to work well with the san diego 3700.

[sideeffect]

pretty sure im going to be able to beat 3ghz now what i thought was the cpu crapping out wasnt at all. Ram is scalling quite well with Cas 3 might be able to do 275 x 11 :D. Cas 3 vs 2.5 isnt really noticable in benchmarks from what ive seen as well.

[Lestat]

If you're talking the DFI NF3 Ultra-D, going to that from the Neo2 Platinum is akin to "Jumping from the frying pan into the fire." I just tried that route. It is as screwy as the Neo2. Sure there are a few lucky souls who get great results and the rest of us, like lemmings, give ourselves high blood pressure chasing their good fortune. A test of a good motherboard is not how well a few people get it to overclock, with a few processors and a few memory types. The test is how well the majority of people get it to overclock. Zoom out, read other forums, the NF3 Ultra chipset is doomed because it does not represent the upcoming revenue stream for these manufacturers. It's a bothersome legacy that the manufacturers throw their excess talent at, when they have a little money left in the budget at the end of the month. I'm certainly not going to throw more good money after bad. I'll enjoy my setup for what it's worth and start saving money for when a real solution comes out. Considering how much money that represents, I'll probably have it about the same time that real solution comes out. I'm feeling pretty tired of whipping this horse and it has put me in a cynical mood. That doesn't mean "Not Objective" nor "Trying just to stir the pot", what I said is heart-felt and based upon 30 years experience in the industry. Over that much time, you get a sense in your gut for things. You just heard my "gut feeling".

Hoot


Hoot i gotta tell you that was very well spoken and 100% true.
and i can only repeat what ive been saying for sometime. If your thinking your gunna run these new chips on this old nf3 chipset your playing a game of roullette your not gunna win. end of story face the fact and go buy a good nf4 or better mobo.



Sideeffect hafve you tried other cpu multipliers. ? like i had mentioned my board with this chip eats up an 8x multi. let me crank an HTT up to 370mhz where as the others dont.

again. this board and a san diego isnt gunna work properly guy i think you know that in your heart of hearts.

[Hoot]

For those who care, here's a picture of where to read the Vagp. If you can't get your bearings from the picture, this monitoring point is once again the missing capacitor between the AGP socket and the CPU.
http://www.frontiernet.net/~hill195/ne2p/vagp.jpg

Hoot

[Lestat]

i was gunna ask if increased stability has been seen when adding that cap back but i realised that 8x agp is only 0.8v

you should set up a website Hoot it dont have to be how to's unless you wanted to but basic info about proper voltage checks on a new 2 and other mobo's that you know about.
it would be quite helpful to so many people.

btw i like those nuts you put on the back of the ATI bracket. you must use a different HSF . I've added a spacer under the center and felt like that added enough pressure to ensure a tighter fit of the hsf. but the nuts are quite a nice touch.
wish i still had my tap and die set id thread tho posts on my stock hsf.



Ya know how wonderful and accurate a DMM reads when you put in a NEW 9v battery ? http://www.myemailemoticons.com/drop_down/funny/funny_20.gif

[corruption]

Ya know how wonderful and accurate a DMM reads when you put in a NEW 9v battery ? http://www.myemailemoticons.com/drop_down/funny/funny_20.gif

LOL.....I was getting wonky readings about 3 weeks ago. I was scratching my head thinking wtf :censored: , as I checked voltage from a new AA battery to verify my meter's accuracy. I should have realized that it needed new batteries when I started to notice that the screen was difficult to read. :brick: :lol:

[Ares323]

If you're talking the DFI NF3 Ultra-D

I was talking about DFI NF4.

I tried to boot with 11x250 , 1.425 + 10% then I upped the Vcore to 1.7 volts and with clockgen to 3 Ghz, ram at 273 @ 3-3-3-8. I get an instant restart when I open SuperPi or Prime.

[sideeffect]

The 8x multi isnt going to cut it even 370 x 8 is only 2960mhz. The 9x i have tested and the 10x they did no better than each other. The 11x is best for me with the 10.5 showing good results to.

Im pretty keen on staying in 1:1 as its much easier to become stable there. My ram is loving 2.65 volts and is more stable there than at 2.75 and 2.8 volts where it would error.

275 x 11 is 2 hours mb2 stable and superpi 32mb stable so far. Going to test it with Doom3 and Far Cry 64 later.

edit - trying 280 x 11 = 3080mhz memory 280 3-3-4-8. Produced fastest pifast and sisoft mem bandwidth results so far even though i lowered trp to 4 :)

[Corbu']

Ares323, you need a new chip to reach 3 Ghz. That 3800+ will never make it :slapass: and I don't even mention 3Ghz and Prime95 in the same sentence.

[Lestat]

can i ask why you assume a 3800 wouldnt do 3ghz ?

you see dozens and dozens of 3000's do 3ghz every day so why would you say it wont do 3ghz ?

is it a 3800 venice ? san diego ? winnie newcastle ? what ?
it may not reach it on air but water or colder sure its possible.

and as for Prime and 3ghz in the same sentence ?

how about 2.65ghz and prime not even close dudes.... but i am at 2750 and i can run any game or 3dApp or other program even Pi and it works fine but prime fails... why is that ?

[sideeffect]

Hi guys Systool (http://www.techpowerup.com/systool/) has been updated to allow exporting to 3rd party apps so you can now make your own motherboard monitors for neo 2 that actually show the right +12v and +3.3v rails. MBM and speedfan and most other tools dont show the correct reading for the neo 2. Systool does though so try it out if you havnt already.

Heres mine :p:
http://www.swjka.com/k8nneo2bios//monitor.jpg

[Hoot]

That sure is the "Swiss Army Knife" of applications. Will it read temperatures off of the SMBus as well as the flawed Winbond 83627 so common on these motherboards?

Hoot

[Lestat]

0.1 doesnt show that. i dont see neo2 listed in the mobo section

what application are you using that has that interface look ?

[Hoot]

The 8x multi isnt going to cut it even 370 x 8 is only 2960mhz. The 9x i have tested and the 10x they did no better than each other. The 11x is best for me with the 10.5 showing good results to.

Im pretty keen on staying in 1:1 as its much easier to become stable there. My ram is loving 2.65 volts and is more stable there than at 2.75 and 2.8 volts where it would error.

275 x 11 is 2 hours mb2 stable and superpi 32mb stable so far. Going to test it with Doom3 and Far Cry 64 later.

edit - trying 280 x 11 = 3080mhz memory 280 3-3-4-8. Produced fastest pifast and sisoft mem bandwidth results so far even though i lowered trp to 4 :)

Not being too proud to try someone elses recipe, I flashed NFM180_3_1T and had a shot at your results. I was disappointed that I couldn't get more than 11x260 3X 1:1 3-3-3-8-1T, but then I re-read your post and realized that you were using a Vapochill, so I didn't feel too bad. It is a promising version of the 1.80 bios none-the-less. One bright spot was though I could not get above 260 HTT, I could drop the CAS to 2.5 and it still worked. SuperPi was only 29.282 though. When I first got this setup, it would do 10x300 @ 1.64V 3X 5:6 2-3-3-8 @ 1.95, but it seems I have "rode it hard and put it away wet" too many times. :) It was fun trying though.

Hoot

[harpyboy]

0.1 doesnt show that. i dont see neo2 listed in the mobo section

what application are you using that has that interface look ?

samurize

[harpyboy]

i said ur posts at systool forums sideeffect.... :)

but im not really towards running systool and samurize at the same time to get get temp readings.... btw tat looks nice.. simple and does things right :)

any idea if the NB temp is correct ? coz even corecenter doesn't have this reading...

[h4nc0]

http://msi-forum.de/thread.php?threadid=18007&sid=14d1f718353c6569518d2e8e2b5b8938

New BIOS out. v1.A3

[Lestat]

A as in Alpha stage bios ? woah no way if its alpha stage

[h4nc0]

A as in Alpha stage bios ? woah no way if its alpha stage


I believe it's a beta version.

[Hoot]

I'm a sucker for a new release. I flashed it and set it to priming before heading off to work. Will see how it's going when I get home.

Hoot

[sideeffect]

Well systool and samurize together only use a small amount of ram maybe 5mb combined and they can both be started on windows start so its not much of a problem.

The Temp readings all seem about right for me Northbridge temp has that 5000rpm fan beating down on it and does feel cool to the touch so i can quite believe its 20 degrees. The harddrive temps and ambient 6800 temps are all read about the same which suggests its accurate.

Neo 2 is added to systool Lestat in the newer versions you look in Hardware monitoring then settings and motherboard.

Gonna check out that new bios now. Thanks for posting it h4nc0 !

Seems that they have fixed bugs in the venice i dont know how much it will help the san diego. I always thought they were the same chips just with venice chips maybe failing 1mb cache test or being speed bins but it seems that they are not.

Think i might wait for Hoot to post back with results. Good luck Hoot :slap:

Seems they have fixed the voltage fluctuations but in doing so have made the max vcore even lower in this bios can anyone confirm that because i really need 1.63 volts to keep this cpu stable.