Overclocking problem?!?!

[engineereeyore]

So I have the following:

MSI K8N Neo4 Platinum motherboard
AMD Athlon 64 3200+ Venice core
2x512MB OCZ EL Platinum DDR400 Ram
Thermaltake 430W power supply
Thermaltake Venus 80mm fan/heatsink

I can't seem to get the processor overclocked any higher than 2.3GHz. I'm using a 230 FSB, 4X HT multiplier. Ram is set to 180/200, so reads around 207. However, this is as high as I can put it. If I try to increase to 240 or even 235, the computer won't boot. The temperature has never, even under load, gone over 42 degrees, so I don't think it's temperature. The motherboard is supposed to be one of the best overclocking motherboards on the market, so I can't believe it's the motherboard. My best guess is that I'm not getting enough power. According to bios readings, the power seems pretty stable and close on all voltages, but I can never check it after increasing the FSB.

I've tried increasing the FSB to 300 and changing the cpu multiplier to 8x, and ram to 133/200, which would be back to 200. This doesn't work either.

If anyone has any ideas, please help.

 

[Bona Fide]

Have you tried increasing the vcore?

 

[engineereeyore]

yep, i have it set at 1.45V

 

[GMtheBest]

Try a newer \ different bios. Loosen you ram timings to 3-3-3-10 (set mem to 100), CPU multi 6X, HT multiplier 3X, try raising the FSB. If it doesnt go any higher it may be a motherboard limitation. What power supply do you have, you may not be getting enough power

 

[NINaudio]

I would also use everest or cpu-z to get a reading of what your vcore actually is. It may be lower than what you have it set at, in which case you may need to raise it more. If you're on air cooling I'd keep the voltage at 1.55v or less. Definitely put the HT at 3X. What voltage are you running the ram at? You may want to bump that up a bit, to around 2.8 or 2.85.

 

[engineereeyore]

I've tried two different sets of ram, and the first I tried at a pretty low speed without result. I've used the BIOS, Fanspeed, and the MSI core utility to check the voltages. The one that seems low, but only on fanspeed, is my 3.3V supply. According to fanspeed, it's barely on the edge, around 3.17. If I recall correctly, most electronics will operate on a swing down to 3.13, so it's pretty low. But only fanspeed shows it, so I'm not sure how reliable that is. My vcore reports in at about 1.43-1.44, so that should be ok, I'd think.

My ram is set for 2.65, but I haven't look into the voltage for that.

I updated my BIOS already thinking maybe that was the problem, but no go.

I'll check the ram voltages and try bumping the FSB up higher with a lower cpu multiplier and see what happens. I'm thinking of just replacing the power supply though with a OCZ Powerstream 520W. I just don't want to do that and then find out that's not the problem. But then, can you really go wrong upgrading to such a nice ps?

 

[the cobbler]

I just bought this board and am testing it as we speak.

strong overclocker, huge problems with the bios. there is a FSB bug with CBBID winchesters week 41 or later and with some of the Venices, though the Venice issue may be solved by later bios revisions (not the Winnie. crap. i have a CBBID 0501 3200+ lol)

bug limits FSB and is related to HTT multi and goes something like this:
Venice cores
FSB bug at 219 mhz in some bios/CPU combos
FSB bug at 229 in some bios
FSB bug at 239 in some bios
you should only have one FSB bug regardless of HTT multi used, though some Venice owners seem to report a range of bug, i.e. 230-240FSB=no post.

Winnie
Htt multi @ 4x= 219 bug
@3x= 229 bug
@2x=229 bug
@1x=239 bug
I tested these all thoroughly and that is what I found. 1mhz deviation upward at each setting resulted in NO POST.
(all bugs from bios 1.6 official. I was able to hit 240@3x multi with 1.71beta, so maybe they are getting closer to solving the problem)

this bug affects all Winnies 3000/3200 week 41 or later, many Venices (hit or miss apparently) and absolutely no SanDiegos whatsoever.

I have the same bug on all bios revisions of my VNF4Ultra, but HTT multi was irrelevant and i could not break 250 FSB no matter what.

If you update the bios and still can't get around it, clock the board through ClockGen via a startup file. Mine is running right now at 304 FSB through ClockGen Prime stable overnight. Go figure.

which bios revision are you using? 1.3 is supposed to be the best OC-er, while 1.61beta and 1.6 officials apparently fix the Venice bug for some additional weeks of that CPU. 1.71/1.72/1.73beta bioses seem better than official 1.6, but you lose a bunch of memory dividers.

I almost got the DFI -DAGF. Starting to wish i had.

 

[Merovingian]

Please post your max HTT, CPU, and RAM speeds using zebos guide and stock voltages or 1.425V CPU and 2.9V TCCD if you have it, then I can help you as I have the same mobo with 3000+ @2.7 on air. PM me if I don't get back soon. My bet is your ram settings are not loose enough and you are not running TCCD.

Edit: Where is Zebo's down and dirty overlclocking guide as a sticky?

 

[engineereeyore]

I'm running BIOS version 1.5. I've looked on the MSI website and don't see a version 1.3 for the MS-7125. There is one for the MS-7030 which even states the it fixes the FSB problem, but there doesn't seem to be one for the MS-7125. If you know where I can download it, please give me a link.

I'll look up the ClockGen program and give it a try also.

I agree with you 100%. Really starting to wish I'd bought the DFI motherboard.

 

[the cobbler]

Do this to isolate the problem:

set CPU multi to 7x or 6x
set memory divider to 100
HT multi to 2x or 3x
bump NF4 voltage up a bit

start bumping FSB 10mhz at a time until you get no post. If you still hit a wall at 240, try booting at exactly 239. If you can boot at 239 but not 240, you have the bios bug. Any limitation you run into at this point is strictly FSB related, memory is completely out of the equation, as is CPU speed.

next:
CPU multi to 10x
mem divider 100
HT multi at 2x or 3x

you know what to do from here. again, any limitations you run into are either FSB or CPU related, memory is completely out of the equation here.

now download and install ClockGenNF4. Set up your voltages and set your HT mult to 3xi and memory settings/divider in BIOS, then boot and run ClockGen. use 8x CPU multi, click 'apply FID', then 'Get values'. start bumping the FSB up again and see how far you get.

I bet you get waaay past 240. I don't think you have a memory problem. Heck I am using cheap RAM and can do 239 with a 180 divider straight out of bios, but then I run into the '239' bug and can't get 240 even with HT multi at 1x and mem at DDR200. Clearly, memory is not the issue.

Through ClockGen, I am running over 300FSBx8, 3x HT multi, DDR 456 on double-sided value ram (2x512mb corsair vs 2.5,3,3,7 1T/dual channel). On a WINCHESTER. My chip will do ~2.65 at 1.7v, 2.43 is good for my stock cooling. You should be able to do my chip's 1.7v max speed on AIR at <1.55v.

there is no way your venice stops at 2.3ghz. seriously. something is wrong with this board's bios.

 

[the cobbler]

Originally posted by: engineereeyore
I'm running BIOS version 1.5. I've looked on the MSI website and don't see a version 1.3 for the MS-7125. There is one for the MS-7030 which even states the it fixes the FSB problem, but there doesn't seem to be one for the MS-7125. If you know where I can download it, please give me a link.

Here you go, from Rebel Haven

 

[engineereeyore]

Cobbler, you're the MAN!!! Changed back to BIOS version 1.3 and lo and behold, I'm at 2.4GHz and climbing.

Thanks everyone for the help.

 

[the cobbler]

Awesome, great to hear that fixed your prob, very happy i could help.

well, i love this damn board now, this thing is a wicked overclocker. I was going to take it back for a DFI but am thinking of getting a San Diego instead and just keeping the board.

i can't believe this is the board that brought down the MSI forums, lol. Rock solid and fast as heck, although it actually runs somewhat warmer than my VNF4 Ultra did.

 

[the cobbler]

just another really nice surprise regarding this board-chipset runs SUPER COLD, must be a heck of a chipset fan on here.

i haven't even AS5-d my NF4 yet and am running 1.80v through it with idle temps in the 20s and load temps~44C after 5 hours of StressCPU(GROMACS). ...bet you could get this thing close to 500FSB with the right RAM and a Danger Den.

 

[imported_LabRat]

Hello,

Had the same problem with my new computer.
Couldn't get any FSB over 225

Motherboard: K8N neo4 Platinum
CPU: AMD winchester 3200* ( 2.0 )

Try the new 1.6 BIOS from MSI. Get it from LIVE update.
No boot problems and i set my Winchester 3200+ to FSB 240 with mulitiplier 10 without any problems last night.
Had to increase the CPU voltage to get at stable system, but it works fine now.


Also tried to boot with FSB 250 and no problems... seems like the "lock" is gone.

 

[the cobbler]

the bug is not gone

only affects Winchester CBBID week 0410 and later

I have the 1.6 official right now and can't get past 239, and that's with a 1x HTT multi. with the 1.71, i could hit 250 with a 4xmulti