Official MSI P7N Diamond (780i) Discussion/Review/Overclock/Guide/BIOS Thread

Welcome to the thread Goldfish.

On other boards, a GTL Ref of 65-67 is generally recommended for Q6600 overclocking. Note that I haven't yet played with GTL Ref on this board yet, so I can offer no direct experience. I would recommend however that you up the GTL Ref and see what results you have.

Also, you may find better results with higher FSBs if you set the FSB Termiantion Voltage between 12-15. Note that on the shipping BIOS, I experienced no FSB holes between 1333 and 1600, but I haven't yet done further tests to try and stabilize at 1600 / 3.6ghz.

Also, which BIOS version are you using? People are beginning to see better results with 1.1B4.

EDIT: Lastly, what is your Q6600's VID and batch number?

Quote:

---

Originally Posted by theYipster

Welcome to the thread Goldfish.

On other boards, a GTL Ref of 65-67 is generally recommended for Q6600 overclocking. Note that I haven't yet played with GTL Ref on this board yet, so I can offer no direct experience. I would recommend however that you up the GTL Ref and see what results you have.

Also, you may find better results with higher FSBs if you set the FSB Termiantion Voltage between 12-15. Note that on the shipping BIOS, I experienced no FSB holes between 1333 and 1600, but I haven't yet done further tests to try and stabilize at 1600 / 3.6ghz.

Also, which BIOS version are you using? People are beginning to see better results with 1.1B4.

EDIT: Lastly, what is your Q6600's VID and batch number?

---

hey does anyone know if the 1.1b4 fix's the 45nm multi problem??

I've heard they fix cold boot issues -- which may be the same thing. I don't have a 45nm chip so I can't confirm.

i will get msi to send the bios and i will try it out,, this is my rig,,,coolermaster 830 stacker,p7n, qx9650 4 gigs corsair 1066, a pair of foxconn 8800gts'640mb clocked to 656/1220,xi-fi card 1000w real power pro and raptor's in raid-0-,,,running 3.8ghz stable for about 1 week now, fsb term is at 8 cpu volt up 3 ,,,nb 1.4 sb 1.55 ram at manual 5-5-5-15 1t and pci freq to 107 i am pretty happy so far but i know it is capible of more

Hello all,

Last week a got this board, because I was having memory issues with an Asus P5E. The good news is that it will run the 2x2GB modules that had previously gone bad in the P5E. But the bad news is that NewEgg sent me the same modules back that I RMA'd. So those modules return errors under MemTest86. Also the 4 Corsair 1GB modules that I wanted to run won't seem to run on this board either. I suspect that those Corsair modules have something programmed on the SPD that prevents them being used together. Though they are exactly the same model.

In other respects I like some things about the M7N Diamond, and don't care for some other "features". Things I do like include; eSATA, though I had to jump through some hoops to get it working, that bundled X-Fi card is nice, though it doesn't work with the front ports on my case, and my Video card is boosted a bit, since it's also an Nvidia chip. It's also a comfort to me that what memory I am running is being used at it's proper speed and timings.

I did have some installation issues with XP. At first I didn't take the time to wipe the drive I was installing to, and I didn't have XP do a full format on that drive. This led to strange issues with phantom devices taking up many of the addresses for hardware. Which in turn prevented XP from properly initializing the JMicron SATA ports. That was corrected by wiping the HDD and reinstalling.

Now the only outstanding issue I'm having is the occasional BSoD, when I use my ROM drives. See the Nvidia SATA Performance Drivers don't play well with SATA ROM drives. So occasionally when I try to use a ROM drive (I have 2 SATA DVD ROM drives, one is read only and the other is a DVD/RW) I get a BSoD for "Driver_IRQ_Not_Less_Or_Equal..." and it flags the "nvgts.sys" file as the culprit. (That "nvgts.sys" is the Nvidia SATA Performance driver.) It seems to happen more when the computer has tried to enter sleep mode. The only consolation is that those Nvidia drivers do allow me to use NCQ on my hard drives, and the BSoDs are infrequent.

The nvgts.sys is a known issue across all nForce boards -- reference and otherwise. I never experienced it on any of my boards until I plugged-in a SATA optical blu-ray drive. Apparently it can be buggy with SATA optical drives.

If you don't want to uninstall the drivers and just use the generic ones provided by Windows (a common solution to this issue,) try relocating the SATA cable to a new port in your motherboard instead. This eliminated the issue on my rig with the blu-ray drive.

So I had a chance to test out the 1B4 Beta BIOS this afternoon. Unfortunately it did not allow me to break the 3.42Ghz barrier on my E8400. I did some pretty intensive voltage experiments w/o success. I have not tried the P05 yet.

New high hit on the E8400. 3.51Ghz using these values:

BIOS: 1B4
FSB: 1560
RAM: 780
VCORE: +0.2v
DIMM: 2.1v
NB: 1.35v
SB: 1.6v
FSBT: 60

Stability unverified, but so far so good.

Edit: Above values refined after experimenting, still able to achieve 3.51Ghz:

BIOS: 1B4
FSB: 1560
RAM: 780
VCORE: +0.15v
DIMM: 2.1v
NB: 1.35v
SB: auto
FSBT: auto

I'm not sure what role the FSBT plays but I was able to get here with it left on auto. SB modification also seemed to be redundant. System would only become stable when NB voltage bumped up. 1.2v = fail 1.25v = fail 1.3v = OS load but not stable 1.35v = stable sweet spot. I'm also noticing that the vcore can remain relatively low ... at least at my current best of 3.51Ghz. More to follow.

Overspeed,

I've asked MSI for a detailed explanation of all BIOS voltage settings. With this information we can shed a bit more light on the OC possibilities of this board.

However, it looks like the new BIOS is doing the trick for you. Two things to keep in mind:

On the reference 780i boards, 1.4v is often needed on the NB to achieve stability between 3.6-4.0ghz. Furthermore, I believe the reference boards default NB voltage to a higher value -- 1.3 or 1.35. Considering that the P7N uses the same NB, I would expect similar voltages are necessary to achieve stability. Try a NB in the range of 1.4-1.45 and see if that can't get you stable at 3.6ghz, let alone 4.0v.

As for your vCore, the consensus on a maximum voltage limit for OC'd 8400s is 1.45v. Unfortunately, thanks to the P7N's unique way of setting vCore, it's not necessarily easy to tell when you've set 1.45v in the BIOS. To do this, you need to know your chips VID. With Speedstep disabled, open (or first download if need be) Core Temp in Windows and read the number for VID. To get the actual voltage set via vCore in the BIOS, simply add the BIOS vCore to your VID. Note that the voltage you set in the BIOS will be higher than the voltage supplied to your chip while idling in Windows, thanks to vOffset, and that the voltage will further drop under load, thanks to vDroop. Your chip, however, should reach 4.0ghz on less than 1.45v, provided it's not a dud.

Also, what is the batch of your E8400? Many can do 4.0ghz on air as easily as a Q6600 G0 can do 3.6ghz. I had one (a Q746A) I sold which apparently can bench 4.5ghz on air on an evga 680i, but it does need a bit of voltage to get there.

theYipster, I have been following your posts on the MSI forums, hopefully Aaron can get us the voltage info. I'm not sure how to get my VID value. Hang on while I Google this... ok I have installed CT 0.96.1. My VID is 1.0375v. I don't think I will be able to find out my batch for the CPU. This PC was assembled for me and I don't have the original retail box or access to the CPU as I have a large aftermarket heatsink on it. I'd rather not remove the heatsink and ruin my thermal compound. Is there any other way to locate the batch via an application?

So it would seem the bottleneck for this CPU is the motherboard. My research shows that the E8400 is capable of easily achieving 4+Ghz with a reasonable heatsink on intel chipsets. Judging by my successes/failures so far, even if a new BIOS helps to get at 3.6Ghz, it seems unlikely that it will ever do 4.0Ghz, my original goal. A shame because when I ordered this PC I was clear that I wanted a good OC board for the E8400 and the P7N is what I was given. Who knows, maybe in time, and with a better undersdtanding of the voltage settings, 4.0 might be within grasp along side refined BIOS's.

I am concerned about putting the NB at 1.4v - 1.5v at it seems to get too hot at those settings. I'm unsure of the cooling potential of the rollercoaster and circupipe, but to touch them at default clocks, they seem hot already. Then again, my finger is not the best for judging temps lol. What feels hot to me might very well be cool for the NB. I have no way of measuring.

Overspeed,

People are indeed having success getting their E8400's to 4.0ghz on 780i-based motherboards, and I am quite confident that it can be done here too -- if not with the latest BIOS, than likely with a future one. Take a look at the 780i and overclocking forums over at evga.com. Right now the evga 780i A1 is the most popular 780i board on the market, and as such has the most OC information available (as well as the biggest community to help people achieve their overclocking goals.)

Now many of the settings for the evga board don't apply to the P7N, due to the different BIOS used. However, you will no doubt see people putting more volts through the NB in order to achieve their goals. I am confident that you will need to do the same. At 1.4v, my circupipe does get rather hot, but it's not burning to the touch. I did order an Antec Spot Cool fan so that I can keep the NB reasonably cool while pumping more voltage through it. If you're goal is 3.6-4ghz, which I do believe is obtainable, then you should look to do the same.

Note that for my quad to boot and benchmark at 3.6ghz, I need 1.4v in the NB

EDIT: Unfortunately if you did not build the computer yourself, getting the batch number will be difficult. It will involve dismantling the cooling system from you cpu so that you can read the numbers across your cpu's heat spreader. I wouldn't recommend doing this -- it's certainly not worth the hassle.

Quote:

---

Originally Posted by Overspeed

Yes I did review your results, I think using shipping BIOSs we have both hit the same wall. I was using the MSI app to monitor my temps. I will try your HW monitor and see if it's different. That said, someone once told me that the E8400's were produced with a sometimes faulty temperature monitoring diode.

Edit: Ok I have CPUID HW running. I also set my bios to your exact settings. Here is what it reads:

vcore: 1.29v
tmpin0: 23C
tmpin1: 47C
core 0: 35C
core 1: 36C
gpu: 56C

What are the tmpin temps, I think 1 may be the mobo temp. I will now put under stress and see where the temps go...... ..... ..... .....
Ok according to CPUID HW monitor my temps hit 42C max after a 15 min prime95 run. The cores cooled back to the low 30`s within a minute. One thing I do notice about your settings is my NB did not get nearly as hot and still seems stable.

---

Quote:

---

Originally Posted by Overspeed

New high hit on the E8400. 3.51Ghz using these values:

BIOS: 1B4
FSB: 1560
RAM: 780
VCORE: +0.2v
DIMM: 2.1v
NB: 1.35v
SB: 1.6v
FSBT: 60

Stability unverified, but so far so good.

Edit: Above values refined after experimenting, still able to achieve 3.51Ghz:

BIOS: 1B4
FSB: 1560
RAM: 780
VCORE: +0.15v
DIMM: 2.1v
NB: 1.35v
SB: auto
FSBT: auto

I'm not sure what role the FSBT plays but I was able to get here with it left on auto. SB modification also seemed to be redundant. System would only become stable when NB voltage bumped up. 1.2v = fail 1.25v = fail 1.3v = OS load but not stable 1.35v = stable sweet spot. I'm also noticing that the vcore can remain relatively low ... at least at my current best of 3.51Ghz. More to follow.

---

Haven't checked in on the thread today, I picked up God of War for PSP and got caught up playing that most of the day. Anyways, yes my NB did seem pretty hot when set at 1.35 volts and under full load. I also had the same thing with the NB voltage, until I bumped it up to 1.35 Windows wouldn't even load without locking up. And I agree about the CPU not needing a lot of voltage.

With CPUID HW tmp0 is system temp and tmp1 is MB temp if I remmeber correctly. You have a pretty cool system temp going there because mine is in the low 30's. I get your kind of system temps in the summer when the air is on but normally get higher system temps in the winter with the heat blasting. And this has been one cold, snowy winter here in MI.

I'm currently back to default settings until probably this weekend. I will also try flashing one of the beta BIOS. I want to get some gaming in for now, still have to complete Hellgate London and Crysis. I would love to hit 4.0 for benchmarking purposes but would like to get a nice 3.4 - 3.6 stable OC for everyday computing.

theYipster, here is a look inside my case. You will notice right away my CPU cooler and the circu-pipe underneath. 2 inches below that is the 8800GTS. I wonder how I could possibly get a fan on that northbridge given my available space?

http://alpha.customfitonline.com/P1000678.jpg

Antec Spot Cool is a small fan on an adjustable stock. I think it would do the trick.

Completed a 26Hr Prime95 Small FFT run on my new rig (see signature). CPU at 3.4GHz, Vcore: 0.2000, FSB 1512, FSB Terminator: 30, NB: 1.5v, Memory 800MHz 4-4-4-13. All other settings are Auto. Temps got as high as 61C on the core and 63C on the NB.

I'm going to start moving up to 3.5GHz.

Overspeed,

From looking at your picture, I would say yhat you have a real airflow problem on the NB radiator. You might be able to mount a small fan like the Antec Spot from the CPU cooler fan housing to at least get some air flow in there.
ASUS provided an optional NB vertical fan that attached to the radiator of their Striker Extreme board. Perhaps something liek that would do the trick, I included a picture.


I have also included a few pictures of my build (named Mercury). Note the clear air path from front to back over the NB radiator.

Stan, I see how you might think that based on my photo but that giant CPU cooler is moving a fair amount of wind which defects off the MB down the the NB rad. Case has rear, top, front and bottom fans all pushing the hot air out. I may still benefit from a spot cooler, if I could only figure out how to mount such a think. Just our of curiousity, how are you able to monitor NB temps?

This is WAY of topic but something is wrong with my keyboard when using this forum. The keys I type and the keys that show up are not the same. For instance, this is what gets printed when I type question mark: É É É É É This is what gets typed if I hold shift and hit numbers 1 through 0: ! " / $ % ? & * ( ) Notice the missing number symbol, at symbol, and the up arrow is replaced by the real question mark. Has anyone ever seen a keyboard behave like this? I swear this only happens when typing in this forum. In notepad, Word, in other forums, my keyboard behaves correctly. GRRRR.

TheYipster, hi and thanks for the info. I'll try to get it up to 3.6GHz later using your tips, but due to busy workschedule I have little time for my own projects.

CoreTemp shows my VID as 1.3125v and have no clue to what the batch number is. Gonna change the coolingpaste and lap the cpu after the weekend, so i'll take a looksy then before I lap it. It's a G0, SLACR revision by the way.

I'm using the BIOS shipped with this board, don't know where to download the beta bios (Well, haven't really tried to find it either) :P

The FSB hole i'm experiencing is strange, seems like i'm the only one.

Goldfish,

Certainly keep us posted on your efforts. If you purchased the chip yourself and have the retail box, you can get the batch number from the white product label. It's under the section and barcode labeled FPO/Batch #. It will begin with an L. For example, mine is L737B266. Batch # together with VID are the best predictors of OC potential.

Unfortunately, a VID of 1.3125 is very high for a Q6600 G0, and you may find it quite difficult to achieve a stable 3.6ghz regardless of the motherboard used. It's not impossible and I have seen it done, but you you will likely need to pass upwards of 1.5v to the chip at load to see any stability. Since there's isn't a way (yet) to reduce vDroop on the P7N, you may need to set vCore in the BIOS upwards of 1.55v or 1.6v, which is a bit much for an air cooled chip.

Note that with my own experience with the four G0s that have passed through my hands, two had VIDs of 1.325v and neither could stabilize at 3.6ghz on an evga 680i with a tuniq tower. By the time I watercooled, I had gotten rid of both chips. One wouldn't post at 3.6ghz regardless of vCore, while the other would post with a reasonable vCore around 1.42, but wouldn't stabilize at anything higher.

Still, it's worth a shot, and if you can get the batch number, you'll at least have a more complete idea of whether or you'll succeed, as VID alone isn't as strong a prediction as VID and batch. Be sure to watch your core temperatuers while OCing. If they exceed 70-75c, then it's time to back down.

You can get Beta BIOSes from MSI's User to User forum: http://forum.msi.com.tw Note that you may have to register to get to the BETA Bioses sub-section, which can be found under the overclocking forum.

Also note that while I haven't experienced any noticeable FSB holes with my P7N, nForce boards are indeed famous for them, so I suppose it's to be expected. Often times new BIOSes will eliminate or at least move FSB holes.

Overspeed,

Look at the ASUS fan picture. The fan is a vertical mount, so the botom of the fan sits on top of the NB radiator (roller coaster). The fan draws air up from the radiator bottom and pushes it out the fan top. Note the paddle wheel fan blades. Since the fan simply clips to the radiator, you would not need any additional mount points. From your picture it looks like that might work. You can get the fans on e-bay for $10. Look for "Optional Heatpipe Fan Asus Striker". I'll try mounting mine on the radiator and let you know if it works.

TheYipster, I trew away the retail box long time ago. I am using some cheap coolingpaste now, my AS5 tupe was unfortunately empty.
I had a EVGA 680i earlier, but the thing died after just 2 weeks of use. The temps was at 70c max @ small FTT's in Prime95 (@3,2GHz 1.32v).

Now can pass 81c with ease using this bad coolingpaste, so I have backed down to 3GHz @ stock for now.

Goldfish,

My apologies, but if you're passing 70c on the 680i at 3.2ghz, then 3.6ghz is really out of the ball park for your chip w/o substiantially better cooling -- an even then it's still iffy. I hate to say it, but you may have gotten a G0 dud. It happens :( Btw, what cooling are you using?

Yipster

I'm using Noctua NH-U12F with a Arctic Cooling PWM fan, max 1500rpm in a Antec P180 case.
BTW, I used the cheap coolingpaste on the evga also. Think things will improve once I have lapped the CPU, I lapped my old P4 3,2GHz and that did the trick. Shaved 5C of idle and about 8C on load.

I got no fancontrol on this board either, but that may be my fans without a rpm-wire so the board can't detect my fans. Next month I will receive a bonus at work, maybe i'll invest in new high-reving fans. I can handle some noise when i'm playing games, but when the machine is idleing I want the fans to slow down. No luck using SpeedFan either.
Got one chassisfan with an rpm-wire, but can't control that one either.


I am not to worried about i can't overclock that high, the cpu is still fast and can with ease run at QX speeds. It's just a hobby :)

Edit: on the evga it was 24h stable @ 3GHz with 1.23volts (My EVGA was vDroop modded)

theYipster, are you currently using the 1B4 bios? Will it use lower multis with your q6600? I've only been able to successfully use 6-7-8 multi with the .P05.

GOLDFISH, even with cheap silicone paste and unlapped surfaces I think those temps are too high for your hs to be working correctly. Are you certain it's making full contact and properly tightened?

Loonym, I'm still using the shipping bios. I've had little time to do anything this week but look over this forum when I can. If I have time tonight I'll flash -- if not, then this weekend.