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

[silent-circuit]

...ok, it has now done this twice.

While running Prime95 on all 4 cores the machine will suddenly lose power. Completely. Just go black. It is on a UPS with AVR (CyberPower 1500VA) and is pulling ~410watts OCed to 3.6Ghz at full load according to the UPS's display (that number includes the 24" LCD) so the PSU isn't anywhere near being overdrawn so far as I can tell -- even assuming 75% efficiency (it is likely doing better) and all power going to the PSU (which it's not) that's still ~308 watts internal draw, which is less than 1/2 of what this thing is rated for.. Don't know what's going on here. Voltages seem to be fine, like zero fluctuation, but I'm just using a POS multimeter.

It did this at stock once, too, so it's not the OC (or not just the OC).

I think I may RMA this board to Newegg and see if I have the same issues with another. Seriously starting to wonder if it's defective or on its way out on top of everything at this point.

[x-built-Stan]

Overspeed, BKA, All, (A little OT)

If you change "Edit Options", "Thread Display Mode" to "Linear - Newest First"
the newest replys always show on page 1.

Makes it very easy to find the new posts.

[BKA]

Overspeed, BKA, All, (A little OT)

If you change "Edit Options", "Thread Display Mode" to "Linear - Newest First"
the newest replys always show on page 1.

Makes it very easy to find the new posts.

Thanks, that works for me.

Been spending the couple of hours doing trial and error with my Q9450. So far I can do 450 with the following settings:

FSB: 1800(450)
VCORE: +.20000
NB: 1.45
CPU GTL: 65
FSB Term: 60
Everything else auto.

Screenshot

I tried to increase the FSB a little with the same settings and Windows froze up on me, so I probably need more Vcore or NB voltage to go higher. Thats it for tonight I guess, I'm going to run it at 3.0(375) for everyday use.

FSB: 1500(375)
VCORE: .06250
CPU GTL: 65
FSB Term: 60
All others on Auto

This was all done with the 1.16 Beta BIOS.

[theYipster]

BKA,

That's good news for the Q9450. I think 3.6ghz is going to be tops for a lot of these chips unless you pay through the nose for a 790i Ultra board. If you can get it any higher, let us know.

[loonym]

And not only the cost of the 790i, which alone I might be able to swallow. Have you priced 2x2gb ddr3 kits?

[theYipster]

And to think I bought 8 gigs of CAS 4 PC-6400 DDR-2 for as little as $105 last December!

[silent-circuit]

Prime stable over night at stock now, no strange power-down, so hopefully that's a thing of the past.

Again, someone mentioned a P07 BIOS -- does anyone here have it, and would they be willing to send it to me?

[dogg]

No success with flashing bios... first I tried to use usb utility.. everything was going right, after restart I've choose usb to boot, but it posted to vista... Than I tried to use AFU but it said that file is not bios file..Downloaded some diffetent bios file P05 and the same thing... Am I doing something wrong?
I got USB enabled in BIOS and boot from other device enabled as well..


__________________________________________________ ________________
MSI P7N Diamond (780i)
E8400 Penryn on air (Zalman CNPS8700NT)
2x2G Geil Evo One DDR2 CL4 800MHz
2x MSI 8800GT 1GB OC in SLi
2x 150GB WD Raptor
750GB Seagate Barracuda
Vista 32Bit

[loonym]

Is the USB stick formatted to be bootable?

If you try and use AFUWIN, after you select Open, in the files of type pull down menu select All Files (*.*), then browse to the bios file you wish to load. The Main BIOS Image box will already be checked. Simply select Flash, when it's completed shutdown, clr cmos, power up.

[silent-circuit]

Now running reasonably stable at 3.4Ghz, 1+ hour Prime95 quad core stable. Vcore is a bit high, still tweaking there, but temps are under 60C load.

I maintain it's the board/BIOS and not my chip, as 3.4 was torture to get stable just as 3.6 was. It took a lot of screwing around with CPU GTL and FSB REF. It ended up that high RSF REF and moderate CPU GTL were necessary. Sadly the same settings don't work at 3.6. Hopefully later BIOS revisions and more info on exactly what the settings mean/do will allow a higher OC.

[dogg]

Is the USB stick formatted to be bootable?

If you try and use AFUWIN, after you select Open, in the files of type pull down menu select All Files (*.*), then browse to the bios file you wish to load. The Main BIOS Image box will already be checked. Simply select Flash, when it's completed shutdown, clr cmos, power up.

Yes, it works... thanx loonym... but there is CMOS Checksum Bad on startup, so I can choose F1 to enter the setup or F2 run default... if I enter the setup, then save and exit, there is the same cmos error and F1, F2 options...



MSI P7N Diamond (780i)
E8400 Penryn on air (Zalman CNPS8700NT)
2x2G Geil Evo One DDR2 CL4 800MHz
2x MSI 8800GT 1GB OC in SLi
2x 150GB WD Raptor
750GB Seagate Barracuda
Power: Coolermaster M series 850W

Vista 32Bit

[theYipster]

Silent,

Out of curiosity, I read a bit about the differences between the X3320 and the Q6600. There isn't nearly as much OC and benchmark information out there as there is on the Q6600, and there's very little info comparing batches (unless they share Q6600 batch numbers.)

However, what I did read is that despite arguably better thermal characteristics with the Xenon, people do seem to be a bit less successful OCing them, particularly on nForce boards. I've seen several instances of the usual conversation regarding the correlation between VID and OC potential, or lack thereof. If your VID is indeed as low as you say, your chip may have a voltage barrier below 3.6ghz. It's known to happen on the lowest-VID Q6600s. They'll be superstars up to 3.4-3.6ghz, and then suddenly hit an almost insurmountable voltage wall. Reviewing the massive amount of Q6600 VID and batch information I've collected, the most successful chips seem to have a VID in the 1.22-1.25 range. Still, batch more telling than VID, and in my experience a decent predictor can only come from knowing both (and having batch comparison info.)

Still, if you were stable at 3.6ghz for any reasonable amount of time, that's good news. Please post your current BIOS settings... Perhaps we can help you get back to where you were.

[loonym]

Did you clear the cmos dogg?

[silent-circuit]

It hard locked and bluescreened after about 1 hour 45 minutes to 2 hours at 3.4Ghz. More IRQ_NOT_LESS stuff.

Settings:
1511 FSB
800 RAM
RAM timings: 5 - 5 - 5 - 20 2T
2.1 VDIMM (stock RAM timings/voltage/speed)
100 PCIe Frequency
.175 vcore
1.45 NB voltage
20 CPU GTL
63 FSB Term

[theYipster]

What happens if you use those same settings, but try an FSB Term of 25?

Hmm, another thing I would consider doing is instead of going straight to prime95, try running a custom 1 hour OCCT 2.0 test set to RAM+NB only. This is the best means I have found to find stable NB settings. If you are unstable, I would solely focus on playing with vNB and FSB Term until that test passes. Then, I'd leave those settings in place, and concentrate on vCore and GTL Ref for the Prime95 small fft test.

Dividing vCore / GTL and NB / FSB Term into discrete sets to tune separately has helped me better manage this BIOS.

[loonym]

I like those numbers though even so it bluescreened. The reason for this is because I've been experimenting hitting fsb holes, which there are some I believe, and I hadn't tried in low 1500's range much after I found a big one after about 1457 I think. It's a pain in the ass trying to find because success means no post cmos clr

[silent-circuit]

FSB Term of 25 does not seem to do well with this board/chip. It seems to need higher ones. Normally it will not get in to Vista with lower FSB Term values.

At FSB Term 25 It will run Small FFTs in Prime for a while, but if I try Large in place FFTs 2+ threads will die pretty much immediately.

Changed nothing else. I'm going to see if 63 makes the large FFT test work better. Hadn't run it before.

[loonym]

If you can pass small fft it's really a good signal your cpu is not the main factor in the failure. I needed increase in nb gtl ref voltage for stability in my benching with high fsb along with high fsb terminator voltage (which you are using).

edit: I also needed lotta vcore for 3600 which I don't necessarily mind but my house is hot and I don't want my cpu running mid 60's c crunching 24hr a day.

[theYipster]

I like those numbers though even so it bluescreened. The reason for this is because I've been experimenting hitting fsb holes, which there are some I believe, and I hadn't tried in low 1500's range much after I found a big one after about 1457 I think. It's a pain in the ass trying to find because success means no post cmos clr

Good point. Remember that when OCing on an nForce, FSB holes come with the territory. Best explanation I've seen for this ever-present phenomenon (they're on the 790i too) is that its due to the manual RAM speed feature.

[silent-circuit]

Gone back to stock after failing repeatedly to get it well and truly stable at 3.4Ghz. I'm about ready to either return the board and get a reference board (the XFX 780i is the same price at Newegg and they may allow a swap) or just wait for a better BIOS and hope. For the last time, I do not believe it is my chip, but the board/BIOS that is causing problems for me. The USB issues alone are so off-putting to me that I may return it for that reason alone.

[dogg]

Did you clear the cmos dogg?

yes I did.. now after some restarts I figured out, that my G15 keyboard was probably causing that... so I changed it with old ps2 kb.... the screen is the same... cmos checksum bad with f1 f2 options but staying there just for a second and system posting.. probably still the same usb device problems as before.. startup beep also changed when usb kboard is plugged off.. strange souds ,anyway ,on this board..
now I can start tweakin a bit, finally

thx loonym and Yipster

[loonym]

lol, I'm just glad you worked it out

[dogg]

lol, I'm just glad you worked it out

not really, swapig the keyboards working ... usb can post through the checksum bad screen but is not able to get to bios, probably is still not mapped while is the right time, so Im usin ps2 one, but with this stay in that screen... anyway,
even when I use one for getting to bios and than i swap them (wondered that it works) it results start in defaults whatever changes I made in bios and than save and exit.. I will continue checking what is behind that... maybe clr cmos button is not working propperly....

[loonym]

here's a spi 8m run

vdroop gets bigger for me after I pass 1.4v

[AaronYuri]

My account was finally activated, thanks Mark.

As I said in the beta/performance BIOS thread, I will continue to upload beta and performance BIOS releases as they come.

I was talking with an MSI engineer and he will ask Taiwan about the integer values. I should have a reply next week.

I only got around 3 hours with this board, as a BIOS flash went wrong (power cut) and bricked my board. But I will post what knowledge I managed to gather from using the board in the short time I had with it.

I found the board is a good clocker and runs really cool compared to the reference designs. There's 2 things I can pick at however. The first being the VDroop, it got to the level of 0.6v at one point. As loonym said, after about 1.4v is you're problematic stage. The second point I can pick is the memory speed settings. You're required to set the memory speed manually, if you fail to calculate it correctly the board won't POST. So you need to have a calculator on hand.

The best way to clock this board is spend quite some time in the BIOS, get a notebook, pen and calculator.

I found this board a challenge, it was even harder than the ICFX3200 - but I love a challenge. Keeping a cool head is most important, if you get pissed and angry, you're only going to do worse. Overclocking is only truly trial and error when you use this board.

The voltage settings that had the most effect on overclocking were: (excluding CPU voltage)

NB Voltage
FSB Terminator
CPU GTL Ref

I don't see why you need to change the others just yet.

As soon as I get a reply about the integer values, I will post it on the MSI forums and here. If you have feedback on the BIOS releases I'm uploading please post them in the thread here: http://forum.msi.com.tw/index.php?topic=115073. Please don't complain, keep it informative.

Thanks guys, keep clocking!