nForce 790i SLI/Ultra chipset mobos - Reviews/OC/Guides

[Guacamole]

Thanks. But don't bother to spend too much time on it. I was just curious that's all. I will probably be running on higher FSB anyway for my 24/7 settings.

Yeah I actually had no luck at 7-6-6-24 only 7-7-7-24 works for me.. I froze and then rebooted a couple of times but I am with you I don't plan on running there anyway...

[tekjunkie]

Yeah I actually had no luck at 7-6-6-24 only 7-7-7-24 works for me.. I froze and then rebooted a couple of times but I am with you I don't plan on running there anyway...

Thanks buddy, for taking time to check it out. Much appreciated.

[mike]

I wanted to - but didn't have the patience to go thru all the posts - how many people have managed so far to push past 585fsb on wolfsdale with a 790i? I found 3 585+ results so far - but I am not sure if these are radnom "special" results or not. Been hoping to find a (any) DDR3 board that can do 600 - but been falling waaaay short so far.

Before I try another 790i ( would be my third...) is it worth the effort with wolfsdale???

[Coupa]

Argh, i just tried EVERY bootable GTLref voltage as well as every green NB, SB, PLL, and vTT voltage.

For some reason, the the best GTLref voltages for 1.20 vTT were + 98/104/104/98, which seems a bit high to me, but that seems to make the difference between playing for 5 minutes and 5 seconds in CoH.

Voltages at the moment are:

CPU Voltage: 1.28125v
Loadline calibration: Enabled
PLL: 1..6
vTT: 1.2
Mem: 1.5 (its cheap memory, anymore then 1.5 heats it up so much the computer restarts after a while)
NB: 1.56
SB: 1.55
GTLref: 98/104/104/98

The PCIe 1/2 are also at 104, cause if use 100 it crashes within 5 second again.

Does anyone have any suggestions? All i can think of now is to up the vTT by .2, then try every bootable GTLref again, then up by .2, etc etc. I dont really want to do that, it takes long enough on just one vTT, but i will if i have to.

[sonofander]

I gave up on trying to stabalize my memory synced at 1811, just couldn't do it...

so I backed it off to unlinked at 1610 and started to explore the boundary conditions of the data corruption. I found that it was dependent on NB, RAM and VTT/GTL voltages. According to sfc scan, I have integrity violations at NB = 1.46 and RAM = 1.9 even though I am prime stable. a bump to NB=1.52 and Ram=1.96 and there are no more data corruption issues per sfc scan!

so here are my settings that are 100% prime stable and have no file integrity issues!!!

FSB: 1811
RAM: Unlinked @ 1610, 7-7-7-24 (all subtimings on auto)
Multi & Vcore - set to your preference (I've tried @ 9 multi for 4.0 and 9.5 multi for 4.3 and they both work)

VTT=1.36
PLL=Auto
RAM=1.96
NB=1.52
SB=1.6
GTLs=-29,-32,-32-,29

on my machine these settings are 100% stable (prime & 3D) and result in NO CORRUPTION!!!

[Guacamole]

I gave up on trying to stabalize my memory synced at 1811, just couldn't do it...

so I backed it off to unlinked at 1610 and started to explore the boundary conditions of the data corruption. I found that it was dependent on NB, RAM and VTT/GTL voltages. According to sfc scan, I have integrity violations at NB = 1.46 and RAM = 1.9 even though I am prime stable. a bump to NB=1.52 and Ram=1.96 and there are no more data corruption issues per sfc scan!

so here are my settings that are 100% prime stable and have no file integrity issues!!!

FSB: 1811
RAM: Unlinked @ 1610, 7-7-7-24 (all subtimings on auto)
Multi & Vcore - set to your preference (I've tried @ 9 multi for 4.0 and 9.5 multi for 4.3 and they both work)

VTT=1.36
PLL=Auto
RAM=1.96
NB=1.52
SB=1.6
GTLs=-29,-32,-32-,29

on my machine these settings are 100% stable (prime & 3D) and result in NO CORRUPTION!!!

I tried those exact settings but I sill have corrruption, but then I have corruption at stock and most people don't have corruption at stock...

[zbogorgon]

Can somebody tell me where tha f**** can I get P02 bios ? Thank you

[Broly]

I am now running 3.64 ghz like last time with a 9 multi and 405 FSB.
The kicker?
1.30 v and only 1.30 VTT! :O
PRIME STABLE?!

anything higher things get ercky. I think it's a bios issue but I'm happy that I'll probably be able to run a 23pct overclock with a 1620 fsb NP with low volts.
Great freaking board.

Waiting on the new COrsair dominator 4 gig so I can buy em

[slim142]

Update #1

Removed heatsink city from SIIE, cleaned crappy thermal paste, applied MX-2, put back heatsink city and now temps are 32c for NB in idle (48c before) and 31c in idle (46 before) for the SB. They go up to 44c when they are in load and since I still dont get my case, I think this temps will go even lower with all the yate loon Im putting in my Armor+.
After doing this, I seriously recommend it to everybody. It is SUPER EASY! as long as you keep your motherboard on the plastic that came with it in the box so the backplates will stay in place. Takes around 40mins if you take your time to do it good. Im glad I did this and now 1.5v wont be a problem for my SIIE.

I have also installed my Thermalright 120 Extreme but so far, I have NO fan on it. Temps for my Quad are 44-46c in idle (shouldnt they be lower?) and for Load they go up to 64c. I have to keep in mind I have no fan on my 120E and also my house is kinda hot today. Will see how well the fans and the case helps on this. I remember for my E6600, when I placed in the Zalman9700, my temps went from 64c(idle) with the intel fan to 26c (idle) with the Zalman and 38c for load (at stock speeds 266Mhz FSB). I think the fan will really make a difference with my Quad.
Also, I want to mention that my heatsink gets REALLY hot, I was worried that I didnt do a good job with the paste and installation but it looks like I did because after 5 mins of orthos my 120E gets really hot to the touch (top part) so I guess the Artic Silver 5 is really working as well as the lapping I did to my 120E and the cleaning with Artic Purifier. Now I just need a fan and place this monster on my case.

Is 2:12pm here in FL and Im still waiting for the FeDeX guy to come. I hope he doesnt trip on the 15-stair set that my building has while carrying my Armor+, what do you think, should I go and help him out? :P

[Big Myke]

glad to see you go through with replacing the factory tim from the heatsinks

[Roger_D25]

Yeah Slim142 that is a huge difference! I wish I would have run my S2E with the stock TIM on it before replacing it(just to see what the difference would be). It was the first thing I did after opening up the box, lol! I was pretty dissapointing with the stock TIM placement, a good part of the SB wasn't even making contact! I must say that having screws and backplates instead of plastic push-pins is a very nice change and makes life much better!

[slim142]

Thanks Myke

And yes Roger, I agree. The push pins were a pain and they werent making the heatsinks have full contact with the chips. BTW, It took me 9 hours to build my rig, is 12:30am here and Ijust started to clean up. I must say this was a hard but worth day. I finally have it ready for the torture tests!!! Tomorrow is mothers day though and I wont be in the pc for obvious reasons, but will start with the tests monday first hour!

[Roger_D25]

Looking forward to it slim, by the way your not the only one taking a day off from computers (at least most of the day), lol!

[M@G!C]

Yeah Slim142 that is a huge difference! I wish I would have run my S2E with the stock TIM on it before replacing it(just to see what the difference would be). It was the first thing I did after opening up the box, lol! I was pretty dissapointing with the stock TIM placement, a good part of the SB wasn't even making contact! I must say that having screws and backplates instead of plastic push-pins is a very nice change and makes life much better!

When i changed TIM to MX-2 i saw a drop of 8ºC in NB temperatures.


Well i think i´ve made some progress...
After several hours of testing, i found some quite simple settings, that until now(fingers crossed), are working quite well for me!
I was previously running my QX9650 at 1400QDR and 1800Mhz memory as i was unable to go further without BSOD, freezes, or messages in boot saying files are missing or corrupt...

Here are my settings now:

1800QDR running LINKED and SINKED (450FSB)
Multiplier 8.5 = 3825Mhz (I think i will be able to raise the multiplier more but i haven´t tested)
Memory timings 7-7-7-18 1T
P1 ENABLED
P2 ENABLED
VCORE - 1,33125V (set in BIOS)
Load Line Calibration - DISABLED
VDIMM - 1,86v (set in BIOS)
NB - 1,40v
SB - 1,55v
GTL ref 0 -29
GTL ref 1 -32
GTL ref 2 -32
GTL ref 3 -29
All spread spectrums DISABLED
All the other settings where left AUTO
BIOS 0603

With this settings i was able to run PRIME95 stable for 2 hours then i manualy interrupt, run SUPER PI 1M and 32M several times without errors and 3 loops of 3DMARK06..
During MAX LOAD in PRIME95 my cores reached a maximum of 49ºC mesured with REALTEMP (I´m on aircooling with TUNIQ TOWER)
I shut down and restarted several times and everything was just fine, no freezes no BSOD!
I will post some screenshots later...

NOTE - My memory (Cellshock BLUE 1866MHZ)or mobo are very sensitive memory voltage settings!!! Going below 1.86 to 1.80 result in errors during post saying files are missing or corrupt(no matter how many times i restart!!). Raising the voltage from 1,86v to 1,90v results in BSOD during post!! (no matter how many times i restart!!)

[slim142]

I am done building my system. At stock speeds, it has been running Orthos for more than 8 hours estable. I cant say how good I feel right now because I thought I was going to run into problems.

My system specs (real quick, most important stuff only)
ASUS Striker II Extreme (...)
Intel Core 2 Quad Q9450
G.Skill DDR3 HZ Series 2Gbs 7-7-7-18
(2) 8800GT SC editions from eVGA
Seagate 7200.9 250Gbs (Using it for now just in case of data corruption if I decide to OC)
Corsair TX750
Windows XP SP2 (no updates since I havent set up internet on it)
NO DRIVERS (keep reading)

So the first thing I noticed is that BIOS 0402 was very buggy. Manual settings were just not working at all, orthos failed in 5 secs so I decided to move on to 0507 (I want to try the closest next BIOS release to see which one feels safer and estable) and everything is going OK so far with stock speeds and RAM running at designed timings/speed.

What did I do?
First, manually imput RAM settings (timings, voltage, speed, CR)
Went to voltages and put everything in AUTO except for RAM (1.86v because it really gives out 1.90) and LLC Disabled.
Unlinked ratio
P1/P2 Enabled
Spread Spectrums Disabled
GTLRefs at default
Multi=8x
CPU options like C1E, speedstep, virtualization etc DISABLED ALL and ALL cores enabled.
Then other settings like Sata, IDE, onboard stuff... I selected the ones I liked, enabled disabled some this really doesnt matter.

Now, I left the system running all night and Orthos has been running perfectly for more than 8 hours and is still running. However I have to mention THERE ARE NO (NO) DRIVERS INSTALLED ON MY SYSTEM YET!!!, NOT EVEN NFORCE OR GEFORCE DRIVERS!. ONLY WINDOWS XP SP2, NO UPDATES AT ALL FRESH INSTALLATION.

Now, I have a feeling nForce might cause system inestability or it can also be eXtreme OC, CANT confirm because I havent tried the drivers yet but just to make sure of this, I wanted to run the system at stock speeds with no drivers and all seems normal.
I will end orthos after a couple minutes more and then Im going to install drivers like nForce and Geforce drivers....

[slim142]

Update #2

Ok guys, something told me this had something to do with the sata drivers and it looks like they are somehow the responsables for the corruption and stuff.
I closed Orthos with 10 hours estable with no drivers as mentioned above, then I decided to install nForce 9.64, GeForce 169.44 and System Tools 6.01. Restarted the PC 3 times (one for each installation) then checked system management to check 8800GT s were recognized and the SMBUS etc. I enabled SLI and then checked temps to make sure everything was still fine. Finally I decided to open orthos and RUN and guess what? Orthos showed

"Couldnt write MBM, No MBM written" (or something like this, this also happened to me with 0402)
Then after 4 seconds ( same as BIOS 0402 )

ERROR SYSTEM HALTED... couldnt write temp file... (same as BIOS 0402)

What!!!, ok so fresh installation, fresh driver installation and then this Hmmmm!. I havent panic yet, so I decided to up a little my NB and SB voltages since this time I had SLI enabled (When I ran orthos for 10 hours, I had no SLI since i had no drivers). I went to BIOS and manually put 1.5v for SB and 1.36v for NB (1.36v gives 1.39v and sometimes 1.37v, but if I leave this in auto, it gives me 1.32v I think).

Restarted and now I have started Orthos again, its been like 3 mins and everything looks fine so far. Fingers crossed wishing this was all SB undervoltage due to SLI.

But again, this is still the very beggining of my testing. I dont want to jump to conclusions but we will see how it goes.
Now I would like to ask you tekjunkie, GAR, 5010, Roger_D25, M@G!C, Big Myke, Broly, Sonofander and everybody else experiencing Orthos failure, data corruption and unestable gaming.... Do you guys install the nForce drivers right after finishing installing windows?

[bluehaze]

Getting ready to install my 790i EVGA in a couple hours (Mothers Day), reading some of these posts has me wary I am on a 780i now with no issues but I dad have plenty of issues if I installed NVidia latest drivers with the "media" something or other checked in the drivers. this is the Sata and raid drivers the board was flaky as all hell with those drivers installed but was stable as a rock without them installed.

Have any of you tried running the board without updating the drivers via Nvidia instead just using the Windows update drivers?

[GAR]

whats a safe pci-e frequency on the eVGA 790's??????? using a 9800GTX

[Broly]

Update #2

Ok guys, something told me this had something to do with the sata drivers and it looks like they are somehow the responsables for the corruption and stuff.
I closed Orthos with 10 hours estable with no drivers as mentioned above, then I decided to install nForce 9.64, GeForce 169.44 and System Tools 6.01. Restarted the PC 3 times (one for each installation) then checked system management to check 8800GT s were recognized and the SMBUS etc. I enabled SLI and then checked temps to make sure everything was still fine. Finally I decided to open orthos and RUN and guess what? Orthos showed

"Couldnt write MBM, No MBM written" (or something like this, this also happened to me with 0402)
Then after 4 seconds ( same as BIOS 0402 )

ERROR SYSTEM HALTED... couldnt write temp file... (same as BIOS 0402)

What!!!, ok so fresh installation, fresh driver installation and then this Hmmmm!. I havent panic yet, so I decided to up a little my NB and SB voltages since this time I had SLI enabled (When I ran orthos for 10 hours, I had no SLI since i had no drivers). I went to BIOS and manually put 1.5v for SB and 1.36v for NB (1.36v gives 1.39v and sometimes 1.37v, but if I leave this in auto, it gives me 1.32v I think).

Restarted and now I have started Orthos again, its been like 3 mins and everything looks fine so far. Fingers crossed wishing this was all SB undervoltage due to SLI.

But again, this is still the very beggining of my testing. I dont want to jump to conclusions but we will see how it goes.
Now I would like to ask you tekjunkie, GAR, 5010, Roger_D25, M@G!C, Big Myke, Broly, Sonofander and everybody else experiencing Orthos failure, data corruption and unestable gaming.... Do you guys install the nForce drivers right after finishing installing windows?

Heya slim. You may be on to something. I'm going to uninstall my chipset drivers and push the clocks. THis is a very very reasonable explanation because my chip can hit 4 ghz no problem but I"d never make it in normal mode, only safe... Will keep you posted. THanks again slimmo

[5010]

The "nforce mediashield" have always been CRAP and until someone proves otherwise I will stay with Windowz drivers. I've still had corruption but not my whole OS, my corruption was quite specific and since adjusting my VTT + GTLVREF I haven't had any corruption issues at all. If anything, having those crap drivers installed can only increase the likelyhood of data corruption.

[bluehaze]

Well I got it installed, took a bit longer than expected as the first time my Tuniq Tower was not seated properly there are a couple capacitors very close to the CPU socket and the heatsink was resting on them so it was not fully seating on one side! Took it back off and fixed it. Haven't done much tweaking just set CPU to 1.5v CPUFSB to 1.45 and Mem to 2.0v to see if it would boot into Vista at 1600FSB Linked and Synced 3.6ghz q6600 with OCZ Platinum DDR3 2x2gb pc1600. So far so good crossing my fingers gonna go play some more now.

Also I bought this board used and it doesn't have a fan for the northbridge like my 780i has. It's an EVGA 790i was it supposed to come with this fan? Bios reports 35c for northbridge so i guess it's not a problem but if it came with the fan i would like to get it from the seller!

Thanks

[tekjunkie]

Update #2

Ok guys, something told me this had something to do with the sata drivers and it looks like they are somehow the responsables for the corruption and stuff.
I closed Orthos with 10 hours estable with no drivers as mentioned above, then I decided to install nForce 9.64, GeForce 169.44 and System Tools 6.01. Restarted the PC 3 times (one for each installation) then checked system management to check 8800GT s were recognized and the SMBUS etc. I enabled SLI and then checked temps to make sure everything was still fine. Finally I decided to open orthos and RUN and guess what? Orthos showed

"Couldnt write MBM, No MBM written" (or something like this, this also happened to me with 0402)
Then after 4 seconds ( same as BIOS 0402 )

ERROR SYSTEM HALTED... couldnt write temp file... (same as BIOS 0402)

What!!!, ok so fresh installation, fresh driver installation and then this Hmmmm!. I havent panic yet, so I decided to up a little my NB and SB voltages since this time I had SLI enabled (When I ran orthos for 10 hours, I had no SLI since i had no drivers). I went to BIOS and manually put 1.5v for SB and 1.36v for NB (1.36v gives 1.39v and sometimes 1.37v, but if I leave this in auto, it gives me 1.32v I think).

Restarted and now I have started Orthos again, its been like 3 mins and everything looks fine so far. Fingers crossed wishing this was all SB undervoltage due to SLI.

But again, this is still the very beggining of my testing. I dont want to jump to conclusions but we will see how it goes.
Now I would like to ask you tekjunkie, GAR, 5010, Roger_D25, M@G!C, Big Myke, Broly, Sonofander and everybody else experiencing Orthos failure, data corruption and unestable gaming.... Do you guys install the nForce drivers right after finishing installing windows?

This is the order that I had installed.

OS installation -> nforce drivers(only LAN and smbus drivers, no mediashield)
->Display drivers->windows updates.

[OBR]

dont use a 0603 bios, my system is after update corrupted and have to reinstall OS

[slim142]

Heya slim. You may be on to something. I'm going to uninstall my chipset drivers and push the clocks. THis is a very very reasonable explanation because my chip can hit 4 ghz no problem but I"d never make it in normal mode, only safe... Will keep you posted. THanks again slimmo

Guess what Broly? My system is 100% ROCK ESTABLE WITH DRIVERS INCLUDING ALL SATA, IDE AND MEDIASHIELD CRAP FROM NVIDIA. Solution? Manually imput 1.50v for the SB (even if you dont have SLI, 1.5v is the default) and upped my NB voltage to 1.36v (which gives out 1.39v in load). Thats all I did and my system is 17 (yes people, 17 hours) estable not a crash not an error. , now I just need to test 3d Estability...
Why might have the drivers cause orthos to crash? Because I think if you leave your NB voltage in auto, the motherboard gives the NB around 1.2v and thats waaaay too low for the NB to work specially with SLI and high-end RAM.
Keep in mind I havent OC my system so I dont know if OC+Nvidia drivers=Corruption, this might be related one to the other since unestable OC will definitely interfere with your OS.

The "nforce mediashield" have always been CRAP and until someone proves otherwise I will stay with Windowz drivers. I've still had corruption but not my whole OS, my corruption was quite specific and since adjusting my VTT + GTLVREF I haven't had any corruption issues at all. If anything, having those crap drivers installed can only increase the likelyhood of data corruption.

I'll be posting my picture of 17 hours estable with the SATA/IDE/Mediashield crap installed. It seems that I was wrong at the beggining and this was all related to NB voltage. Probably, when I begin OCing, I will need to increase VTT and GTLVRefs.
However, as mentioned above, I havent OC my system so I dont know if OC+Nvidia drivers=Corruption, this might be related one to the other since unestable OC will definitely interfere with your OS.

This is the order that I had installed.

OS installation -> nforce drivers(only LAN and smbus drivers, no mediashield)
->Display drivers->windows updates.

This isthe right way to install drivers. If you want, dont include the SATA controllers etc to prevent corruption problems.

dont use a 0603 bios, my system is after update corrupted and have to reinstall OS

You know what makes me think that your problem is? The Striker II BIOSes seem to be really sensitive when they are updated. Thats why some people find 0601 and 0603 crap compared to 0511, because they found the spot to make it work on 0511 but on 0603 it wont work until you find the right voltages, we all know this but usually is not that serious, but it seems this motherboard BIOS change a lot.
Ex: 0402 wouldnt be estable, upgraded to 0507 and it was rock solid at the same settings. Enabled SLI and I had to up voltage from the NB to estabilize it.

[Roger_D25]

Slim - So your current stable system settings are mostly stock (Q9450 @ stock, GSkill @ stock, etc...)? That is a great place to start when testing for stability, now as you increase your clockspeeds on everything it will be easier to figure out what might be causeing instabilitiy! Most people don't like doing it this way becasue its time consuming and a bit borning but it will pay off on the other end when your running at 5Ghz, lol! Good job man!