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

Quote:

---

Originally Posted by Broly

Hmm I'm still tweaking my GTL1 and GTL3 refs but it seems core4 is my slouch and is always dying in prime.
Anyone know some good VTT settings for the qx9650? Right now I'm running 1.36 VTT and 1.3625 Vcore.. I'm just trying to tune the GTLs so that I get a solid prime and then I'll bring down the volts.

So far I've tried +3 percent (so 70 percent, but 70-67 percent requires a +3percent*vtt*1000 = x mV increase) and I've tried negative values too.
Anyone else have any experience with the GTLs, have negatives worked well for anyone on quads? Not duals? Does more GTLn ref ruin stability or just feed excess volts and not affect stability? Really trying to fine tune here and seeing how far off I am to get the right values.

---

Broly I have had the SIIE for about a week and althought I still have nagging issues with data corruption or the appearance of such I can tell you this..

After reading techjunkies post I have had great success with overclocking my QX9650 Stable rock solid at 1780, 1640, 1680, 1800 and a few others

Here are the settings I have used to great succes prime stable for hours
QDR/RAM=1780 Unlinked
PCIE1/PCIE2 = 100
SPP<-> MCP= Auto
vCORE=1.375(1.344 in windows per cpuz)
Loadline =Enabled
vTT = 1.28
PLL=Auto
vDIMM=1.9
vSPP =1.46
vMCP = 1.55
CPU GTL_REF0 = -29mV
CPU GTL_REF1=-32mV
This is at a muli of 9 but I am also rock solid at 9.5 with 1.425 vcore set in bios memory unlinked at 1600 or as close as I can get it.. Try these settings they work.. My other buddy has this board and we are both having great success...:up::up::up:

P.S. Ignore my signature I just haven't updated it yet....

Quote:

---

Originally Posted by Guacomole

Broly I have had the SIIE for about a week and althought I still have nagging issues with data corruption or the appearance of such I can tell you this..

After reading techjunkies post I have had great success with overclocking my QX9650 Stable rock solid at 1780, 1640, 1680, 1800 and a few others

Here are the settings I have used to great succes prime stable for hours
QDR/RAM=1780 Unlinked
PCIE1/PCIE2 = 100
SPP<-> MCP= Auto
vCORE=1.375(1.344 in windows per cpuz)
Loadline =Enabled
vTT = 1.28
PLL=Auto
vDIMM=1.9
vSPP =1.46
vMCP = 1.55
CPU GTL_REF0 = -29mV
CPU GTL_REF1=-32mV
This is at a muli of 9 but I am also rock solid at 9.5 with 1.425 vcore set in bios memory unlinked at 1600 or as close as I can get it.. Try these settings they work.. My other buddy has this board and we are both having great success...:up::up::up:

P.S. Ignore my signature I just haven't updated it yet....

---

Good to hear that I could of help. LOL, the irony here is that I am having problems with my OCZ 2x2GB kits getting it stable, even passing memtest without any errors is a struggle @ 1333QDR/1600 RAM. Go figure. I got it Monday and haven't had time to play around much with it.


For the people who have been asking for a matrix for VTT/GTL, I have made one. Here it is


http://img146.imageshack.us/img146/2...tgtlgy8.th.jpg

anyway for those who running qx9650 on s2e found a :banana::banana::banana::banana:ty bug...dunno on other cpus..but no matter what voltages u give when you're running at 1600 fsb...won't work..i had also a bug on striker extreme...you have to pass 1600 fsb atm running 1610X10 and works fine

My 790i freezes after 2-5 minutes while playing CoD4 with 8800GT in SLI. It works fine non-sli. So I though the pcie bus was overloaded. I clocked the cards down to 500/750 and now cod4 works for hours. It's not a solution, but it might cast some light on the issue.

Can anybody else with gaming problems try this out?

this board is amazing when it doesn t give corruption problems

Quote:

---

Originally Posted by |SiLA|

anyway for those who running qx9650 on s2e found a :banana::banana::banana::banana:ty bug...dunno on other cpus..but no matter what voltages u give when you're running at 1600 fsb...won't work..i had also a bug on striker extreme...you have to pass 1600 fsb atm running 1610X10 and works fine

---

Yes I agree 1600 is a FSB hole and there seems to be no way to stabilize it but 1610-1680 I can get stable no problem...

Quote:

---

Originally Posted by Guacomole

Yes I agree 1600 is a FSB hole and there seems to be no way to stabilize it but 1610-1680 I can get stable no problem...

---

Guys what BIOS are you running? I ask because I've posted the same thing, with my E8400 FSB1600 was no good at all. Once I went up to FSB1605 though all issues went away (BIOS version 0511). Interestingly however FSB1600 works just fine with my QX9770 (same BIOS)? I only installed this chip late last night but I gamed for a couple hours, did some minor 3D benching, and ran Prime for a bit over 1hr with no problems. Granted I didn't Prime overnight (although I will very soon) I wasn't even able to boot into Windows at FSB1600 with E8400!

Techjunkie - Thanks for posted that table, much apprecaited! I recently printed off Anandtech's S2E review where they get pretty deep into the GTL settings with 45nm dual and quad core chips but your spreadsheet makes the calculations much easier so thanks!

Here is my main concern with this board SIIE.. It overclocks awesome everything is great in games and performance but with normal mundane tasks like data transfer, creating ISO's, installing programs I get constant CRC errors, memory fault in quickpar, etc... I have a small window to return this board and my question is honestly are these issue cureable with a bios update.. More specifically will these issues be resolved without a hardware revision?? I just don't feel confident with ASUS they have left me out in the cold a number of times they are not a great company.. I don't want to be stuck having to down clock my machine all the time to do regular things.. Not Cool.. Honest opinion would be appreciated... My opinion this is another fatal flaw from Nvidia and although it will get fixed eventually SIIE owners will get totally screwed in the typical ASUS fashion.. Evga will offer great support and ASUS will give you the finger.. What do you think???

If I was in the case where I should decide, I would wait until next weekend to see if we can at least receive any news since that nvidia guy told dinos22 2 weeks. I dont know, we can also spam pmp to send us another BIOS to play with, jk :P

Guacomole - That is a very good question and one I think everyone would like answered! From what I have read at various places there is a more than likely chance that this "Data Coruption" problem is caused by something a BIOS update could take care of (although no one is willing to commit to that and I haven't heard anything concrete from vendors or Nvidia)? Granted this doesn't answer your question but like Slim142 said I would hold off as long as you can and see what happens. As you mentioned this board is quite impressive despite these few major issues, if they got fixed this board would be a great board!

As you also mentioned however Asus customer service is all but a joke (coming from my own personal experience) so if these issues turn out to be hardware related I think we're all going to be left in the cold. My own plan is to wait a few more weeks (maybe even a month) and hope for the best, if there is not at the very least talk about a fix in the works than I'm most likely going to put the board up for sale and take my chances somewhere else. I don't want to turn this into a rant against Asus but I'm always surprised that a company like them doesn't take customer service more seriously as their products are for the most part great! Last but not least, whats your plan if you don't stick with the S2E?

he just ordered a 780i board... LOL

I mean I really love this board but I am highly considering going back to the 780i for the interim.. I just need my computer to work all the time under all conditions and my real concern is being stuck with a $450.00 Mobo that has corruption issues while OC'd.. I would totally wait it out if somebody said hey bios update will fix the corruption.. I really don't want to get stuck with $450 paper weight or have to RMA through ASUS and wait 3 weeks by then I will have to buy another board anyway.. So I am thinking about sending this thing back, picking up SII Formula and waiting it out that way.. Not the best choice but at least everything will work for awhile until they get it figured out.. .

Here's he thing ASUS has never offered to fix exchange or swap anything when their products don't hold up.. They got me on the entire P5N32-SLI series of boards I think there were 3, then again on the Premium "my fault" and the Striker Extreme. The only reason I bought this board is because it will be my last board until the new series of processors hit and Nvidia can't get me anymore but here we are same old S*&^&#37;t..

Don't get me wrong this board is awesome it OC's better than any intel I have used and I am profoundly happy, overall its amazing.. Love the DDRIII and all the features especially the WB on the NB... It flies my memory performance is great I just can't wrap my head around the corruption..

BTW bios 603 is where I have all my success in OCing and my buddy just called me to say he has his board rock solid at 1811 FSB and his OCZ 2X2GB 1600Mhz memory linked and synced at 1811 prime stable which gives him 4.3 on the CPU read of 12K + and latency in the low 40's that's pretty good now I just wish he could burn a movie, or transfer some files..

Guacomole, we have to keep in mind that at least nvidia is aware of the issue and "supposedly" they are trying to fix this issue. I wish next week they come up with a couple of new BIOS because this is seriously getting out of control. This is not only here but on the ASUS forums and everywhere else.

On saturday, I will experience my own SIIE and will come back and give you a review of what my experience was like. What I can tell you is that even when I couldnt check if my OC was 100&#37; estable (just ran orthos for 6 mins, lol), I havent install the nForce drivers and everything seemed good. Dont know if this might be also a cause for the BSoDs but Im just saying what I experienced with the board for the hour 20 mins I used it. Basically I just wanted to make sure my hardware was working and see if I needed to RMA anything.

Quote:

---

Originally Posted by GAR

OK fellas, ive been running for 2 weeks without ANY issues on my eVGA at 475fsb, i dont wanna jinks myself LOL, everytime i post my stable settings, the next day i get a failure or some sort, anyway here it goes.........

Voltages used are as follows in the bios for other people with similar setup to help you out with 475fsb............Following settings is what i use, my computer sometimes stays on for days, i have played crysis for over 2 hours to test, played team fortress 2 for hours, every game i have thrown at it, handles with no problems, passes prime95 for hours, blend and small....good luck

FSB = 475x8, all spread spectrums disabled, cpu features disabled, P1/P2 enables.
cpu = 1.4250 in windows 1.39 idle, 1.33 load
memory = 1.900
fsb = 1.400
spp = 1.500
mcp = 1.600

---

Nice to see you back gar, and also nice to hear those good news and that you are finally estable (so far). 475Mhz is really nice for a quad specially with 8x.
Sorry for the question but FSB is also CPU VTT right? Remember Anandtech mentioned on the little table that 1.36v for the VTT is the maximum recommended for 24/7 operation. Try to see if you are estable at 1.36v because you dont want to fry that $380 cpu.

Quote:

---

Originally Posted by Roger_D25

Guys what BIOS are you running? I ask because I've posted the same thing, with my E8400 FSB1600 was no good at all. Once I went up to FSB1605 though all issues went away (BIOS version 0511). Interestingly however FSB1600 works just fine with my QX9770 (same BIOS)? I only installed this chip late last night but I gamed for a couple hours, did some minor 3D benching, and ran Prime for a bit over 1hr with no problems. Granted I didn't Prime overnight (although I will very soon) I wasn't even able to boot into Windows at FSB1600 with E8400!

Techjunkie - Thanks for posted that table, much apprecaited! I recently printed off Anandtech's S2E review where they get pretty deep into the GTL settings with 45nm dual and quad core chips but your spreadsheet makes the calculations much easier so thanks!

---

here i'm running 0601 with qx9650 my friend has the saqme prob with q6600 and e8500 same bios

1620 PRIMING GREAT
I did 9 multi as TEST RUNS GREAT!!!!!!!!!!

Quote:

---

Originally Posted by Roger_D25

Guys what BIOS are you running? I ask because I've posted the same thing, with my E8400 FSB1600 was no good at all. Once I went up to FSB1605 though all issues went away (BIOS version 0511). Interestingly however FSB1600 works just fine with my QX9770 (same BIOS)? I only installed this chip late last night but I gamed for a couple hours, did some minor 3D benching, and ran Prime for a bit over 1hr with no problems. Granted I didn't Prime overnight (although I will very soon) I wasn't even able to boot into Windows at FSB1600 with E8400!

Techjunkie - Thanks for posted that table, much apprecaited! I recently printed off Anandtech's S2E review where they get pretty deep into the GTL settings with 45nm dual and quad core chips but your spreadsheet makes the calculations much easier so thanks!

---

Same for me...1600MHz memory failed Memtest with QX9650 irrespective of FSB, with QX9770 it runs fine at 1600MHz in Memtest!
I don't understand!!

Just wanted to mention real quick (I just noticed) BIOS 0512 is in the ASUS FTP if anybody wants to officially download it from the ASUS server. However, no changelog available.

guys, I have some amazingly good news!
I'll let the picture speak for itself :D

http://i37.photobucket.com/albums/e8...ntitled-17.jpg

Hey sono, you forgot to post your settings!!!
did you mess with the GTLrefs???

Ok guys, here's my settings:

FSB: 1811 - linked & synced
9.5 multi for 4.3 on my proc
PCI 1 & 2 - 100
SPP - MCP - Auto
all spread spectrum and cpu oprions disabled

Voltages as set in bios:
vcore - 1.468
pll - auto
vtt - 1.36
ram - 2.0
NB - 1.56
SB - 1.6
GTL ref 0 - -29
GTL ref 1 - -32
GTL ref 2 - -32
GTL ref 3 - -29

I was completely prime stable with the above settings, but with my memory at 1600 (unlinked) and much lower volts on mem and NB but still had corruptio issues (AEB sfc scan)... I've been busting my @ss trying to get this ram running at 1:2 w/ my fsb but it really is tough work. As I neared stability, I decided to run sfc scan for sh1ts and giggles and low and behold - "no integrity violations"!!!

I'm still working on getting my system stable at this point though. I'm golden in prime between 1024 and 384 and small FFTs are rock solid. 2000-4096 I get freezes and blue screens though. I'll get this figured out though :)

whats really interesting is that at 1811 FSB w/ ram @ 1600 = corruption issues, but 1811 synced = no integrity violations... trip out huh?

after my day of playing here, I really do believe this is all caused by VTT/GTL issues. the fact that I can dial in "correct" GTL settings and be rid of corruption problems speaks to that. I'll keep working and we'll see what happens

I dunno why anything above 10 multi is BSOd, that's what I'm experiencing.
I'm already down to 1.32v for my 9 multi x 405 for 3.64 GHz...
Anything above 9 multi (Haven't tried .5s yet) is BSOD in windows... Any reason to why?

This qx is a monster, I can't believe it takes such low volts to OC it properly... I"m going to clock it down even more soon!

Any good recommendations for an HSF? I thought the Intel Extreme one was good but on the quiet fan mode i'm getting 87ish deg c :(

@ brolly - I had similar issues trying to move from 9 multi to 9.5 changing MHz from 4.07 to 4.3. VTT was the answer for me

Anyway I can't seem to get 2056-4096 stable in prime. Trying NB from 1.5-1.6 and ram from 1.9 to 2.06v... Nothing but blue screens :( any ideas? Keep in mind, my overclock is 100&#37; stable w/ my ram unlinked at 1600 with the above settings, only nb at 1.46 and ram at 1.9

Really wanna figure this out as this combination of settings has eliminated corupption in my machine

Help!!! LOL

your volts are very high for your CPU though man. I am happy with the current votls (1.325 vcore 1.32 vtt) and I'm pretty sure I can move it to 9.5 np.
What kind of temps are you getting on H2o?

Update with my QX9650 and 4GB OCZ EB sticks.

I got it to pass memtest for 3 complete passes with 1610 Linked and synced with 7-6-6-24 2T timings. But cant get it pass SPI 32M. Either SPI crashes orI get a BSOD. I noticed that if you are running tight timings, it requires more SPP voltage. For example I needed 1.44 SPP voltage for 7-6-6-24 to pass Memtest, but since I cant do squat with these timings in windows, I relaxed the timings to 7-7-7-24 2T and 1.40 SPP.I passed SPI 32M and and am priming now.

http://img142.imageshack.us/img142/1...2fsbck6.th.jpg

I don't think it's my RAM though , as I am on my 2nd set of sticks. Now I am convinced its the chipset. Maybe it will get better as the BIOS matures.

Anybody here able to run the OCZ EB 4GB sticks @ their rated timings(1600 @ 7-6-6-24)??