You should never use PLL Auto. Never trust it!
Printable View
You should never use PLL Auto. Never trust it!
I know :)
OMG, so I tried -61mv on core 1 which is the one that gives me more trouble and it froze after 3 hours. I restarted, selected -58mv for core 1 (which lasted 13 hours of prime) and after 5 mins of use I was about to post an answer here, it froze :(Quote:
Originally Posted by golfjam
Why! :shrug:
Right now, all my hopes are on 0703. Yes, I also see weird that after 13 hours, you just freeze or restart. I will try 0703 and will see how this goes. I also want to ask something, can an OC be unestable even if you have the right GTLRef, but too much let say vcore or VTT?Quote:
Originally Posted by Guacomole
Thanks for the settings :up:Quote:
Originally Posted by Guacomole
Quote:
Originally Posted by [cTx]Warboy
Wait, whats wrong about leaving it in auto? I always get 1.52 in PLL according to PC Probe II.Quote:
Originally Posted by Broly
One last question, I havent turn off my pc since the last 2 weeks. I mean, I only restart it when I freeze while OC but since I build it 2 weeks ago, its been running with no resting time. My NB temps are on 62c and Im wondering why is so high, probably because I havent leave my pc off and its been getting tortured with no brake time? I know of people getting affected because of NB temps...
Im assuming that if you change your vTT you have to adjust the GTLref so its the same % of the vTT - or at least thats the way i have come to understand it.Quote:
Originally Posted by slim142
And my NB runs at 62 degrees too, but i always crash with SLI, so i can hardly ok that for you.
However, one point of good news, i seem to be 3d stable with 1 9800GTX now. Just SLI giving me problems, so this does mean that the BIOS's are getting better. Hopefully soon ill be able to use 2 cards. I mean, theres just something special about 200fps in CoD4.
Well yeah thats what I thought about the VTT and GTLRefs.Quote:
Originally Posted by Coupa
How come the NB runs so hot? I have applied MX-2 and cleaned them with Artic Purifier. I remember I used to have 38c in NB, well back then I didnt have any OC...
Many people seem to have SLI problems. Are you OCed? Have you tried fresh windows installation?
Really dissapointed with 0702 so far, with same settings but P1 and P2 in AUTO the system froze in 5 mins lol (with 0701 is 100% estable)
Still testing now with changed GTLRefs...
Coupa - many people reports problems with 9800 GTX SLI ...but i have no problem with two 8800 Ultras in SLI ... working perfect, but i cant to find right GTL voltages for 1600MHz FSB ...
Completely stock, and its the same windows installation ive had since i started, even when i oced to 3.2ghz for a time, ive never had corruption, so thats something.Quote:
Originally Posted by slim142
Yeah i know, theyll eventually crack it, so ill just wait it out, now that i can actually play games im happy. I mean, im only running a 22"W monitor, so i dont really need the second card to play most games maxed out.Quote:
Originally Posted by OBR
Im still to scared to load up CoH, cause that always freezes me so fast, and i can play games with one card, and if it freezes there then its not stable, and well, ignorance is bliss.
Idk coupa, I could only suggest you wait this week to see if the new upcoming BIOS fixes any of your problems.
slim try the following settings man:
NB GTL = +40mV
CPU GTL = -21mV
CPU GTL = -24mV
CPU GTL = -24mV
CPU GTL = -21mV
CPU VTT = 1.4V
NB Voltage = 1.66V
SB Voltage = 1.65V
Place a fan over the northbridge cooler to keep it cool, and try various FSBs.
The following settings should be ok for FSBs up to 475MHz ( QDR1900 )
Will those GTLs be acceptable for an e8400 45nm dual core and the like? If So, I'll test out those settigns for sure.Quote:
Originally Posted by BenchZowner
Zowner - I've been working on 4050Mhz for a few days now with only limited success (450Mhz x 9.0x). I'm going to post my major settings below, do you think they are a bit conservative for the clockspeed I'm working towards? I ask because the settings you list above seem quite aggressive, especially for anything under 4Ghz? I should mention that all settings below are direct BIOS settings (not actual).
CPU Multipler: 9.0x
FSB - Memory Clock Mode: Linked
FSB - Memory Ratio: Sync Mode
FSB (QDR), Mhz: 1800
========================
CPU Voltage: 1.375v
Loadline Calibration: Disabled
CPU PLL: 1.5v
CPU VTT: 1.30v
VDIMM: 1.96v
NB: 1.50v
SB: 1.50v
CPU GTL_REF0 Ratio: Auto
CPU GTL_REF1 Ratio: Auto
CPU GTL_REF2 Ratio: Auto
CPU GTL_REF3 Ratio: Auto
NB GTL_REF Ratio: Auto
I have tested memory extensively and its rock solid at given clockspeed and timings (7-7-7-21-1T-1-26-12-Auto-24-Auto-Auto-60-Auto). I should mention that no matter what I choose in BIOS tWR will not change, I also have limited results with tRFC? I guess what I'm getting at here, am I expecting too much with the above settings at 4.05ghz? Granted I could just crank up some settings and see what happens but I've been working hard to not have to do that, lol! I'm trying to figure out a good combination of voltages (using a little finese, instead of just cranking up the voltages)! Last but not least, you mentioned running the CPU VTT at a whopping 1.4v, is that a somewhat safe voltage for 45nm quads if properly cooled?
For the E8xxx CPUs that I've tested on the Striker II Extreme it didn't require a manual GTL selection until I surpassed 550MHz FSB ( QDR2200 ).Quote:
Originally Posted by Big Myke
Yeah, those settings can work for a dual-core as well.
However the NB GTL at +40mV was funky with the dual-core, but it did settle down & worked like a charm at NB GTL +30mV.
The GTL, and NB/SB voltages are "bonded" to the FSB, not the actual CPU's frequency.Quote:
Originally Posted by Roger_D25
With a quad-core you're putting more stress on the BUSes and the NB, thus you need higher voltages to get it stable at the same FSB & settings that a dual-core can run at near stock settings.
Depending on your luck, your CPU will have a FSB wall between 440 and 480MHz FSB ( assuming that you're using the QX9770 @ sig system )
There will always be a specific FSB that you won't be able to get past with your current motherboard + CPU combo.
Suggested changes next to your settings.Quote:
Originally Posted by Roger_D25
1.4V VTT seems safe to me ( I've been using it on all my 45nm parts since day 1 and none died or degraded at all ).
Try the suggested settings without any fear, don't sweat, it's gonna be ok ;)
Thats a massive voltage on the NB and SB isnt it?Quote:
Originally Posted by BenchZowner
If you want to keep P1 & P2 enabled at high FSBs, there's nothing you can do, you have to raise the NB & SB voltages.Quote:
Originally Posted by Coupa
1.6V might be enough, but I haven't tested it thoroughly yet ( just benching on the S2E board, regular/normal stuff when I start reviewing it ).
If you have a fan over the NB you don't have a problem with those voltages ( the 1.66/1.65V )
Once again your experience and help are greatly apprecaited, thanks!
Thanks bench, i'll give it a whirl, I too have the same problems with p1 and p2 as slim. I'll try upping the north and south to those voltages. see if i have any luck
I must have a good chipset here... 1.40V SPP @ 444FSB Linked & Synced / 1777 - 7-7-7-21 1T w/ P1 & P2 | No additional MCP voltage | Kentsfield & SLI
Wow, I really love it:eek::clap:Quote:
Originally Posted by Andreoid
the corruption problem with the 790 boards.
yes and no... at the common fsb/mem combinations the corrption issues seem to be almost completely fixed with the new bios releases
when ocing it still happens... but if you tweak the board properly most people can get rid of it even at high speeds from what i can tell.
Bench, I tried using 1.6v on the northbridge and I get corruption messages on bootup. When I bring the Northbridge back down to 1.45v I can boot back into windows fine. :rolleyes: any ideas why the northbridge voltage would corrupt windows at bootup with voltage of 1.52 through 1.6v
I respectfully disagree with this statement, the corruption issues have been solved at stock I think that is the only true place we have seen improvement (I am speaking only for the SIIE).. I personally own the board and so does my friend and neighbor and we have both tried unlimited combinations of GTL, and VTT combos to no avail.. I personally have been under the assumption that my board has been rock solid and corruption free with a heavy OC for weeks. When my friend called me to tell me he couldn't stabilize the board without corruption I ran sfc scan and the corruption was present. CHanged 1voltage the NB, rebooted ran sfc scan no corruption. Rebooted changed nothing and guess what CORRUPTION.. Whatever, I wouldn't touch this board with a ten foot poll until they work it out, because ASUS will leave you high and dry if they decide to refresh the chipset.. Just my 2 cents..Quote:
Originally Posted by saaya
1) Have you installed a fan on the northbridge cooler ? Because 1.6V without a fan is too much for the stock cooler, the chipset would be hitting 60C + temps easily, and the results can be seen ( instability, corruption, BSODs ).Quote:
Originally Posted by Big Myke
2) You mean that you used the same settings, only changed the NB Voltage from 1.4V to 1.6V and you couldn't boot into windows ?
Yes, I booted with 450*9 fsb with 1.45v to the nb and have been booting in windows just fine. I only change the northbridge to 1.52 through 1.6 and I cannot boot into windows because of corruption. I did not put a little fan over the northbridge. I would think I should have enough airflow to at least boot into windows for atleast a minute from using the side of my case fan.