Printable View
What would you say the max NB volts should be put through a DFI P35? I'm currently running 1.71v (everest reporting) on mine with a fan directly on it (50c is max I see with fan on it full load). When I buy my new chip I will be WCing the NB.
q6600 at 3.6ghz (400x9) took 1.47v (borrowing from work till I get my new chip)
8gb (4x2gb sticks) PC28000 (2.0v)
NV 1.71v 1.6v VTT
Took a lot of volts to get stable with 8gigs of ram. How much more you think I can go or is it all based on temps?
Thanks
Raising the voltage, especially on the NB is not an option for me.
As you guys might now that on the Asus Striker II Formula, their NB chipsets run quite hot at default voltages.
Mine can range from anywhere from 60 to 70. :(
Vio|37
Maybe... but before this, I really did have to use high voltages for 3.80GHz. Stupid me... that wasn't necessary. There is a golden voltage ratio that I discovered. If I left everything like a mess at high or near high voltages, the system would be very unstable. But IF I go with just enough vCore, vPLL, vFSB, and vNB, the system would be rock-solid, and I believe that can happen with other boards as well. Of course it's trials and errors, and I tried over 20 different combos for several hours (actually ten days now) before I could find this perfect ratio for 3.80GHz at 1.304-1.312v. It wasn't even stable at 1.388v before.Quote:
Originally Posted by sofarfrome
I like :D. Now I just need to get my hands on one of those.
And yes just enough voltage is more stable then too much because there is less heat generated. That is why just throwing 1.5 volts at something is not the answer even if it can take it you need the cooling to keep up with it.
Quote:
Originally Posted by RunawayPrisoner
Care to share what voltages you came up with? I know boards will vary but it will give us a place to start.
vCore: 1.3500v (1.328v real)Quote:
Originally Posted by Delirious.
vPLL: 1.58v
vFSB (VTT): 1.32v
vNB: 1.55v
3.80GHz rock solid. The only thing that has problem is OCCT, and I wouldn't bother... since I won't be running at anything that is as intense as OCCT. Primed 10K for at least 6 hours straight.
how long before you fail in OCCT?Quote:
Originally Posted by RunawayPrisoner
Yeah, the x3350 works fine in it. It is stable at this speed.Quote:
Originally Posted by graysky
http://www.xtremesystems.org/forums/...d.php?t=182445
Though it was prime_blend stable then, I had to make some changes since then. Running Prime Large FFTs, Folding SMP, 1080p video, and mp3, I would not fail prime, but would crash. I have since corrected this by upping my NB GTL and NB voltage. it is now 1.63 in windows. Been running the same load for the last 13 hours without issue.
I like the board, but I see other people getting higher speeds and I cannot go higher because my VTT is already maxed. I would probably go higher on 4GB ram, but I need 8 damnit!
yes, I have spent many hours on this with no luck. I have a different chipset than you, and I also have 8GB of ram. I have a feeling this is related. The chipset has to work harder with 8GB.Quote:
Mmm... did you guys ever try... backing down the VTT and the vPLL? I think 1.6 VTT is just too overkill. Without any GTL modification (I don't even have the option to), it just requires me 1.32v VTT for 3.80GHz stable. Are you sure... that that's necessary?
40 minutes. Does so every single time without any reason at all, and I can't excuse the fact that I was running OCCT in Windows Vista. What bugs me is that the system is completely stable with all other stress tests no matter what I threw at it. Actually, 450 would fail that way as well. I'm suspecting my RAM sticks, so I'm getting another pair to see what would happen.Quote:
Originally Posted by Nebulus
Well, I have 4 sticks of mixed Kingston and Crucial RAM running in here. Needless to say... compatibility may be an issue already.Quote:
Originally Posted by Seraphi
Cool man... I think I might actually go with the P35 version (the LT P35) since I really want a P45 board and am unwilling to wait... for $170 I can use it for 6 months or whatever and get the P45 and hopefully sell it for 90 or so.Quote:
Originally Posted by seraphi
What are you doing that you need 8 gigs of RAM?
yes, I have spent many hours on this with no luck. I have a different chipset than you, and I also have 8GB of ram. I have a feeling this is related. The chipset has to work harder with 8GB.[/QUOTE]
Just got my Q9450 last Friday-- 3.45 stable on stock volts
Pushed it a little to 3.5-- but backed off. I may try again this week to get to 3.8 or so.
Asus Maximus- 4 Gb G.skill ddr1066
I'm also stuck at quite low FSBs with a Q9450(ES) and a DFI LP X38. With aircooling the most I can get Prime Blend stable is 440x8 with these voltages:
CPU Core: 1.3625V
CPU VTT: 1.503V
NB: 1.504V
RAM: 2.190V
CPU GTL 1/3: 102
CPU GTL 0/2: 103
NB GTL: 103
At 450x8 the most I can do is a quick 3DM06 run. Any longer load at that speed will lock it up.
I had high hopes for this chip since it went rock solid at 400x8 with all stock voltages... :(
Recommendations?
Update:
After cranking up the voltages I got the FSB to 465 (using SetFSB). Anything higher would lock instantly.
CPU Core: 1.40V
CPU VTT: 1.603V
NB: 1.76V
RAM: 2.190V
CPU GTL 1/3: 110
CPU GTL 0/2: 109
NB GTL: 118
Trying at 458 now to see if Prime will run with these voltages.
Those are insanely high IMO... try tweaking the GTL?
This is what I current have on mine ... 3.6GHz
vcore: 1.2500v
vPLL: 1.6
vdimm: 2.15v
vFSB: 1.4v
vNB: 1.55v
vSB: 1.05v
FSB: 450
RAM: 1124 (OCZ PC2-8500)
Load line Calibration: Enable
Trying to go higher but failed OCCT test when FSB set to 460 .... Don't know if problem from PSU or not (PSU: Targan 480W)
Ambient: 25.6*C, loaded 100%CPU: 60*C (max)
Try Prime95 instead of OCCT?
Picture please :)Quote:
Originally Posted by sbinh
On first page .. :p
I don't have any prog. to display voltages ... but CPUz should displays vcore ..
http://img215.imageshack.us/img215/8...maxtempaf8.jpg
With OCCT display the vcore ... I ran OCCT, you can see the pic on first page.
http://img229.imageshack.us/img229/4...emp2kf3.th.jpg
Okay... I volt-modded the motherboard. OCCT seems stable now. :) Does it mean... 3.80GHz is FINALLY stable (with all stress tests)?
If that's so, I'll tackle 490 immediately. ;)
Edit:
http://i235.photobucket.com/albums/e...OCCTstable.jpg
I am content. :) It's official. 1.304v for 3.80GHz is... achievable!
BUMP!
http://i235.photobucket.com/albums/e.../PC/OCCT39.jpg
Let's see what happens. :)
So has anybody hit a fsb wall yet?
I was thinking of buying a q9550 for the higher multiplier. But after reading all the posts of people having to raise NB voltage like crazy i am having second thoughts.
I currently just upgraded to an EVGA 780i and water on the NB is not an option for me right now so im concerned about NB heat issues playing a big role on how high i get my oc. I want to at least hit the big 4Ghz mark. Should i just go wit the E8500?
Get the E8400 if you just want 4GHz.
Get the QX9650 if you want easy Quad 4GHz.
And get one of these chips if you want to race with the FSB. ;) And only half an hour before the final result of 3.90GHz!
Edit: Failed... awww...
yea i really been wantin a quad for a while now. So do you think i might be able to come close to or hit 4Ghz wit a q9550 and still manage to keep my NB temps under control with the stock EVGA heatsink an fan? And use it for a 24/7 overclock?
Uhm... I'm not sure, to be honest. If you have a golden chip, you might be able to hit 475MHz stable, and that'll be just 4GHz. But the chances are slim, and even at 475MHz, I still need my NB volt at 1.55v versus 1.25v default. How is your NB running with default volts on water cooling?
well thats the thing, im not under water up north. Its all stock