Printable View
Have you guys been able to get decent fsb with quads with this mobo? I was never able to get higher than 450 on my older DFI P35 UT board or even the DFI X48 LT board. I just played with a few settings and was easily able to boot at 470 with Maximus II formula...I haven't done any stress testing yet...that will follow soon. I'm thinking I may even get this CPU to 500! I haven't even changed any GTL settings yet.
I try 2 Q6600, they both reach 510 "easilly", don't try higher fo now. On MF1 with rampage Mod, I was able to reach 470 MAX (even with really high Vnb). But they don't boot with Strap 400 on MF2.Quote:
Originally Posted by ACE76
On a X3350, max reached: 490. But really hard to do something nice with it. There's some strange errors :mad:.
And it need too much Vnb to run at high FSB. Something like 1.60v...
e8400 is the easiest to OC :p:.
But I have to look closer to do something better than FSB 540.
By the way, in auto on CPU PLL, the board gives some 1.72v :eek::shocked:, that's crazy.
Bios 802, 701 were tested but I prefer 604.
:)
Hiya Mr-rob, how's bios0604 better than the others, could you elaborate a bit more ?Quote:
Originally Posted by mr-rob
Strap 400 don't work on either 0503 and 0801 for me. I already tested with Ballastix pc8500, Mushkin and Dominator and all failed.
My Q9450 reached a max of 475 on bios0801 and Prime95 blend-test stable for 12+hrs but failed OCCT within 30 mins. The best I
cab do atm is 450 to be OCCT stable.
I was gonna try 0802 but perhaps I may backflash to 0604 and see it goes.....
My current settings:
fsb=450, memory @ ddr1080 on strap333. AI Booster=strong, PL=8
vcore=1.3125
vTT=1.325
vPLL=1.6325
vNB=1.51075. yeah I need more
The northbridge does get hot under load. Was planning to watercool the NB but now I won't bother to because the heatpipe assembly is all in 1 piece. Too much work for me. :(
Priming a 8400@445*9 atm with
vcore=1.325 (cpuz 1.304 under load)
vTT=1.325
vPLL=1.59xx
vNB=1.3
I think i need to play around some with GLT but I cant find any good reference on how this works. All i can find are people recomneding to set them at x61/x63/67 and similar but this board gives mV changes... does anyone have any good reference on how to set these up or is it just trial and error?
I run some tests with bios 701 & 802. No way to do something nice.Quote:
Originally Posted by clee
That is the reason why I prefer bios 604.
Strap 400 works with X3350, but I can't get high FSB.
This are some screens with bios 604:
-X3350:
http://images4.hiboox.com/vignettes/...fc0a5a9496.jpg
-e8400:
http://images4.hiboox.com/vignettes/...9d90af115d.jpg http://images4.hiboox.com/vignettes/...bc1efef874.jpg
-Q6600:
http://images4.hiboox.com/vignettes/...47565fa846.jpg http://images4.hiboox.com/vignettes/...b87ddc4f36.jpg
This tests are not the maximum reached on this board, just some quick parameters...
Quick toughts:
-This mobo needs in most of case less voltage than MF1 (my Q6600 is able to run fully stable @8x450, Vcore 1.30v Cas4 on Muskhin Xp˛ and also with Red Tracer PC6400.
-X3350 needs more Vnb to run high FSB. Also high CPU PLL...
-A nice mobo for OC, but I don't konw yet if it can be a real performer just like MF1.
:cool: This is my solution for WCing the NB:
http://www.xtremesystems.org/forums/...d.php?t=194858
:D
That's all for now. Next time, I'll try to make sonething nice with my e8400
:rolleyes:
Enjoy :)
y??Quote:
Originally Posted by Seewolf
how u think i was running at 400fsb 1066mhz??
strap 400 is usable on p45's more than x38's/x48's since the tRD is now same as strap 333...
on x38's/x48's it takes a tRD hit..
hmm i've been looking at all u guys on ure settings...
why arent u all adjusting ure gtl on 45nm proc.. u should.. if not ure just overvolting ure proc..
calculation is simple
this is for 500fsb on crossfire...
vtt 1.2v nb 1.2v
gtl 0/2 = 1.2v x 0.635 - leave this at auto
gtl 1/3 = 1.2v x0.667 ( too high)
to change
1.2v x 0.667 + Y= 1.2v x 0.63
Y = 1.2v(0.63-0.667)
= -0.044
so u should change gtl 1/3 to -45vm
for 400fsb on crossfire
vtt 1.12v nb 1.12v
1.12v x 0.667 + Y = 1.2v x 0.63
Y = 1.12v( 0.63-0.667)
= - 0.04144
so u should change gtl 1/3 to either -40vm or -45nm depending on ure voltage flux on load..
have fun maximus 2 users
for 65nm
u should run 0.667v for gtl 0/2..
so u should + vm on it..
u will find this board much more stable.. and less voltage on vcore if u do this and make sure ure vtt is = or more than ure nb..
follow this settings first to find ure vcoreQuote:
Originally Posted by Legacy
strap = 333
Dram Freq 1000
Ram Setting - leave it auto for now
dram static read control - disable
dram read training - disabled
mem oc charger - enabled
ai clock twister - moderate
ram .. everything on auto right up to
transaction booster = 10
vtt = 1.2v
vnb = 1.2v
pll=1.5
ram= 2.1
sb etc 1.1,1.5v
gtl 1&4 = -40/-45 depending on ure vcore flux
gtl 0/3 = auto
loadline enabled
spread disabled
cpu settings .. all of them disabled..
have fun
keep increasing ure vcore till ure stable.. i advice using orthos blend priority 10.. with no background programs running..
Can you give an example for 65nm cpu, in your equations above you have their volts as the same, so no reference for me to figure which goes where.Quote:
calculation is simple
given VTT = 1.35, vnb = 1.30
ok say ure using vtt 1.35vQuote:
Originally Posted by nettwerk2006
so adjustment for gtl 1/4 is null
therefore u need to adjust gtl 0/3
so how much easy
1.35v x 0.667 = 1.35v x 0.635 + Y
Y= 1.35v( 0.667-0.635)
Y=0.0432v
= 43.2mv
i usually roundup.. so u should +45vm on gtl 0/2
havent really tested a 65nm quaddies on p45's but from experience in the past with other boards.. 65nm works best on gtl 0.667/0.67
Thanks for info, I figured it outQuote:
Originally Posted by cstkl1
What about for NB GTL?
would I vtt, and .667, board default appears .64?
1.30 (.667) = 1.3(.64) + y
y = .0351
Attachment 82496
?
cstkl1 ... does changing that bring a stable higher OC? or just higher FSB with a lower multi?
Quote:
Originally Posted by cstkl1
Thanks for all your help so Far, but I have a question. you say to change my GTL 1&4 to -40/45 but the only GTL's I see are 0,1,2 and 3 I don't see 4 unless NB GTL is 4 ..
I have changed everything in the Bios to the settings you have suggested.
waiting on your reply
The ones you would need to adjust are CPU GTL Reference 1 & 3, because they are set at .67 which would be good for a 65nm cpu, but not a 45nm like yours.Quote:
Originally Posted by Legacy
Ok so I set the GTL Reference 1 & 3 to -45 and used all the other settings that were mentioned up above with a Vcore of Bios 1.41 CPU-Z 1.40Quote:
Originally Posted by nettwerk2006
Failed orthos :( 7 min
Do you think my chip just sucks?
I have seen people get there E8400 to 4 Ghz with low voltages
could be ure proc...Quote:
Originally Posted by Legacy
could be ure rams..
hmm try upping the vcore till u get it stable..
ah sorry guys
i meant 65nm change gtl 0/2
45n change gtl 1/3
sorry about the mixup
nb leave it.. cause this is chip dependent..
this are p45's they are suppose to be running low voltages on nb..
I raised the vcore as High as 1.45 with no luck, comp lock up when it starts.
I don't think I want to raise the Vcore any higher :D
I even get weird little graphic glitches on the screen
hmm have u tried 3.5ghz??Quote:
Originally Posted by Legacy
and are u sure ure rams are okay??
Yeah , I ran memtest for 36 hrs with no errors.. I can run 3.6 stable 450x8 1.33 Vcore 1.2 Nb 1.2 FSB always 1:1 Ratio
@cstkl1
What GTL values for 45nm Quad and Dual Core. with VTT 1,35 and VNB 1,25 do you recommend ?
I am running your FSB 500 Settings with VTT 1,35 VNB 1,2 and GTL 1/3 +50 on a 45nm DC and CPU PLL 1,6 Loadline disabled.
Your help is appreciated very much - saved me some hours of trying.
The board pooped on me this morning :D tried to turn on the computer and it failed to boot, tried the crash free bios but nothing worked just a blank screen even tried to pull the cmos battery and still nothing so I sent it back for replacement .. hopefully I will have a better experience with the next board :)
Well, first off I'll have to say Hi!, since I'm new here.
You'll have to excuse me if these questions are already answered somewhere else, but it feels like cracking your head into a brick wall trying to find concrete answers to these, IMO, fairly simple questions... I'm not really looking to do any overclocking (guess I'll be slapped around a bit here, for that).
Will this board run 8Gb RAM (4x2) stable? Would you recommend 800 or 1066 Mhz?
Is this, occasional, NB temp issue I keep reading so much about a out-of-the-box fault, or something related to overclocking?
Will it actually impair real world gaming, that the P45 only supports 8x PCI-e when running CF? I'm looking to go 2x 4870. There's a lot of sites coming to different conclusions on this...
Any answers at all would be greatly appreciated.
Quote:
Originally Posted by silent_kebab
well for one the board has 2 PCI express 2.0 slot so the extra bandwith will make up for the 8x8 when throw in crossfire mode from what I understand..
If you don't plan on extreme overclocking your cpu than DDR 800 should suit you well and It's hella cheap !
My northbridge was at 38c out of th box.
This board can hold 4x4 GB sticks, you should have no problem running 8GB, I wouldn't recomend mis matching ram sticks though
i meant 500x7.Quote:
Originally Posted by Legacy
use the vtt/vnb voltage i gave u with the gtl's
adjust the ram and vcore accordingly.get orthos blend priority 10 stable for the first two test.. 1024k and 8k.. better if u do custom and set form 8k to 4096 with 30 mins on each test.. run for an hour on 1024k 30 mins and 8k 30 mins..
ok here is the problem with quaddiesQuote:
Originally Posted by IronAge
p45 with dualcore.. single card vs dual card.. different vnb
4 sticks.. single sided vs 2 sticks single sided.. different vnb
4 sticks.. dual sided vs 4 sticks dual sided.. really different vnb
throw in a quaddy.. .. well u know what i mean..
each setup with quaddies are different.. as with dual core..
quaddies has pll voltage to come in play.. while dual core will bsod with any increase in pll voltage..on ram stress
ok to answer someQuote:
Originally Posted by silent_kebab
p45's vs x48's..
reason y ppl are liking p45's is cause all the motherboard temps are low. hence it can take a lot of punishment with the low voltages its running to do 500 fsb..
if ure not into clocking ure bus speeds.. than x48's with the slight advantage on 16x pcie is the way to go..
but if ure into clocking and a acceptable ( with almost negligable performance difference with crossfire on x48) crossfire performance than this board will not fail u.. its easy to clock on this without much cooling or worry on the board
seems like the newer bios rocks..:D everything is cooled with stock cooling.
http://i163.photobucket.com/albums/t...tos/600fsb.jpg