Im running my 3110 at 520x7.5 with 4 gigs and perfectly stable. Great board. Longest lasting board Ive had in a while
Printable View
Im running my 3110 at 520x7.5 with 4 gigs and perfectly stable. Great board. Longest lasting board Ive had in a while
I'm having a tough time overclocking my Q9550 on my Abit IP35 Pro board.
First question. While researching my problem I've seen some references to the CPU PLL. This isn't in my BIOS. Does the IP35 Pro have this under another name? Is it the same as ICHIO 1.5?
Second question. Are P35 boards bad at overclocking 45nm quad core CPUs or something? Would I have better luck with a P45 board?
I can get my Q9550 up to 3.825GHz, which makes me happy. But I'm not happy with the 1.39vcore/1.23vtt or 1.38vcore/1.30vtt I need to do it with.
cpu multiplier: 8.5x
fsb: 450MHz
speed: 3825MHz
vcore: 1.38v
ddr2: 2.1v
vtt: 1.30v
mch: 1.25v
ich: 1.05
ichio: 1.5
ddr2 ref v: 0
gtlref 0/2: 67%
gtlref 1/3: 67%
tcl: 5
trcd: 5
trp: 5
tras: 18
At first I was real confused by the talk of GTLREF. But I believed most of my confusion was because other motherboards have more GTLREF options, which made me wonder what they were talking about. With the IP35 Pro we just have the two percentages, one for each core to worry about.
I read I should try to make sure the vtt + gtlref % = 0.8 to 1.0. So with my VTT of 1.3 with GTLREF percentages at 67% equals 0.871.
I'm still confused about what vcore and vtt voltages are safe for 45nm CPUs. I've seen people say don't go over 1.36 vcore while others say just stay below 1.45. Vtt seems to be over lower, I was told my 1.3vtt was even too high. *confused*
Anyway, if anyone has any tips, I'd love the help. So far I seem to be stable (really need more testing though), but it seems like I'd like to find a way to do this with lower voltage if possible, even if a newer P45 board is the answer.
I have a Q9550 like you and I am very happy with this board. ICHIO helps to increase the memory bus speed. CPU GTLREF has a small effect on the overcloking, I found that 63% was better than 67% for 45nm.Quote:
Originally Posted by Ravynmagi
I did some search about max CPU VTT, it's recommended not to exceed 1.4V
For max Vcore with a good cooling (water), it's recommended not to exceed 1.5V.
I am using 1.37 for CPU VTT and 1.46 for Vcore (not drop, I have done the Vdroop mod), it's running fine for few month so far and prime95, orthos, OCCT stable.
http://pagesperso-orange.fr/sebastien.vignieu/uGuru.png
http://pagesperso-orange.fr/sebastie...eu/Everest.png
You voltage look safer than mine. I wouldn't be worry if I were you, getting a new motherboard for lower voltage is definitely not worth it.
Official bios 17 is one of the best for the Q9550.
Hope my post help you a little.
Show us some benchmarks, VIPE!Quote:
Originally Posted by The Viper
Hi guys, I was wondering if the problem persists for all of you or just me. System is stable when i OC in Bios, but when I use uGuru and try to OC it hangs that same second. Now some days it doesn't and its really great i can just OC really quick for some games and then downclock to save power. Is there any fix for the uGuru to work. I got the latest Abit uGuru v3.109
please! thanks
Quote:
Originally Posted by The Viper
Rebels Haven has posted a mod of the 18 Beta bios.
Ace,
Question for you? I had the board IP35 Pro for almost a year now running great with bios 11 system include Q6600 oc to 3.4 vcore 1.360 425x8 gskill memory 4gig at pc1066 and corsair hx620 ps never care to update the bios. I just upgrade the bios to 14 heard a lot of good thing about it and swap ps supply to corsair HX520 and now can't oc anymore have tried all bios 11, 14, 16, 17 no go will boot to windows but as soon run OCCT only last less than 5 minutes lock up, had the vcore up to 1.470 still no go, is there a problem with this board running with the corsair hx520? or I should do more of proper cmos clear like pull out the battery and such....
Any info will be grealy appreciated.
do you have CPUz opened while you have Uguru running?...if so, close CPUz out and then try again.Quote:
Originally Posted by Goone
if that is not the case, i would clear CMOS with the PSU power cord unplugged, CMOS battery out and then push the power button to unload all the caps...leave it out for about 5 minutes.
what bios are you using?Quote:
Originally Posted by lecter
have you cleared CMOS for 5 minutes with the battery out, unplugged the PSU and hit the power button to unload the Caps?
Thanks Ace for the quick reply, yes no cpuz running with uguru only uguru by itself heard they don't go together quiet well, and still no go I will give the proper cmos clearing battery? anything else I should be aware of
thanks again
Maybe with your GPU and other peripherals, the HX520 just isn't enough.Quote:
Originally Posted by Goone
Only had one gpu 9800gt, 2 hd, and 2 sata dvd-rom
have you tried putting the old PSU back in to see if it returns back to your previous OC without any interruptions?Quote:
Originally Posted by Goone
OK Ace, put the old power supply and can't get the previous OC here is mine current setting
Q600
Vcore 1.3850, MCH 1.44, CPUVTT 1.23, ICH 1.05, DDR 2.10, ICHIO 1.60, DDR REF 0%, CPU CTLREF0,1 67, CPU CTLREF1, 2 65
rest of system Q6600, IP-PRO35 BIOS 14 corsair hx 620 2 sata hd, 2 sata dvd, GPU 9800GT, what could be wrong
I could do this easily at 1.3650 before.
Goone:
Maybe these links can help:
http://forums.HEXUS.net/abit-care-he...-boot-now.html
http://forum.uabit.com/showthread.php?t=129194
I read several more at the time about issues with the Corsair HX series -- and a few other multi +12v rail PSs. However, I think I only bookmarked these two.
-- Hans
what that tells me you were on the very edge of the OC and disturbing the delicate balance, you can not return there...SO...you have to accept something less for an OC....OR....try the below:Quote:
Originally Posted by Goone
1) i would not use differential CPU GTLREF's.
2) try higher MCH and higher ICHIO.
3) AND...Vcore might need to be higher.
Thanks Ace, what do you mean by would not use differential CPU GTREF? you mean by number for those is no good or just leave at default which is I believe 67?
I will give it a try with higher vcore, mch, and ichio
My system has been running extremely sluggish when burning. The mouse also starts to get jerky and it takes over an hour to burn a dvd when it should only take about 6. Any idea what the problem could be? Running and IP35 Pro with a q6600 with XP Pro SP3
Ok, I'm back again! Did the bolt mod on the mainboard finally and the PWM temps dropped 20C.
Updated rig specs in sig. I am going to tweak the MCH and ICHIO voltages when I find the max FSB.
Using IntelBurnTest for stability testing. Current results so far:
9x Multiplier
430MHz FSB
Vcore 1.2625 in uguru, about 1.29-1.30V displayed when loaded. Minimum I can do at this FSB/Multi combination. Temps max around 70C
DDR2 2.2V, 5-5-5-15 timings (still loose to find max FSB)
MCH 1.25V: 1.41V
ICHIO 1.5V: 1.70V
Vtt: 1.10 (found no difference in stability with this raised.
DDR2 Ref: 0%
ICHIO 1.05V: 1.05V
GTL refs both at 67%
Got my PWMs up to ~70C when max loaded. Think I need some spot cooling on them and the NB before I up the CPU voltage too much more.
Going to lower the multi and gun for a higher FSB. Any truth to the "FSB holes" encountered in the 430MHz area?
Dropped the multi to 8 and started bumping the FSB more.
Got to 440 stable but hard locks at 450.
Maxed out Vtt, moved ICH 1.05 to 1.09 and bumped Vcore and still locked.
*edit*
Lowered multi to 7, upped MCH to 1.44 and ICH to 1.8, lowered GTLREFs to 63%. Got to 470 stable. 475 hard locks.
*edit 2*
470 is completely stable lowering all voltages to default but leaving MCH at 1.44V. I lowered the FSB to 450 to give some headroom and am now bumping the multi higher. Lowering the voltage slowly on Vcore to see where the chip is happy. Right now I have 1.245V at 450x8, 3600Mhz.
Can I really expect to get to 500FSB if I adequately cool the NB or should I be content with 450-470?
These would appear to be my sweet spot settings:
450x8, 3600 MHz
Vcore on load: 1.205V
Vmem: 1.9V
4-4-4-12-1T timings
MCH 1.44V
GTLREF 63%
All else stock
These settings give no errors in IntelBurnTest. Temps are 60C on that program, 50C on Boinc Seti. Trying for 450x9 (4050MHz) required 1.35V and caused PWM temps to rise to 80C, CPU to 75C. Unacceptable temperatures without water cooling on CPU and air/water on Mosfets.
Considered running 4:5 or 5:6 memory timings but that will probably require loosening timings.
I'm very happy with what I've achieved today with this, but if anyone has any other suggestions for tweaking things for further gains, I'm open ears.
Hey guys
I'm new around here and i would like some of your opinion...
I have this board and the best settings i got with my Q6600 G0 are the following:
378x9 = 3.4GHz
Vcore = 1.49 BIOS 1.47 with vdrop 1.43 with vdroop
CPU VTT and all mobo voltages are default...
Reading the first ten pages of this topic (and some others, found in google), i have decided to play a bit with the mobo voltages, since you all guys said it can make the overclock more estable...
Problem is.. it didn't help me at all... I still can't make my Q6600 to fly higher without a prime failure or a blue screen... I'm getting dismotivated already...
What do you guys think? Did I reach my cpu's FSB wall?
I was thinking about making a vdroop mod in the motherboard... would it help?
Thanks..
Quote:
Originally Posted by cadupradoo
when you move up the FSB, you must to increase MCH and ICHIO voltages (NB & SB respectivelly) too. Moreover, you must to find the best configuration of CPU VTT voltage and CPU GTLREF parameters, because they 'll help you to stabilize the system. What DDR2 voltage are you using to memories? Maybe, the DDR2 voltage isn't enough (set it at 2.1-2.2v to discard to memory inestability) and you can relax the latencies at 5-5-5-12 or 5-5-5-15 to reach the stability.
In my configuration, i found that CPU GTLREF 0&2 65% and CPU GTLREF 1&3 63% are more stable than 67% (default value). Even so i think that i'm using a vcore too high.
http://img212.imageshack.us/img212/1...5vv2es3.th.jpghttp://img212.imageshack.us/images/thpix.gif
http://img230.imageshack.us/img230/6...o599zo8.th.jpghttp://img230.imageshack.us/images/thpix.gif
Anyone here have IP35 Pro + E8600?Any issue??Because AFAIK there are some issue on it.Is it true??
Unfortunately, it doesn't seem to work, sorry.Quote:
Originally Posted by OC4/3