Printable View
My project going to 4GHz 24/7 @ room temp has started... I've ordered custom made cooling block over mofsets and backside of MB. I've also purchased lots of silicon tube and some splitters for PWM cooling.
If someone wants exact pics to made PWM block I can send you 3D project made with Google Sketchup.
post some pictures! and wow! thats some nice ram @ 573mhz and CAS4!
Here is two pics... I hope my friend does these blocks soon so I can make some tests with temps... I also need to buy new waterpump + block for cpu... So posting test results will take some time...Quote:
Originally Posted by xdrift0rx
http://picasaweb.google.com/Tuomo.Ve...600IP35Pro4GHz
I'll update more pics when get some...
@fatguy1992 i have just purchased an ep45-ud3p (running a q6600) and some corsair dominator 2x2 pc8500 v1.1. What are your bios setting for 3.6 as im needing a lot of voltage for it to boot into windows but be unstable.
Managed to have it stable at 3.2 with 2x1 flex XLC PC9600 @ DDR2 1000. but anything over 1000 on the ram, and it seems to become unstable even with +0.2v on the MCH
Same with the dominators will rum stable at 1066 but the cpu has problems has OCing is there something im missing (im also a little confused with all the voltage settings of this board).
BTW over 4GHZ is unreal!!!
Mine Q6600 G0 was stable enough to run 3D Mark 2006 @4.1GHz, Super Pi @4.2GHz, boot into Windows @4.3GHz with ~1.55v:
http://valid.canardpc.com/show_oc.php?id=304720
http://service.futuremark.com/result...eResultType=14
I used Gigabyte P35-DS4, 4x1GB Kingmax DDR2-1066 CL5, custom watter cooling. Was running @3.6GHz 1.35v, DRAM @1GHz 2.0v (+0.2v more than RAM specs) without an issue 24/7.
Try to incrase vDIMM. If you add +0.2v to vMCH then add +0.25v to vDIMM although your RAM should work stable with less voltage. That worked for me. Furhter, increase vFSB.Quote:
Originally Posted by arapab79
Also, with FSB : DRAM ratio set to 1:1 the OC should be more stable. Try with lower multiplier and higher FSB. Just avoid 400~450MHz FSB, try above 450.
gOJDO
Very nice :clap:
I'll be hooking my phase change cooler up tomorrow and doing some more xtreme Q6600 overclocking.
these chips arent old quite yet - nice oc.
Thanks for the reply gOJDO.
Im not real sure how to get a 1:1 ratio with this board as ive set the MCH memory latch to 266mhz as recommended and the only options are 2.50A, 3.00A and 4.00A.
Whats the safe voltage for memory, recommended is 2.1, when i add .25 the figure becomes red.
Can you tell us your full system specs(including cooling system)?
Gigabyte EP45-UD3P
Lapped Q6600
Cosair Dominators 555 15, ddr2 1066 with dominator fan
GTX260
4 x 36gb raptors
1 x 120gb hdd
1 x 500gb hdd
Gigabyte Odin Gt 800w PSU
Water cooling consisiting of 2 x 2x120 rads, dtek water block, MCP655 pump
Nice setup you have there and high voltages shouldn't be a problem with such cooling.
You should set the "System Memory Multiplier (SPD)" to 2, so when FSB is set to 400MHz, the memory frequency is going to be 800MHz (400MHz DDR). That is 1:1 ratio of FSB : DRAM. Set as such it makes the least load on the MCH and that means better stability. Your RAM already rated at high vDIMM, so we'll start with lower vDIMM.
You should have updated the the mainboard firmware to the latest.
First, the following settings should be disabled: Limit CPUID Max. to 3, NX bit, C1E, C2/C2E, C4/C4E, TM2, EIST, VT, CPU spread spectrum, PCI-e spread spectrum.
Set the following voltages:
vDIMM=2.25v.
vCore=1.45v
vCPU Termination=1.4v
vCPU PLL=1.85v
vMCH Core=1.3v
First: try FSB 400MHz, if it boots into Windows test for stability. If stable, start reducing voltages one by one in 0.05v steps and every time test for stability. That's how you'll determinate the min voltages required for stability of the given OC. Use OCCT and 1 hour test. If it passes than your CPU is going to be stable in any real-world scenario. Then you may start incrasing FSB and voltages(one by one in 0.05v steps) to find another stable OC. Reducing multiplier may give you better results(higher OC and/or lower volages).
If the system won't boot into Windows or it boots but it's not stable, then try reducing FSB or reduce multiplier and increase FSB.
What are your thoughts on throwing my q6600 in my dfi p35? I think my dfi 680i might be holding me back
Could well be. I have both the ASUS P5K-E-WiFi-AP and the ASUS P5N32-E-SLI+ motherboards. My Q6600 goes to 4100mhz + benchable in the P5K. The exact same cpu when in the P5N32 won't pass 3400mhz :(Quote:
Originally Posted by xdrift0rx
Definately worth a try IMO.
The fun begins. ATM my Q6600 is running 4.05GHz W/ 1.536vcore, 1.57v PLL and 1.3 VTT and temps r -44C idle and -38C load with my phase change unit.
:shocked: vCPU PLL is set too high, isn't it???Quote:
Originally Posted by gOJDO
It is high, but as long as it is within 0.4v to vCPU it's not dangerous. These voltages are not going to be used 24/7. These are the start(MAX) voltages, used to determine the system stability at a given CPU mutli and FSB. Then he'll reduce these voltages, one by one in small steps, until he discovers the least voltages at which his CPU is stable.Quote:
Originally Posted by mr-rob
I just did some Q6600 overclocking before.
Heres my result
http://valid.canardpc.com/show_oc.php?id=506194
But I forgot to upload my screen shot(i'm on my other PC now).
which mobo are you running that Q6600 in fatguy?Quote:
Originally Posted by fatguy1992
My UD3P rev 1.0, like it says in the link :p:
BTW my BIOS version is F4.
@ gOJDO
Set the following voltages:
vDIMM=2.25v.
vCore=1.45v
vCPU Termination=1.4v
vCPU PLL=1.85v
vMCH Core=1.3v
Thanks for the info i tried those (and occt tested - small data set) and it seemed to make my system more unstable. I kept adjusting the voltages/fsb/ up and down but the system would always lockup within the first five minutes.
I decided to load optimised defults then turn off all the C1E, Limit CPUID max to 3 ect then do a occt test - small data set, and CORE 1 came back with an error.
Im about to do another test im using the latest occt. If i get an error again is it safe to assume that my cpu stuffed, this would explain a lot of my OC troubles.
That's very impressive and If you can manage to save a CPU-Z validation I'm pretty sure that will be #1 in the HWBot DB. I think I'm in there around #3 or #4.
I put 1.68v into my Q6600 at 4.102Ghz before I called it quits with no damage done. My whole motherboard was putting out a LOT of heat though, even with everything on water but the mosfets, which have copper sinks on them.
Very well done!
Oh also, I think the Q6600 is the absolute best Intel Quad as far as OC ability and price performance goes, to me anyways. I was thinking about going to my local Microcenter to try and hand pick another Q6600 just to keep before Intel discontinues them.
im running my q6600 3.8 1.58vcore on air dont wana go further bit worryed love to get to the magic 4ghz
Do u need 1.58v for 3.8GHz, like have u tried lowering it?
That does seem very high but you're using just air cooling. For 3.8Ghz stable my Q6600 needs 1.4875 on water.
It seems you are not lucky with your CPU. What FSB and multi have you tried?Quote:
Originally Posted by arapab79