Overclocking a new Pentium D 920

[RangerXLT8]

Originally Posted by Lead Head

Im still thinking its the NF4 intel edition....

Is there any chance that the PCI-E bus may be OCing along with the FSB causing instability? like maby just a weird bug with the PCI-E bus and dual cores on the NF4?

Yes I agree 100%. EVERY 920 I've seen has hit >240 FSB when In a capable MB, and that is on default vcore.

PCI-E bus over about 110MHz will cause instability on most MBs.

Then again 533MHz memory, is not guaranteed to run any higher, but Corsair will have overclocking potential, how much is the question.

I've found it easiest to overclock with the fewest variables, that is run your memory at a low devider to eliminate that as an issue and it's one less factor. This is untill a stable CPU is overclocked, then you can work up the memory with a FSB:memory ratio.

[jph1589]

The pcie is locked at 100mhz.

The memory is at 533mhz because it really isn't stable at anything over that. It might run at 566 but i don't think that is keeping me from setting the FSB to 1000.

I've turned off all the spread spectrums.........disabled everything that isn't being used..........i've tried disabling the Vanderpool technology also.

This board doesn't have dividers that you can set. It allows you to either lock the memory at one speed or link it to the FSB. I have tried both ways. It also allows you to set the memory timings yourself or to let the bios set them loosely. I've tried both. I have also tried vcore up to 1.45v and memory at 2.0v. I won't take either of them higher than that because I just dont think they are the problem.

None of it seems to make a difference.

The top stable fsb with the 920 installed is 900. It just wont post any higher than that.

The top stable fsb with the 520 install is 1000. It will post all the way to 1050 but isn't stable.

[Lead Head]

Originally Posted by jph1589

The pcie is locked at 100mhz.

The memory is at 533mhz because it really isn't stable at anything over that. It might run at 566 but i don't think that is keeping me from setting the FSB to 1000.

I've turned off all the spread spectrums.........disabled everything that isn't being used..........i've tried disabling the Vanderpool technology also.

This board doesn't have dividers that you can set. It allows you to either lock the memory at one speed or link it to the FSB. I have tried both ways. It also allows you to set the memory timings yourself or to let the bios set them loosely. I've tried both. I have also tried vcore up to 1.45v and memory at 2.0v. I won't take either of them higher than that because I just dont think they are the problem.

None of it seems to make a difference.

The top stable fsb with the 920 installed is 900. It just wont post any higher than that.

The top stable fsb with the 520 install is 1000. It will post all the way to 1050 but isn't stable.

Its is the board...It has to be! Im wondering if the vregs arent upto the task of supplying the current needed?

[jph1589]

"Its is the board...It has to be! Im wondering if the vregs arent upto the task of supplying the current needed?"


I think the 920 actually draws less wattage than the 520 did. But I keep coming back to the fact that the NB and FSB have no voltage adjustments and that they now have two cores to keep up with.

I have a gut feel that it is the motherboard also. But I keep reading that C19-A review that Ranger did and then I'm not so sure anymore.

[gvblake22]

Originally Posted by jph1589

I have a gut feel that it is the motherboard also. But I keep reading that C19-A review that Ranger did and then I'm not so sure anymore.

You can't base your overclocking achievments off of someone elses success with similar components. Even two system with the exact same components will yield different overclocks. While it is curious to see someone else overclock the same component better than you, I can almost guarantee that there are other people who could not overclock as high as you with the same board. It all prettymuch ends up coming down to luck of the draw and how well all of your components play with eachother.

[jph1589]

Actually just to clarify, it wasn't the board review........it was the review of the XMS6400 pro.

And I also understand that there will be different results from one component to the next. This thread started just to try and determine if everything was being done right and if so, which component was the likely culprit.

I think the people who posted here will agree that it is probably the motherboard or the memory or the cpu. Anything else you guys want to throw onto the list? LOL

[gvblake22]

Originally Posted by jph1589

Actually just to clarify, it wasn't the board review........it was the review of the XMS6400 pro.

AHHH, yes, gotchya.

Originally Posted by jph1589

And I also understand that there will be different results from one component to the next. This thread started just to try and determine if everything was being done right and if so, which component was the likely culprit.

I think the people who posted here will agree that it is probably the motherboard or the memory or the cpu. Anything else you guys want to throw onto the list? LOL

HAHAHAHA, weeeellll, you can always add cooling and PSU to the list!

[jph1589]

Originally Posted by gvblake22

AHHH, yes, gotchya.

HAHAHAHA, weeeellll, you can always add cooling and PSU to the list!

Well if we want to rewind a few posts, I mentioned that the 920 is idling about 12c lower than the 520...... 26c to be exact. Someone must have replaced my Katana 775 with a watercooling system when I wasn't looking.

The PSU is probably OK since the 920 draws lower wattage than the 520. So I won't add those to the list LOL

[gvblake22]

Originally Posted by jph1589

Well if we want to rewind a few posts, I mentioned that the 920 is idling about 12c lower than the 520...... 26c to be exact. Someone must have replaced my Katana 775 with a watercooling system when I wasn't looking.

The PSU is probably OK since the 920 draws lower wattage than the 520. So I won't add those to the list LOL

Yeah, I was mostly joking since you left the question as pretty open ended. I've plum run out of ideas for ya man. I think you have just hit the wall with your hardware. :?

[Lead Head]

You cannot trust internal CPU sensors, the only way to get acurate readings would be to cut a narrow channel in the IHS, then stick a thermal probe within the channel, fill the channel with thermal epoxy, then mount the heatsink

Also, the 520's TDP is 85w, while the 920's is 95w, so the 920 DOES draw more power